Communication

Evaluation of Electric Vehicle Performance Characteristics for

Adaptive Supervisory Self-Learning-Based SR Motor Energy
Management Controller under Real-Time Driving Conditions
Pemmareddy Saiteja, Bragadeshwaran Ashok * and Dharmik Upadhyay

School of Mechanical Engineering, Vellore Institute of Technology, Vellore 632014, India;

pemmareddy.saiteja2020@vitstudent.ac.in (P.S.); dharmic.upadhyay2021@vitstudent.ac.in (D.U.)
* Correspondence: ashokmts@gmail.com; Tel.: +91-9865467729; Fax: +91-4162243092

Abstract: The performance of an electric vehicle (EV) notably depends on an energy management
controller. This study developed several energy management controllers (EMCs) to optimize the
efficiency of EVs in real-time driving conditions. Also, this study employed an innovative methodol-
ogy to create EMCs, efficiency maps, and real-time driving cycles under actual driving conditions.
The various EMCs such as PID, intelligent, hybrid, and supervisory controllers are designed using
MATLAB/Simulink and examined under real-time conditions. In this instance, a mathematical model
of an EV with a switched reluctance (SR) motor is developed to optimize energy consumption using
different energy management controllers. Further, an inventive experimental approach is employed
to generate efficiency maps for the SR motor and above-mentioned controllers. Then, the generated
efficiency maps are integrated into a model-in-loop (MIL)-based EV test platform to analyze the
performance under real-time conditions. Additionally, to verify EV model, a real-time driving cycle
(DC) has been developed, encompassing various road conditions such as highway, urban, and rural.
Subsequently, the developed models are included into an MIL-based EV test platform to optimize
the performance of the electric motor and battery consumption in real-time conditions. The results
indicate that the proposed supervisory controller (59.1%) has a lower EOT SOC drop compared to the
PID (3.6%), intelligent (21.5%), and hybrid (44.9%) controllers. Also, the suggested controller achieves
Citation: Saiteja, P.; Ashok, B.;
minimal energy consumption (44.67 Wh/km) and enhances energy recovery (−58.28 Wh) under
Upadhyay, D. Evaluation of Electric
different real-time conditions. Therefore, it will enhance the driving range and battery discharge
Vehicle Performance Characteristics
characteristics of EVs across various real-time driving conditions.
for Adaptive Supervisory
Self-Learning-Based SR Motor Energy
Management Controller under
Keywords: electric vehicles; energy management controllers; adaptive supervisory self-learning
Real-Time Driving Conditions. controller; efficiency maps; driving cycle
Vehicles 2024, 6, 509–538. https://
doi.org/10.3390/vehicles6010023

Academic Editor: Teresa Donateo

1. Introduction
Received: 28 January 2024 Electric vehicle (EV) adoption has been identified as the most optimal strategy for
Revised: 1 March 2024 mitigating air pollution and greenhouse gas (GHG) emissions. Despite the increasing
Accepted: 4 March 2024 scarcity of fossil fuels and the ever-growing demand for the same, it is crucial that the
Published: 8 March 2024
automotive industry moves toward more environmentally friendly approaches [1]. Hybrid
electric vehicles (HEVs) and electric vehicles (EVs) have emerged as pivotal solutions for
the current environmental hardships. EVs have several advantages such as negligible
Copyright: © 2024 by the authors.
pollutants, high efficiency, and quiet driving experience with minimal environmental noise,
Licensee MDPI, Basel, Switzerland. with being battery powered being the main reason for this [2]. However, to make EVs more
This article is an open access article acceptable and feasible for daily life, more advancements are needed, particularly in range
distributed under the terms and extension, battery, and motor technology. Thus, the development of EVs and the assessment
conditions of the Creative Commons of their performance remain significant areas of interest for both the automotive sector and
Attribution (CC BY) license (https:// the research community [3]. Moreover, it is crucial that the power required for propulsion
creativecommons.org/licenses/by/ must be analyzed under real-word operating conditions to evaluate EV performance and
4.0/). optimize the characteristics of the electric motor (EM), battery, controller, and converter. The

Vehicles 2024, 6, 509–538. https://doi.org/10.3390/vehicles6010023 https://www.mdpi.com/journal/vehicles

Vehicles 2024, 6 510

actual operation of the vehicle is highly unpredictable, influenced by factors such as road
conditions, energy consumption (EC), temperature, road grade, and driving behavior [4,5].
Therefore, standardized driving cycles (DCs) mandated by legislation are employed for
EC assessments of EVs. Several DCs are used globally, which include the New European
Driving Cycle (NEDC), the Worldwide Harmonized Light Vehicles Test Cycle (WLTC), the
Federal Test Procedure (FTP), the Indian Modified Driving Cycle (IMDC), US06, etc. [6,7].
In the study conducted in [8], the Urban Dynamometer Driving Cycle (UDDC) and NEDC
were used to explore the impact of temperature on EC; this study revealed optimal EC
values of 1.547 kWh and 1.648 kWh, respectively. Additionally, another study [9] delved into
the influence of the gearbox on the EC of EVs, finding that dual and continuously variable
gearbox systems conserve more energy than single-gear transmissions. However, the
energy usage of electric vehicles is primarily influenced by factors such as the dimensions
of components, operational zones of powertrain elements (motor and battery), state of
charge (SOC), driving distance, and so on. Addressing these issues requires an optimal
modeling environment to enhance vehicle performance and EC under various real-time
driving conditions. Also, ensuring the effectiveness of vehicle modeling and validation
procedures requires considering critical design and control decisions. The EV propulsion
system’s primary components include an electric motor (EM), battery, controller, and
power converters. Various types of EMs, such as a brushless direct current (BLDC) motor,
permanent magnet synchronous motor (PMSM), switched reluctance motor (SRM), and
induction motor (IM), are employed in EV propulsion systems. The choice of EM and
its associated controller significantly impacts EV performance. Therefore, to maximize
performance and energy efficiency, considering performance needs, economic concerns,
and the desired driving experience, appropriate motor and controller selection is crucial.
As well, the practice of longitudinal vehicle platooning plays a crucial role in autonomous
intelligent transportation systems by ensuring the maintenance of an optimal longitudinal
spacing with preceding vehicles [10–13]. In this study, the SR motor has been chosen
due to its beneficial characteristics, including high starting torque, efficiency, and power
density. As a result, the creation of an effective energy management system (EMS) is a
widespread priority, especially for EVs, aiming to distribute power demand efficiently
while maintaining drivability and performance.
Presently, various energy management system (EMS) techniques aim to enhance elec-
tric vehicle (EV) performance across diverse real-time conditions; these include proportional–
integral–derivative (PID) control, direct torque control (DTC), model predictive control
(MPC), field-oriented control (FOC), fuzzy logic control (FLC), hybrid control, etc. These
controllers can be characterized by their ability to achieve specific goals, such as minimiz-
ing energy consumption, optimizing dynamic responsiveness, and enhancing drivability.
Although DTC offers superior torque control, especially for SR motors in EVs, it has
drawbacks like torque and current ripples at low speeds, making it challenging to attain
maximal vehicle performance and minimal energy consumption under real-time driving
conditions [14,15]. To achieve high performance while maintaining the battery state of
charge (SOC) around the desired value during depletion, the integration of DTC and FOC
techniques into SR motors for EV propulsion is explored. However, using both controllers
results in poor energy consumption and regenerative braking efficiency in various driving
conditions [16]. Consequently, the model predictive control (MPC) approach is employed
to minimize EV energy consumption in real-time driving scenarios, forecasting future EV
behavior and identifying optimal operating conditions. Despite its high computational
costs and the need for prior knowledge of future driving actions, MPC fails to set appro-
priate battery discharge limits in real-time driving circumstances due to the complexity
of EV systems [17,18]. In contrast, the proportional–integral–derivative (PID) approach
gains attention for implementing an optimal EMS in real-world driving scenarios due
to its reliability, short processing time, and efficient memory resource utilization. This
approach dynamically adjusts control signals to enhance EV efficiency and driving range
under various conditions. Optimization of PID variables using particle swarm optimiza-
Vehicles 2024, 6 511

tion (PSO) and the genetic algorithm (GA) has shown improved transient response under
real-time driving circumstances [19]. An optimally adjusted PID controller aids in seamless
transitions between diverse driving conditions, optimizing the recovery of energy during
regenerative braking and improving the efficiency of the battery. Unlike traditional vector
control approaches like DTC and FOC, the PID controller showcases its ability to reduce
energy consumption (EC) and expand the operational range of electric vehicles (EVs) in
dynamic situations. The tuning of PID parameters and energy preservation in unpre-
dictable conditions is accomplished in [20] through the utilization of battery state-of-charge
(SOC) feedback and vehicle velocity. Despite its advantages in precise tracking of desired
speed and torque for smooth acceleration and deceleration, PID control may struggle with
nonlinearities and uncertainties in EV systems, impacting battery state-of-charge (SOC)
estimation accuracy. Due to these reasons, the tuning of PID parameters is difficult in
various transient conditions. Hence, designers have turned to intelligent controllers such
as the fuzzy logic controller (FLC) and neural network (NN) to enhance EV performance
under unpredictable conditions [21,22]. According to the literature, in order to regulate
the nonlinear operations of electric vehicles (EVs), fuzzy logic control (FLC) utilizes a
rule base and membership functions that incorporate input and output variables under
different conditions. The precise calibration of the membership functions and rule base
enhances accuracy in responses under diverse dynamic conditions. As a result, a fuzzy
logic controller (FLC) proves more effective than a PID controller across various attributes
like energy consumption (EC), state of charge (SOC), regenerative efficiency, etc., under
changing speed and load conditions [23]. Nevertheless, the nonlinear behavior of battery
usage introduces multiple uncertainties into the battery SOC feedback system. Therefore,
adjusting FLC settings to enhance electric vehicle (EV) performance becomes crucial, a
topic that researchers have yet to fully address.
To tackle uncertainties in real-time battery state of charge (SOC) while driving, an
innovative hybrid learning approach is proposed [24]. This method combines fuzzy logic
control (FLC) and proportional–integral–derivative (PID) techniques to govern electric
vehicle (EV) transient error responses in various scenarios. The FLC’s role is to adapt
PID parameters based on suitable rules and membership functions. This fusion of FLC
and PID controllers is considered a ground-breaking hybrid strategy for reducing energy
consumption (EC) and enhancing battery performance in diverse transient conditions. It
offers the advantages of both accuracy and robustness, potentially improving battery health
and lifespan. Furthermore, the hybrid approach is employed to enhance the performance
of electric vehicles (EVs) across diverse road conditions (urban, rural, and highway). This
is achieved by incorporating feedback from battery state of charge (SOC), vehicle speed,
driving behavior, and the direction of current flow (regeneration) [25]. While various hybrid
control techniques have been applied in research studies to manage unknown parameters
in nonlinear systems operating under limited conditions, their data-driven nature makes
it challenging to establish suitable mathematical models for real-time driving situations.
Neural networks (NNs) represent a type of control mechanism capable of adapting and
learning through adjustments in neuron weights, sizes, relationships between layers, and
activation functions [26]. However, the precision of neural network control relies on both
the quality and quantity of training data, with longer training times and potential issues
outside the training domain. Recognizing the limitations of relying solely on a single
learning approach, the integration of multiple control strategies (FLC and NN) through
hybridization holds promise for achieving greater efficiency in real-time conditions [27]. In
this context, the adaptive supervisory self-learning controller (ASSC) emerges as a hybrid
learning control approach, combining FLC’s reasoning mechanism with NN’s self-learning
capability. NN improves the output decisions of the fuzzy inference system by defining
optimal membership functions based on training data, aiming to enhance EV performance
and minimize EC in real-time scenarios. The ASSC approach quickly reduces variations
in speed, torque, and unfavorable chattering consequences under different conditions.
Implementing a supervisory control technique improves the dynamic behavior of the
Vehicles 2024, 6 512

EV concerning battery SOC, EC, battery C-rate, regenerative efficiency, etc. [28]. Hence,
the ASSC proves to be a more efficient control strategy than FLC and NN for managing
battery state-of-charge (SOC) fluctuations in real-time scenarios. Implementing a nonlinear
ASSC strategy holds the potential to improve the utilization profile of the battery SOC in
electric vehicles (EVs), thereby extending their operational range. The ASSC approach is
implemented with an SR motor used in EV applications under real-time conditions. In
this study, the complete process of design and performance evaluation is carried out using
the model-in-the-loop (MIL) simulation. MIL simulation is typically more efficient and
computationally straightforward as it involves defining system behavior using a mapped
experimental response technique, rather than relying on numerical representations to
illustrate the behavior of the SR motor, controller, and battery systems. A model-based
calibration technique is used for the development of an efficient SR motor and controller
maps. This procedure encompasses several stages, including the design of experiments
(DOE), model fitting, optimization, and lookup table generation. In the generation of maps
(lookup tables), the application of experimental design enables a systematic exploration
of the effects of the SR motor and controller behavior under diverse conditions, while
minimizing the number of necessary test cases. This DOE-based approach systematically
examines the impact of the SR motor and controller behavior under various conditions,
reducing the complexity, time, and expense associated with map development. As the EV
field evolves, advancements in control strategies will continue to be pivotal in achieving
greater energy efficiency, prolonged battery life, and enhanced driving experiences.
Nowadays, automakers all agree that an effective energy management system (EMS)
is necessary for EVs to operate optimally under real-time driving conditions. This study
emphasizes the significance of a sophisticated EMS in enhancing the energy efficiency
of EVs under practical scenarios. The numerical simulation of EV performance using
various EMS techniques in real-time driving situations has been the subject of numerous
studies. Moreover, precisely determining the driving range through an assessment of
electric vehicles’ EC is crucial in eliminating driver anxiety. However, identifying the
optimal discharge path for EVs is challenging due to factors such as driving actions, road
grade, travel distance, and initial SOC variability. This study aims to accomplish the
goal of efficient energy management and battery utilization in EVs by introducing an
advanced self-learning control strategy. The objective of this is to bridge the gap between
theoretical advancements and practical implementation by evaluating the effectiveness
of the optimal EMS in an SR motor-equipped EV under real-world driving conditions.
Despite numerous studies on EV performance with different control methods, there is
a lack of comprehensive publications examining the performance characteristics of EVs
with various energy management approaches. In this context, the objective of the present
study aims to develop a mathematical model of an EV by integrating a mapped SR motor
and controller efficiency with different EMS techniques. This research paper introduces
energy management strategies such as proportional–integral–derivative (PID) control,
fuzzy logic control (FLC), hybrid control, and adaptive supervisory self-learning controllers
(ASSCs) to minimize energy consumption and extend EV range under diverse operating
conditions. The ASSC approach is found to enhance the dynamic behavior of EVs compared
to traditional approaches like PID, FLC, and hybrid, considering parameters such as battery
SOC, energy consumption, battery C-rate, and regenerative efficiency. A novel aspect of
this work is the development of a real-time driving cycle in urban, rural, and highway
conditions to assess the effectiveness of EVs in practical operating scenarios. Additionally,
this research paper employs an innovative methodology to create DCs and efficiency
maps of the controller and SR motor under real-time conditions. These maps are then
incorporated into the EV model to evaluate various performance parameters, including
energy consumption, regeneration efficiency, motor power, battery SOC, battery current,
and C-rate, using different energy management controllers. This study concludes by
comparing the performance of various energy management controllers (PID, fuzzy, hybrid,
and ASSC) across the beforementioned parameters to validate the real-time performance
Vehicles 2024, 6 513

superiority of the proposed adaptive supervisory self-learning controller. It is anticipated

that the ASSC approach will significantly improve battery utilization, ultimately enhancing
the real-time operational performance of electric vehicles. The specified objectives of this
study aim to provide readers with knowledge and a critical perspective on the design and
development of effective energy management controllers for electric vehicles while also
suggesting potential areas for future research.

2. Proposed Methodology
The methodology involves the usage of various energy management controllers with
the integration of efficiency maps and a real-time driving cycle for different operating
conditions to evaluate the performance of the EVs. In this study, the entire design and
performance evaluation process is conducted using the model-in-the-loop (MIL) simulation.
The workflow and proposed methodology are organized into four aspects, as illustrated in
Figure 1. Using the parameters of Ather 450 plus the vehicle as a reference, the EV model is
created with an SR motor using MATLAB/Simulink. Table 1 contains detailed specifications
of the developed EV model. This study builds an electric vehicle configuration using a
backward-looking modeling approach. It starts with sub-models for longitudinal blocks,
transmission blocks, battery blocks, mapped motor blocks, and controller blocks. A real-
time driving cycle and a variety of efficiency maps of energy management controllers (PID,
FLZ, hybrid, and ASSC) are installed in the EV model in order to validate the performance
of the EVs. Second, to guarantee the battery utilization path and extend the driving range
of electric vehicles, this research work develops several energy management controllers,
such as PID, FLZ, hybrid, and ASSC, under real-time operating conditions. Using a
point-by-point model-based calibration technique, the motor and controller behavioral
maps (lookup tables) are produced experimentally using various energy management
controllers in transient scenarios. In order to assess the effectiveness of different energy
management controllers under real-time operating conditions, a real-time DC covering
a range of road situations, including urban, rural, and highway applications, is built. In
order to confirm the performance of EC, regeneration efficiency, motor power, battery
SOC, battery current, C-rate, etc., the created efficiency maps and DC are also loaded into
the EV simulation model. On the other hand, depending on the initial battery SOC and
trip distance, different battery usage characteristics apply. The battery utilization path is
constant throughout real-time operation when the various energy management controllers
are utilized to minimize sudden energy distribution, increasing the vehicle’s operating
range. In order to ensure the superior real-time performance of energy management, a
variety of performance characteristics are finally compared between the various energy
management controllers. It is expected that the suggested ASSC approach will result in
better battery utilization, which will eventually improve the operational performance of
electric vehicles in real time.

Table 1. Technical specifications of the developed electric vehicle.

SNO Parameters Features Value Unit

Model Electric Scooter -
Rolling resistance 0.015 -
Mass 111 kg
1 Vehicle
Garde Angle 0 degree
Area 0.875 m2
Velocity Driving Cycle kmph
Type Belt Drive -
Gear ratio 7.8:1 -
2 Transmission
Transmission Efficiency 85 %
Vehicles 2024,
Vehicles 2024, 66, FOR PEER REVIEW 6
514

Table 1. Cont. Gear ratio 7.8:1 -

SNO Transmission Efficiency
Parameters Features 85 Value % Unit
Type SR Motor -
3 Motor Type SR Motor -
3 Motor Driving Cycle Time 33.45 km
Driving Cycle Time 33.45 km
Type Lithium-ion -
BatteryType
Capacity Lithium-ion
2400 Wh-
Battery Capacity 2400 Wh
BatteryBattery
Voltage Voltage 51.1
51.1 VV
4 4 Battery
Battery
BatteryBattery
Initial Initial
SOC SOC 100
100 %%
Cell Voltage
Cell Voltage 3.6
3.6 VV
Cell Capacity 2.7 Ah
Cell Capacity 2.7 Ah

Figure 1.
Figure 1. Proposed methodology of
Proposed methodology of the
the present
present research.
research.

3. Development of Two-Wheeler Electric Vehicle Vehicle Model

Model toto Validate
Validate the
the Performance
Performance of
the Proposed Control System
of the Proposed Control System
The
The development
developmentand andanalysis of of
analysis a backward-facing
a backward-facing Matlab–Simulink
Matlab–SimulinkEV model with
EV model
an SR motor arrangement are presented in this article. The schematic representation
with an SR motor arrangement are presented in this article. The schematic representation of the
EV configuration is illustrated in Figure 2. The Simulink model consists of five
of the EV configuration is illustrated in Figure 2. The Simulink model consists of five es- essential
functional blocks, each
sential functional derived
blocks, each from the physical
derived from thecomponents incorporated
physical components into the electric
incorporated into
vehicle (EV) system. These blocks include the longitudinal block, transmission block,
the electric vehicle (EV) system. These blocks include the longitudinal block, transmission
battery block, mapped motor, and controller block. The architecture of the developed
block, battery block, mapped motor, and controller block. The architecture of the devel-
Simulink function block, designed for simulating the EV model, is illustrated in Figure 2.
oped Simulink function block, designed for simulating the EV model, is illustrated in Fig-
In this section, the significance and mathematical relationships governing the vehicle
ure 2. In this section, the significance and mathematical relationships governing the vehi-
dynamics, battery, motor, and controller are analyzed to enhance the precision of energy
cle dynamics, battery, motor, and controller are analyzed to enhance the precision of en-
consumption (EC) and driving range estimations.
ergy consumption (EC) and driving range estimations.
Vehicles 2024, 66, FOR PEER REVIEW
Vehicles 2024, 7
515

Figure
Figure 2.
2. Schematic
Schematic representation of EV
representation of EV simulation
simulation model
model and
and EV
EV powertrain
powertrain configuration.
configuration.

3.1. Longitudinal Vehicle Dynamics Model

Analysis of the dynamics and the impact of different parameters on the performance
of the vehicle is the first step in the modeling process. The forces acting on the vehicle are
determined to to compute
computethese theseparameters.
parameters.The Themodel
modelisischaracterized
characterized bybya single
a singledegree
degree of
freedom (DOF) focusing solely on longitudinal motion, with lateral
of freedom (DOF) focusing solely on longitudinal motion, with lateral and vertical mo- and vertical motions
beingbeing
tions disregarded [29]. In[29].
disregarded the Infundamental powertrain
the fundamental modeling,
powertrain the evaluation
modeling, of electric
the evaluation of
vehicle (EV) performance
electric vehicle (EV) performance centerscenters
on its longitudinal
on its longitudinaldynamics. Longitudinal
dynamics. resistive
Longitudinal re-
forces, forces,
sistive including aerodynamic
including force, acceleration
aerodynamic force, rolling
force, acceleration force, and
force, rolling gradient
force, forces,
and gradient
are taken
forces, areinto
takenaccount [30]. The
into account driving
[30]. cycle plays
The driving cyclea plays
pivotal role in the
a pivotal rolelongitudinal vehicle
in the longitudinal
dynamics
vehicle block asblock
dynamics it provides a criticalainput.
as it provides criticalThis involves
input. a comprehensive
This involves assessment
a comprehensive as-
of various resistive forces exerted on the vehicle during its longitudinal
sessment of various resistive forces exerted on the vehicle during its longitudinal motion. motion. These
encompass
These rolling rolling
encompass forces (Equation (1)), arising
forces (Equation (1)), from
arisingthefrom
interaction between between
the interaction the tires and
the
tires and the road surface; aerodynamic forces (Equation (2)), influenced by theshape
the road surface; aerodynamic forces (Equation (2)), influenced by the vehicle’s and
vehicle’s
speed;and
shape gradient
speed; forces (Equation
gradient forces(3)), stemming
(Equation (3)),from inclines
stemming or declines
from inclinesin orthe terrain;
declines inand
the
acceleration forces (Equation (4)), reflective of changes in velocity. After
terrain; and acceleration forces (Equation (4)), reflective of changes in velocity. After that, that, the estimation
of the
the total tractive
estimation of theforce is illustrated
total tractive forcein Equation (5). Finally,
is illustrated the output
in Equation of the longitudinal
(5). Finally, the output
vehicle dynamic block is wheel speed and torque, as specified
of the longitudinal vehicle dynamic block is wheel speed and torque, as specified in Equations (6) andin (7).
Equa-
tions (6) and (7).
Fr = Cr f ∗ GVW ∗ Cos (θ ) (1)
𝐹𝑟 = 𝐶𝑟𝑓 ∗ 𝐺𝑉𝑊 ∗ 𝐶𝑜𝑠 2(𝜃) (1)
Fa = 0.5 ∗ ρ ∗ A f ∗ Cd ∗ V (2)
𝐹𝑎 =Fg0.5 ∗ 𝜌(θ∗) ∗𝐴𝑓
= Sin ∗ 𝐶𝑑 ∗ 𝑉 2
GVW (2)
(3)
Facc = a ∗ GV M (4)
𝐹𝑔 = 𝑆𝑖𝑛(𝜃) ∗ 𝐺𝑉𝑊 (3)
Ft = Fr + Fa + Fg + Fa (5)
𝐹𝑎𝑐𝑐 = 𝑎 ∗ 𝐺𝑉𝑀
Wt = Ft ∗ Rw
(4)
(6)
V ∗ 60
𝐹𝑡 =Ws𝐹𝑟=+ 𝐹𝑎 + 𝐹𝑔 + 𝐹𝑎 (5)
(7)
2 ∗ π ∗ Rw
where: GVW—grass vehicle weight, GVM—grass vehicle mass, Crf—coefficient of rolling
𝑊𝑡 = 𝐹𝑡 ∗ 𝑅𝑤 (6)
resistance, Fr—rolling force, Fa—aerodynamic force, Fg—gradient force, Facc—acceleration
Vehicles 2024, 6 516

force, Ft—total tractive force, Cd—drag coefficient, Af —frontal area, a—acceleration, ρ—

density, Wt—wheel torque, Ws—wheel speed, Rw—wheel radius, and V—velocity (rpm).

3.2. Transmission Model

To cater to diverse EV tractive requirements, an effective power electronics-controlled
SR motor is employed in EVs as a substitute for multi-speed transmissions, opting for
gearless or single-speed gear transmission [31]. In the context of this study, the transmission
model is structured around a single-speed transmission with a gear ratio of 7.8:1. It utilizes
inputs from the longitudinal model, namely, wheel speed and torque, to estimate the motor
speed and torque, which are then transmitted to the SR motor model. The outputs of the
transmission model are articulated in Equations (8) and (9).

Wt
Mt = (8)
GR ∗ Te f f

Ms = Ws ∗ GR (9)
where: GR—gear ratio, Mt—motor speed, Ms—motor speed (rpm), and Teff—transmission
efficiency.

3.3. Mapped SR Motor and Controller Model

The SR motor and controller’s operating range determines an EV’s acceleration, maxi-
mum speed, passing ability, and gradeability, among other performance metrics. However,
integrating a mathematically modeled SR motor and controller into an EV simulation
introduces an additional computational challenge and encapsulates the intricate dynamic
features of the transient SR motor [32]. As a result, utilizing the steady-state empirical
model of the SR motor and controller behavior, optimal response behavior maps are created.
In the part that follows, the testing and map-building process will be thoroughly explained.
In consideration of these factors, the developed SR motor and controller efficiency maps,
incorporating different energy management systems (EMSs) such as PID, FLC, hybrid, and
ASSC, are employed in the EV simulation model. Furthermore, based on inputs from the
transmission model (Mt and Ms), the mapped SR motor model estimates motor mechanical
power (MMP) (Equation (10)) and motor electrical power (MEP) (Equation (11)) under
various real-time operating conditions. MEP is determined by integrating the efficiency
maps of different energy management controllers for the SR motor. Additionally, the motor
model calculates motor regenerative power (MRP) (Equation (12)) by incorporating the
regenerative efficiency maps of different controllers under diverse operating conditions.
Subsequently, to assess the performance of the mapped controller in real-time conditions,
the outputs of the motor block (MEP and MRP) are transmitted to the controller block.

2 ∗ π ∗ Ms ∗ Mt
MMP = (10)
60
MEP = MMP ÷ Me f f (11)
MRP = MMP ∗ Me f f (12)
Here, Meff represents motor efficiency, S denotes motor speed, and T represents motor
torque. The operations of the SR motor and battery are regulated by the controller. It
determines any modifications in the vehicle’s operation related to energy demand based
on signals received from the motor model block. The controller block integrates different
motor controller and regenerative controller efficiency maps to estimate battery motoring
power (BMP) and battery regeneration power (BRP) using inputs (MEP and MRP) from the
motor block. Equations (13) and (14) express the functions. Ultimately, the battery block
receives the outputs of the mapped controller block (BMP and BRP) in order to evaluate
the energy consumption (EC) and regenerative efficiency of various control algorithms
Vehicles 2024, 6 517

under the current operating conditions. Comprehensive details on the suggested control
algorithms are provided in Section 4.

BMP = MEP ÷ MCe f f = f (S, T > 0) (13)

BRP = MRP ∗ MCe f f = f (S, T < 0) (14)

where: MCeff—motor controller efficiency, BMP—battery motoring power, and
BRP—battery regenerative power.

3.4. Battery Model

The battery, being a complex and nonlinear system, presents challenges in modeling
due to its dependence on SOC, temperature, aging, and internal resistance. Understanding
the variations in and limitations of parameters related to battery performance is crucial
under real-time driving conditions [33]. However, to simplify the battery model, this
work does not delve into the study of the influence of battery aging and temperature. The
lithium-ion battery is a common energy source in electric vehicles (EVs) due to its distinctive
characteristics, including high voltage potential, high energy density, and being lightweight,
with minimal self-discharge. Utilizing inputs (BMP and BRP) from the controller block, the
battery model forecasts E/km, battery current, driving range, SOC, C-rate, and regeneration
efficiency with different energy management controllers (PID, FLC, hybrid, and ASSC)
in real-time driving cycles. The magnitude and direction of battery current in EVs vary
based on the accelerator throttle and brake position. Additionally, battery SOC serves as a
direct indicator of the total available energy in the battery during trips, a crucial factor in
evaluating the remaining driving range of EVs. Furthermore, the overall driving range of
EVs is directly associated with the energy consumed and recovered during acceleration
and braking. This dependence is primarily influenced by atmospheric conditions, road
segment characteristics, vehicle physical parameters, speed, and acceleration. Lastly, the
functional equations describing E/km and SOC are provided in Equations (15) and (16).

E
Z  
= BMP > 0 ÷ distance ∗ 3.6 ∗ 106 (15)
km
Z 
SOC = ISOC − Bc ÷ 3600 ∗ Bcap (16)

where: BMP—battery motoring power, ISOC—initial SOC, Bc—battery current, and Bcap—
battery capacity.

4. Design and Development of Efficient Control Systems to Evaluate the Performance of

Two-Wheeler Electric Vehicles
An effective energy management controller increases EV range and efficiency while
decreasing EC under a variety of operating circumstances. This study looks into several
energy management controllers, such as PID, fuzzy, hybrid, and supervisory, to improve
EV performance in real-time driving scenarios. Under a range of dynamic conditions,
these controllers can accurately forecast the battery utilization route while minimizing EC
and enhancing regeneration efficiency. In order to develop efficiency maps for the motor
and controller, these controllers are built and tested (steady-state experimentation) in a
variety of dynamic situations. Various energy management controllers are used to develop
efficiency maps for the motor and controller in real time. The experimental technique and
development of the control system for the SR motor are explained in the following parts.

4.1. PID Controller

The most popular controller for managing EV SR motor speed and industrial processes
is the PID controller. In order to achieve the intended output with little to no volatility
in the values despite the various disturbances, it uses a closed-loop system to continu-
 026D ɭ \textrethookbelow{l}
\textrtaill
026E ɮ \textlyoghlig
\textOlyoghlig
026F ɯ \textturnm
Vehicles 2024, 6 0270 ɰ \textturnmrleg 518
0271 ɱ \m{m}
\textltailm
0272 ɲ \m{j}
ously examine the output and modify the input values. The PID controller eliminates
\textltailn
the inaccurate reaction resulting from the difference in \textnhookleft
reference and feedback speeds. To
0273
estimate the final controller output, the erroneousɳ response
\m{n}
is multiplied by the controller’s
\textrtailn
proportional (Kp), integral (Ki), and derivative
0274 ɴ (Kd) sections,
\textscn and then summed [34]. The
duty cycle of the PWM pulses required for
0275 the SR
ɵ motor to operate in a continuous loop is
\textbaro
controlled by this output. Equation (17)0276represents
ɶ the PID controller’s continuous output
\textscoelig
control signal (u(t)). Nonetheless, the0277
EV SR motor’s
ɷ efficiency is impacted by the chosen
\textcloseomega
PID parameters. Under real-time operating
0278 conditions,
ɸ it results in a drop in regenerative
\textphi
efficiency and an improvement in the EC. Therefore, \textniphi
calculating PID settings is crucial
to an EV SR motor’s smooth operation. 0279In theɹproposed\textturnr
work, using the Ziegler Nichols
027A
approach, the PID parameters are selected. When theɺ \textturnlonglegr
SR motor dynamics are unknown
027B
or not readily available, the P, I, PI, and ɻ \textrethookbelow{r}
PID controllers are fine-tuned using the Ziegler
\textturnrrtail
Nichols approach [35]. Based on the 027Cdifferent transient
ɼ responses
\textlonglegr
of the EV SR motor, the
PID controller gains are estimated using027Dthe Ziegler
ɽ Nichols rules. In a closed-loop system,
\textrethookbelow{r}
the SR motor is coupled to the proportional controller.\textrtailr
The Ki and Kd gains are first zeroed
in this manner, and the Kp value is 027E ɾ
subsequently \textfishhookr
increased from zero to the maximum
value until the system displays steady 027F ɿ
oscillations. Tcr\textlhti
(Tcr—critical period of oscillations)
\textlhtlongi
represents the period of oscillations, whereas Kcr (Kcr—critical value of Kp) represents the
0280 ʀ \textscr
greatest value of Kp. The Ziegler Nichols tuning methodology is illustrated in Figure 3,
0281 ʁ \textinvscr
and the PID tuning parameters are established by optimizing the Kcr and Tcr values.
0282 ʂ \textrethookbelow{s}
\textrtails
Vehicles 2024, 6, FOR PEER REVIEW d 11
t) + Ki ∗ ʃ e(t)dt\m{s}
U (t) = K p ∗ e(0283 + Kd ∗ e(t) (17)
\esh dt
\textesh
0284 ʄ \texthtbardotlessj
\texthtObardotlessj
\texthtbardotlessjvar
0285 ʅ \m{S}
\textvibyi
0286 ʆ \textctesh
0287 ʇ \textturnt
0288 ʈ \M{t}
\textrtailt
\texttretroflexhook

14

Figure3.3.Tuning
Figure Tuningmethodology
methodologyofofPID
PIDcontroller
controllergains
gainsusing
usingZiegler
ZieglerNichols
Nicholsapproach.
approach.

4.2. Intelligent
Additionally,Controller
during steady-state experimentation, the mathematical model of the EV
SR motor with a PID controller
For complicated linear andisnonlinear
built using EVMATLAB/Simulink. To assess
systems, the intelligent the different
controller is widely
output
accepted as an acceptable controller. It produces effective and satisfactory results bytoman-
characteristics and create efficiency maps, the EV SR motor is run according the
best experimental design. EV SR motor and PID controller efficiency maps in
aging a wide range of input and output factors. Ref. [36] discusses the intelligent control- various dy-
namic situations are developed as a result of the findings. Finally, the created
ler’s design and development process in depth. Additionally, this work generates intelli-PID controller
efficiency maps efficiency
gent controller have beenmapsintegrated into an EV
under real-time simulation
driving modelbytocombining
conditions examine various
the pro-
performance factorscontroller
posed intelligent such as motor
with power, EC, battery
an experiment SOC, and
involving regeneration electric
a steady-state efficiency. The
vehicle
results of the simulation show that the complex nature of the EV system prevents
SR motor. The EV SR motor and controller efficiency maps are created for a range of real- the PID
controller from achieving
time operating situationsthe desired
using levels of EC,
the intelligent SOC approach.
control drop, and regenerative efficiency.
In order to estimate the
EC and regenerative efficiency with a real-time driving cycle, the obtained efficiency maps
are then fed into an EV simulation model. The intricacy of real-time electric vehicle oper-
ation is a challenge for the intelligent controller’s rule tuning. Therefore, the intelligent
controller displays a less than ideal result with the EV simulation model. However, the
results show that in comparison to the PID controller, the intelligent controller has a lower
EC, a higher regeneration efficiency, and a lower battery SOC loss. Ultimately, in real-time
driving conditions, the intelligent controller outperforms the PID controller in terms of
EV performance.
 \texthth
0267 ɧ \texthookabove{\textheng}
\texththeng
0268 ɨ \B{i}
\textbari
Vehicles 2024, 6 0269 ɩ \m{i} 519
\textiota
\textiotalatin
\textniiota
4.2. Intelligent Controller 026A ɪ \textsci
026B ɫ \textltilde
For complicated linear and nonlinear EV systems, the intelligent controller is widely
026C ɬ \textbeltl
accepted as an acceptable controller. It produces effective and satisfactory results by
026D ɭ \textrethookbelow{l}
managing a wide range of input and output factors.\textrtaill Ref. [36] discusses the intelligent
controller’s design and development 026E process inɮ depth. Additionally, this work generates
\textlyoghlig
intelligent controller efficiency maps under real-time\textOlyoghlig
driving conditions by combining
the proposed intelligent controller with ɯ
026Fan experiment \textturnm
involving a steady-state electric
0270 ɰ
vehicle SR motor. The EV SR motor and controller efficiency \textturnmrleg
maps are created for a range
0271 ɱ \m{m}
of real-time operating situations using the intelligent control approach. In order to estimate
\textltailm
the EC and regenerative efficiency with 0272a real-time
ɲ driving
\m{j} cycle, the obtained efficiency
maps are then fed into an EV simulation model. The intricacy\textltailnof real-time electric vehicle
\textnhookleft
operation is a challenge for the intelligent controller’s rule tuning. Therefore, the intelligent
0273 withɳ the EV
controller displays a less than ideal result \m{n}
simulation model. However, the
\textrtailn
results show that in comparison to the0274PID controller,
ɴ the intelligent controller has a lower
\textscn
EC, a higher regeneration efficiency, and0275a lowerɵ battery SOC
\textbaro loss. Ultimately, in real-time
driving conditions, the intelligent controller
0276 outperforms
ɶ the PID controller in terms of
\textscoelig
EV performance. 0277 ɷ \textcloseomega
0278 ɸ \textphi
4.3. Hybrid Controller \textniphi
The creation of and improvement0279 ɹ
in intelligent \textturnrin real time usually involve
controllers
027A ɺ
a difficult procedure since numerous components, including \textturnlonglegr
MFs, control rules, input
027B ɻ \textrethookbelow{r}
and output gains, and so forth, need to be changed. Furthermore, choosing the right PID
\textturnrrtail
controller parameters is crucial, and there
027C
are several
ɼ approaches
\textlonglegr
suggested for estimating
the benefits of PID controllers. Even so, the
027D PID ɽcontroller’s
\textrethookbelow{r}can be enhanced
performance
by modifying the controller gains. Consequently, in order to adjust the PID gains in
\textrtailr
response to the EV SR motor’s static and027Edynamicɾ speeds,
\textfishhookr
a self-tuning hybrid controller
is designed. It is the result of combining ɿ
027F an intelligent
\textlhti
controller with a PID one. An
\textlhtlongi
intelligent controller is used to change the PID controller’s gains in real time. Ref. [36]
0280 ʀ \textscr
discusses the hybrid controller’s in-depth design and development procedures. Therefore,
0281 ʁ \textinvscr
Equation (18) can be used to characterize the hybrid controller’s output control signal:
0282 ʂ \textrethookbelow{s}
\textrtails
ʃ d
UPID = K p2 ∗ e(t0283
)+K \m{s}
i2 ∗ e ( t ) dt + Kd2 ∗ e(t) (18)
\esh dt
\textesh
where Kp2, Ki2, and Kd2 represent the PID controller’s
0284 ʄ modified gains. Under real-time
\texthtbardotlessj
conditions, the self-learning controller is constructed \texthtObardotlessj
based on the optimal tuning of PID
\texthtbardotlessjvar
gains through fuzzy rules. Next, an EV SR motor steady-state experiment is merged with
0285 ʅ \m{S}
the developed hybrid controller in an attempt to assess\textvibyi
and generate efficiency maps in real
time. In steady-state testing, under various
0286 dynamic
ʆ conditions,
\textctesha hybrid technique is used
to develop the controller maps and motor 0287maps ʇfor the\textturnt
EV SR motor. The created efficiency
maps are then loaded into an EV simulation
0288 toʈanalyze the vehicle’s performance across
\M{t}
an actual driving cycle. The data indicate that in terms \textrtailt
of battery SOC, EC, and energy
\texttretroflexhook
recovery, the self-learning controller performs better than the PID and intelligent controllers.
It might be challenging to adjust the rules and gains of a self-learning controller in real-time
operating conditions. Thus, an adaptive supervisory self-learning controller is presented
in this work to enhance vehicle performance under different dynamic conditions with
14
minimal EC and maximum energy recovery. A thorough explanation of the supervisory
self-learning controller is given in the next section.

4.4. Adaptive Supervisory Self-Learning Controller

For EV analysis in a range of real-time circumstances, this research combines NNs
and FLC into an adaptive supervisory self-learning controller. The combination of FLC
and NNs yields a novel approach that combines the advantages of both methodologies,
leading to a notable advancement in nonlinear modeling, learning, and mapping. Because
of the latest advancements in SR motor technology, optimizing controller parameters for
 For EV analysis in a range of real-time circumstances, this research combines NNs
and FLC into an adaptive supervisory self-learning controller. The combination of FLC
and NNs yields a novel approach that combines the advantages of both methodologies,
Vehicles 2024, 6 leading to a notable advancement in nonlinear modeling, learning, and mapping. Because 520
of the latest advancements in SR motor technology, optimizing controller parameters for
electric vehicles has become a difficult undertaking. Therefore, controller engineers need
to implement
electric vehiclessome complexaself-learning
has become control algorithms
difficult undertaking. Therefore,incontroller
order to maximize
engineersthe need ef-
ficiency of SR motors in EVs. Because of its adaptability, the ASSC
to implement some complex self-learning control algorithms in order to maximize the can be used in a wide
variety ofofcontrol
efficiency applications.
SR motors in EVs. Because of its adaptability, the ASSC can be used in a wide
Figure 4 shows the internal architecture of the ASSC method. The ASSC approach’s
variety of control applications.
comprehensive
Figure 4 shows design and development
the internal architectureprocesses
of the ASSCare method.
covered The
in [36].
ASSC Also, the ad-
approach’s
vantages and disadvantages
comprehensive of the above-mentioned
design and development controllers
processes are covered are Also,
in [36]. presented in Table 2.
the advantages
Figure 5 shows a schematic illustration of the suggested energy management
and disadvantages of the above-mentioned controllers are presented in Table 2. Figure 5 controllers
in realatime.
shows The created
schematic ASSCof
illustration is the
thensuggested
incorporated intomanagement
energy a steady-statecontrollers
experiment inusing
real
time. The created ASSC is then incorporated into a steady-state experiment using an plan
an EV SR motor. The investigations are carried out in compliance with the design EV SR of
the DoE. The efficiency maps for the EV SR motor and controller are
motor. The investigations are carried out in compliance with the design plan of the DoE. then produced using
the efficiency
The ASSC approachmaps for forthevarious
EV SR dynamic
motor andcircumstances. Subsequently,
controller are then the created
produced using the ASSC effi-
ciency maps
approach are incorporated
for various into an electric
dynamic circumstances. vehicle simulation
Subsequently, model
the created to examine
efficiency maps are the
vehicle’s performance
incorporated attributes
into an electric vehiclethroughout
simulationamodel
real-time drivingthe
to examine cycle. The ASSC-assisted
vehicle’s performance
EV simulation
attributes model ayielded
throughout findings
real-time drivingthat showThe
cycle. optimal performance
ASSC-assisted in a variety model
EV simulation of real-
time driving scenarios for a number of performance metrics, including
yielded findings that show optimal performance in a variety of real-time driving scenarios motor power, bat-
terya state
for number of charge, energy recovery,
of performance metrics,battery current,
including and power,
motor battery power.
battery In real-time
state driv-
of charge,
ing situations,
energy recovery,the adaptive
battery supervisory
current, and batteryself-learning
power. Incontroller
real-time outperforms PID, FLC,
driving situations, the
and hybrid
adaptive controllers
supervisory in terms of
self-learning performance.
controller The experimental
outperforms PID, FLC, and setup
hybridandcontrollers
the proce-
dure
in termsof developing
of performance. the efficiency map for thesetup
The experimental EV SRand
motorthe and controller
procedure are described
of developing thein
efficiency mapnext
detail in the for the EV SR motor and controller are described in detail in the next section.
section.

Figure 4. Internal architecture of the proposed supervisory controller.

Table 2. Advantages and disadvantages of different proposed control approaches.

SNO Controller Type Advantages Disadvantages Ref

• Simple
• Derivative noise amplifications
1 PID Controller • Provides decent stability [6–10]
• Complex for nonlinear systems
• Easy to tune the parameters
• Lack of real-time response
• High precision
• Instability when tuning the fuzzy rules
• Rapid operation
2 Intelligent Controller and MF parameters [12–15]
• Easy to implement for
• Low speed and long run time of the
nonlinear systems
systems
• Easy to implement • No extra filtering effects
3 Hybrid Controller • Fast response • Lack of robustness [16–18]
• Low computation time • Limited operational range
• Complex to design
• High filtering effect
• Requires prior knowledge of initial
4 Supervisory Controller • High performance in real-time [20–24]
conditions
• Fast response and accurate
• Large number of datasets are required
Vehicles 2024, 6, FOR PEER REVIEW 13

Vehicles 2024, 6 521

Figure 4. Internal architecture of the proposed supervisory controller.

Figure 5.5. Schematic

Figure Schematicrepresentation
representationofof
a proposed energy
a proposed management
energy controller
management for the
controller for development
the develop-
of SR motor
ment efficiency
of SR motor map. map.
efficiency

5. Development
Table 2. Advantages of and
Controller and Motor
disadvantages Operating
of different Map
proposed under
control Various Real-Time
approaches.
Operating Conditions

SNO This Controller

section outlines theAdvantages
experimental setup of optimal SR motor and controller
Disadvantages Ref
Type
efficiency maps, which will be used to build the Simulink EV simulation model within the
MATLAB software 2018b. • Simple
• Derivative noise amplifi-
• Provides decent sta-
5.1.1Experimental Setup for Map Development cations
PID Controller bility [6–10]
• Complex for nonlinear
In this work, a 4.5 •kW SR Easy to tune
motor is usedthetopa-
create efficiency maps for real-time driving
systems
situations for a variety of energy management controllers (PID, FLC, hybrid, and ASSC).
rameters
As shown in Figure 6, the SR motor is firmly mounted • Lackin theof testbed,
real-timeandre- the 12 kW
eddy current dynamometer is securely attached to thesponse SR motor’s shaft. As the SR motor
operates at 60 V DC, the • stepdown
High precision
transformer•in Figure 6 takes 230 V AC electricity
Instability when tuning
and2converts it to 60 V•AC. Rapid
Intelligent Table 3operation
displays the SR motor’s
the fuzzytechnical
rules specifications.
and MF [12–15] The
rectifierController
receives the lower• Easy
voltage to and
implement
uses it to convert 60 V AC to 60 V DC, which
parameters
powers the SR motor. In thisfor nonlinear
instance, thesystems •
DC voltage of thespeed
Low SR motor
and is stabilized
long run under
varying load and speed conditions using a 400 V capacitor. The SR
time of the systems motor’s drive or
inverter is directly connected
• to the 60 V
Easy to implement • DC output in order to test the
No extra filtering effectsefficacy of the
different energy
Hybrid Con- •management controllers.
Fast response The above-mentioned
• controllers
Lack of robustness (PID, FLC,
3
hybrid, troller
and ASSC) are •connected [16–18]
Low to computation
the STM32 microcontroller
• Limited or DAC via an encrypted
operational
USB connection. They were time designed in real-time using MATLAB/Simulink.
range The rectifier
receives the lower voltage and uses it to convert 60 V AC to 60 V DC, which powers the
• Complex to design
SR motor. In this instance, • High
the DCfiltering
voltage effect
of the SR motor is stabilized under varying
• Requires prior
load and speed conditions • High
using performance in
a 400 V capacitor. The SR motor’s drive or inverter is
Supervisory knowledge of initial con-
4
directly connected to the 60 V real-time
DC output in order to test the efficacy of the different[20–24] energy
Controller ditions
management controllers. • The Fast response and ac-controllers (PID, FLC, hybrid, and ASSC)
above-mentioned
• Large number of datasets
are connected to the STM32curate microcontroller or DAC via an encrypted USB connection.
are required
They were designed in real-time using MATLAB/Simulink. After that, a serial port (RS 232)
connects the STM32 microcontroller to the drive/inverter so it may transmit and receive
5. Development of Controller and Motor Operating Map under Various Real-Time
data for a variety of operational scenarios. Furthermore, an SR motor can have its load
Operating Conditions
manually adjusted at various speeds thanks to the dynamometer control interface. At that
point,This section outlines
the developed energythemanagement
experimental setup of optimal
controllers, such as SRPID,motor
FLC, and controller
hybrid, effi-
and ASSC,
ciency maps,
provide controlwhich
signalswill be used
to the to build
SR motor the Simulink
drive/inverter EV simulation
in various model
real-world withinThe
scenarios. the
MATLAB software.
operating parameters of the SR motor and drive vary depending on the control signals
received from different EMCs. Ultimately, the feedback data obtained from the real-time
system are used to estimate the output responses (current and voltage) of the controller and
motor. After that, based on the experimentally predicted output responses of the motor and
 energy management controllers, such as PID, FLC, hybrid, and ASSC, provide control sig-
nals to the SR motor drive/inverter in various real-world scenarios. The operating param-
eters of the SR motor and drive vary depending on the control signals received from dif-
ferent EMCs. Ultimately, the feedback data obtained from the real-time system are used
Vehicles 2024, 6 to estimate the output responses (current and voltage) of the controller and motor. After 522
that, based on the experimentally predicted output responses of the motor and controller
in various dynamic scenarios, the efficiency maps of the controller/drive and motor are
controller
developedin variousadynamic
utilizing variety ofscenarios, the efficiency
energy management maps of the
controllers, controller/drive
including and
fuzzy, hybrid,
motor are developed
supervisory, and PIDutilizing a variety
techniques. of energy
The next management
section controllers,
explores the including
specific process fuzzy,
of devel-
hybrid,
oping a supervisory,
map. and PID techniques. The next section explores the specific process of
developing a map.

Figure 6. Experimental setup for SR motor and controller efficiency map development with different
strategies.

Table 3. Technical specifications of SR motor used in EVs.

SNO Parameter Value Units

1 Rated DC voltage 90 V
2 Rated speed 10,000 RPM
3 Rated power 4500 W
4 Rated current 90 A
5 Motor phases 3 -
6 Stator phase resistance 2.875 Ohm
7 Stator phase inductance 0.0085 H
8 Rotor moment of inertia 0.08 Kgm2
9 Friction coefficient 0.045 Nms
10 Back EMF coefficient 1.3 V/rad/s

5.2. Optimal Energy Efficient Response Maps under Various Dynamic Operating Conditions
In this work, motor and controller efficiency maps with different EMCs are developed
for the EV simulation model using the model-based calibration (MBC) technique. The
step-by-step method of the MBC technique comprises the creation of maps, modeling,
optimization, and DoE. This study uses a sophisticated method called the MBC method-
ology in MATLAB to examine how different variables affect future results. This study
generates motor and controller efficiency maps for PID, intelligent, hybrid, and supervisory
controllers under real-time operating conditions using a one-stage model technique. The
operating and control settings of the motor and controller have a significant impact on
their behavior. The motor’s operational parameters are torque and speed, and its control
parameters are energy management controllers, which are control algorithms like PID, FLC,
and so on. It is necessary to tune the control parameters (control algorithms) under a variety
of dynamic conditions in order to obtain the best SR motor and controller response maps.
 Vehicles 2024, 6 523

To develop the model, the DoE technique is used to capture experimental motor response
characteristics. Table 4 displays the operational parameter range and experimental control.
Using the I-optimal technique, the design plan is developed in cooperation with the DoE
(50 test situations). The Sobol-series DoE is used to collect the motor and controller data
for different EMCs based on the test conditions, and the tests are conducted in compliance
with the design plan. The collected data are transformed through processing, such that it
approximates a normal distribution, improving the effectiveness of the model’s predicting
capability. In this study, the empirical model behavior of the motor and controller are
generated under dynamic conditions using a Gaussian elimination technique. With the
empirical models of the motor and controller, the optimal efficiency maps for different
EMCs of the SR motor and controller are developed under real-time operating conditions.
The created SR motor and controller efficiency maps, which are used in vehicle modeling,
are shown in Figure 7a–d using various energy management strategies, including PID,
FLC, hybrid, and ASSC, under real-time conditions. Lastly, an EV simulation model is
Vehicles 2024, 6, FOR PEER REVIEW loaded with the produced controller and motor efficiency maps to analyze the vehicle’s16
performance characteristics.

(a) (b)

(c) (d)

(a)

Figure 7. Cont.
Vehicles 2024,
Vehicles 6 6, FOR PEER REVIEW
2024, 17
524

(b)

(c)

Figure 7. Cont.
Vehicles 2024,
Vehicles 2024, 66, FOR PEER REVIEW 18
525

(d)
Figure 7. (a) SR motor and controller efficiency maps for PID controller under real-time conditions.
Figure 7. (a) SR motor and controller efficiency maps for PID controller under real-time conditions.
(b) SR motor and controller efficiency maps for intelligent controller under real-time conditions. (c)
(b) SR motor
SR motor andand controller
controller efficiency
efficiency mapsmaps for intelligent
for hybrid controller
controller under real-time
under real-time conditions.
conditions. (d) SR
(c) SR motor and controller efficiency maps for hybrid controller under real-time conditions.
motor and controller efficiency maps for adaptive supervisory self-learning controller under real-
(d)
timeSRconditions.
motor and controller efficiency maps for adaptive supervisory self-learning controller under
real-time conditions.
6. Simulation and Validation of the Developed SR Motor and Controller Maps with
Table 4. Range of the
the Real-World operating
Driving parameters under different dynamic conditions.
Cycle
The driving
Parameter cycle source is necessary MotortoSpeed
simulate the EC and battery discharge be-
Motor Load
havior. In this study,
Units
a real-time DC is created
rpm
for all road situations, including
N-m
urban,
rural, and highway. Figure 8 shows the experimental process for a generating real-time
Factor type Numeric Numeric
DC in diverse driving situations, including urban, rural, and highway. This study exam-
Range DC design and development
ines the real-time 250–9500 process, as well as route 3–52
selection, trip
timing, and experimental technique. The first and most crucial step in creating a driving
cycle is deciding
6. Simulation andonValidation
a route [37]. TheDeveloped
of the driving routeSR in Vellore,
Motor andIndia, is determined
Controller Maps with based
the
Real-World
on knowledge Driving Cycle
of local road and traffic conditions, as seen in Figure 9. Table 5 discusses
the developed
The drivingdriving cycle ischaracterizations
cycle source for different
necessary to simulate driving
the EC and routes.
battery The behavior.
discharge selected
driving route acomprises
In this study, real-timeall DCthree types of
is created forroad conditions:
all road rural,
situations, highway,
including and urban.
urban, The
rural, and
whole
highway. length of 8the
Figure driving
shows the trip is aroundprocess
experimental 33.45 kmfor with differentreal-time
a generating real-timeDCroad condi-
in diverse
driving situations, including urban, rural, and highway. This study examines the real-
tions.
time DC design and development process, as well as route selection, trip timing, and
Table 5. Developed
experimental driving cycle
technique. Thecharacteristics.
first and most crucial step in creating a driving cycle is
deciding on a route [37]. The driving route in Vellore, India, is determined based on
Parameter Units Description
knowledge of local road and traffic conditions, as seen in Figure 9. Table 5 discusses the
Total Distance
developed km
driving cycle characterizations 33.45for different driving routes. The selected driving
Urban Driving Distance km 4.86
Vehicles 2024, 6, FOR PEER REVIEW 19

Rural Driving Distance km 14.56

Vehicles 2024, 6
Rural Driving Distance
Highway Driving Dis- km 14.56
526
Highway Driving Dis- km 14.03
tance km 14.03
tance
No. of Stops - 3 (urban, rural, and highway)
No. ofcomprises
route
Maximum Stops
Speed all three types
Limit km/h -of road
3 (urban,
80 rural,rural,
conditions: and highway)
highway, and urban. The whole
Maximum Speed Limit km/h 80
length of the driving trip is around 33.45 km with different real-time road conditions.

Figure 8. Methodology for development of real-time driving cycle under different road conditions.
Figure 8. Methodology for development of real-time driving cycle under different
different road
road conditions.
conditions.

Selecteddriving
Figure9.9.Selected
Figure drivingroute
routefor
fordevelopment
developmentofofreal-time
real-timedriving
drivingcycle
cyclefor
forurban,
urban,rural,
rural,and
and
Figure 9.conditions.
highway
highway Selected driving route for development of real-time driving cycle for urban, rural, and
conditions.
highway conditions.
Furthermore, the electric two-wheeler is used to develop a real-time driving cycle for a
variety of road situations. Table 1 shows the selected vehicle specifications. The chosen EV
is connected to a microcontroller and mobile phone GPS. The microcontroller contains the
vehicle’s speed as well as battery and motor performance statistics under different real-time
driving conditions. In addition, the data collected from the mobile GPS include vehicle
speed and location with X, Y, and Z directions along the different driving routes such as
Vehicles 2024, 6 527

urban, rural and highway. During driving, the microcontroller stores the real-time data for
different real-time driving conditions. Following that, a real-time DC is created using the
Vehicles 2024, 6, FOR PEER REVIEW collected data, with different driving routes including urban, rural, and highway. Figure 20 10
shows the profile of the created real-time driving cycle for different road conditions. A
driving cycle is a set of data points that show the relationship between a vehicle’s speed
and Furthermore,
time. It is usedthetoelectric
evaluate the performance
two-wheeler is usedoftocars in several
develop aspects
a real-time including
driving cyclefuel
for
aefficiency,
variety ofelectric vehicle range,
road situations. Table 1and harmful
shows emissions.
the selected Furthermore,
vehicle the The
specifications. established
chosen
real-time driving cycle is used with the EV simulation model to estimate performance
EV is connected to a microcontroller and mobile phone GPS. The microcontroller contains
metrics such as power, C-rate, EC, battery discharge behavior, regeneration efficiency, and
the vehicle’s speed as well as battery and motor performance statistics under different
so on. Finally, this study integrates a real-time DC, a BLDC motor, and various energy
real-time driving conditions. In addition, the data collected from the mobile GPS include
management controller (such as PID, fuzzy, hybrid, and supervisory) efficiency maps with
vehicle speed and location with X, Y, and Z directions along the different driving routes
an EV simulation model to analyze the motor and battery’s performance in real-world
such as urban, rural and highway. During driving, the microcontroller stores the real-time
driving situations.
data for different real-time driving conditions. Following that, a real-time DC is created
using the collected data, with different driving routes including urban, rural, and high-
Table 5. Developed driving cycle characteristics.
way. Figure 10 shows the profile of the created real-time driving cycle for different road
conditions.Parameter
A driving cycle is a set of Unitsdata points that show the Description
relationship between a
vehicle’s speed
Total Distance and time. It is used to evaluate
km the performance
33.45 of cars in several aspects
including fuel efficiency, electric vehicle range, and harmful emissions. Furthermore, the
Urban Driving Distance km 4.86
established real-time driving cycle is used with the EV simulation model to estimate per-
Rural Driving
formance metricsDistance
such as power, C-rate,km 14.56
EC, battery discharge behavior, regeneration ef-
ficiency,
Highway and so on.Distance
Driving Finally, this study integrates
km a real-time
14.03 DC, a BLDC motor, and vari-
ous energy
No. of Stopsmanagement controller (such
- as PID, fuzzy, hybrid,
3 (urban, rural, and supervisory) effi-
and highway)
ciency maps with an EV simulation model to analyze the motor and battery’s performance
Maximum Speed Limit km/h 80
in real-world driving situations.

Figure10.
Figure Real-timedriving
10.Real-time drivingcycle
cycleprofile
profilefor
fordifferent
different road
road conditions.
conditions.

7. Results and Discussion

7. Results and Discussion
The objective of this study is to validate the effectiveness of an EV by integrating a real-
The objective of this study is to validate the effectiveness of an EV by integrating a
time DC and various EMC efficiency maps into a simulation model. This section comprises
real-time DC and various EMC efficiency maps into a simulation model. This section com-
an analysis and comparison of various EMCs with respect to the EV’s power output, battery
prises an analysis and comparison of various EMCs with respect to the EV’s power output,
current, C-rate, EC, and regeneration efficiency, all of which are evaluated under real-time
battery current, C-rate, EC, and regeneration efficiency, all of which are evaluated under
operating conditions. It is also necessary to analyze the parameters associated with battery
real-time operating conditions. It is also necessary to analyze the parameters associated
with battery and motor performance to comprehend the fluctuations in and limitations of
EVs under real-time operating conditions.

7.1. Motor Power

The intended speed and torque of EVs under different driving conditions have an
 Vehicles 2024, 6 528

and motor performance to comprehend the fluctuations in and limitations of EVs under
Vehicles 2024, 6, FOR PEER REVIEW 21
real-time operating conditions.

7.1. Motor Power

in thisThestudyintended speed and
using various torque
energy of EVs under
management different driving
controllers. The variousconditions
EMCshave suchan as
impact on the motor power. Real-time analysis of motor power fluctuations
intelligent, ASSC, PID, and hybrid controllers were used under different road conditions, is performed
in this
and study using
the resulting various
motor energy are
fluctuations management
illustratedcontrollers.
in Figure 11.The various to
According EMCs such as
the findings,
intelligent, ASSC, PID, and hybrid controllers were used under different
the average motor power of the PID, intelligent, hybrid, and supervisory controllers is 5.8, road conditions,
and3.1,
4.3, theandresulting
2.3 kW motor fluctuations
in various aresituations.
driving illustrated The
in Figure 11.show
results According to the
that the lackfindings,
of real-
the average motor power of the PID, intelligent, hybrid, and supervisory
time parameter modification in conventional controllers (PID, intelligent, and hybrid) controllers is 5.8,
4.3, 3.1, and 2.3 kW in various driving situations. The results show
causes the maximal motor power fluctuations compared to the proposed controller under that the lack of real-time
parameter modification in conventional controllers (PID, intelligent, and hybrid) causes the
a variety of road conditions. Thus, the proposed supervisory controller will reduce EC
maximal motor power fluctuations compared to the proposed controller under a variety of
and increase the driving range of EVs under various driving conditions. Furthermore, the
road conditions. Thus, the proposed supervisory controller will reduce EC and increase
maximum motor power of the PID, intelligent, hybrid, and supervisory controllers is 12.2,
the driving range of EVs under various driving conditions. Furthermore, the maximum
11.1, 9.2, and 7.8 kW, respectively. The suggested supervisory controller shows less fluc-
motor power of the PID, intelligent, hybrid, and supervisory controllers is 12.2, 11.1, 9.2,
tuations in maximal motor power than the standard controllers, such as PID, intelligent,
and 7.8 kW, respectively. The suggested supervisory controller shows less fluctuations in
and hybrid
maximal motor controllers.
powerBy introducing
than the standardlarge variationssuch
controllers, in motor
as PID, power, a traditional
intelligent, con-
and hybrid
troller reduces driving range and increases battery discharge rate and energy
controllers. By introducing large variations in motor power, a traditional controller reduces consump-
tion under
driving different
range driving battery
and increases conditions. To overcome
discharge rate and this problem,
energy the proposed
consumption control-
under different
ler uses its self-learning capabilities to effectively mitigate the real-time
driving conditions. To overcome this problem, the proposed controller uses its self-learning power variation
ofcapabilities
the motor. to In effectively
the end, inmitigate
real-timethedriving situations,
real-time power the suggested
variation supervisory
of the motor. In con-the
troller yields excellent performance results in terms of discharge
end, in real-time driving situations, the suggested supervisory controller yields rate, EC, and regenera-
excellent
tion efficiency.results in terms of discharge rate, EC, and regeneration efficiency.
performance

Figure
Figure11.
11.Variations
Variationsininmotor
motorpower
powerwith
with different
different energy management controllers
controllers under
under varied
varied
driving
drivingconditions.
conditions.

7.2. Battery Power

The maximal velocity of an EV during performance is affected by the electricity pro-
duced by its battery at a specific instant. Battery power varies according to the temperature
and discharge rate under various dynamic conditions. The various EMCs such as intelli-
gent, ASSC, PID, and hybrid were used under different road conditions such as urban,
 Vehicles 2024, 6 529

7.2. Battery Power

Vehicles 2024, 6, FOR PEER REVIEW 22
The maximal velocity of an EV during performance is affected by the electricity pro-
duced by its battery at a specific instant. Battery power varies according to the temperature
and discharge rate under various dynamic conditions. The various EMCs such as intel-
rural, and
ligent, highway,
ASSC, and hybrid
PID, and the resulting battery
were used power
under fluctuations
different are illustrated
road conditions such as in urban,
Figure
12. Theand
rural, resultant graphs
highway, and show that thebattery
the resulting proposedpower controller,
fluctuationsin comparison
are illustratedto in
other tradi-
Figure 12.
tional controllers
The resultant (PID,
graphs showintelligent, and hybrid),
that the proposed showsinfewer
controller, variations
comparison under
to other battery
traditional
controllers
power (PID,rural,
in urban, intelligent, and hybrid),
and highway drivingshows fewer In
conditions. variations
differentunder battery power
road circumstances,
in average
the urban, rural,
batteryand highway
power of thedriving
differentconditions.
EMCs is 9.3, In different road2.7circumstances,
6.4, 4.6, and kW, respectively.the
average
The battery power
conventional of thehave
controllers different
highEMCs
batteryis 9.3,
power 6.4, variations,
4.6, and 2.7hence
kW, respectively.
the battery Thedis-
conventional
charge controllers
characteristics willhave high battery
be affected under power
real-time variations, hence the battery
driving conditions. discharge
This will lead to
decreases in the operational range of EVs due to the increased energy consumption. to
characteristics will be affected under real-time driving conditions. This will lead de-
Then,
creases in the operational range of EVs due to the increased energy consumption.
in a variety of driving circumstances, the supervisory controller exhibits the minimal bat- Then, in
a variety of driving circumstances, the supervisory controller exhibits
tery power variation and thus will aid in reducing EC and increasing driving range. Also, the minimal battery
power
the variation
maximum and thus
battery power will
of aid
the in reducing
PID, EC and
intelligent, increasing
hybrid, driving range.
and supervisory Also, the
controllers is
maximum battery power of the PID, intelligent, hybrid, and supervisory
19.6, 16.3, 10.7, and 9.1 kW, respectively. The conventional controllers (PID, intelligent, controllers is
19.6, 16.3, 10.7, and 9.1 kW, respectively. The conventional controllers (PID, intelligent,
and hybrid) generate the greatest battery power fluctuations in response to varying road
and hybrid) generate the greatest battery power fluctuations in response to varying road
conditions, resulting in an increase in EC and battery depletion rate. However, the super-
conditions, resulting in an increase in EC and battery depletion rate. However, the supervi-
visory controller shows minimal battery power variations compared to the other tradi-
sory controller shows minimal battery power variations compared to the other traditional
tional controllers under different real-time driving conditions. As a result, the proposed
controllers under different real-time driving conditions. As a result, the proposed controller
controller will help lower battery power deviation, improving EVs’ EC and driving range
will help lower battery power deviation, improving EVs’ EC and driving range in a range
inofa real-world
range of real-world driving scenarios.
driving scenarios.

Figure
Figure12.
12.Variations
Variationsininbattery
batterypower
powerwith
withdifferent
different energy
energy management
management controllers under varied
controllers under varied
driving conditions.
driving conditions.

7.3. Battery Current

In electric vehicles, the accelerator and brake pedal positions affect the magnitude
and direction of battery current. Figure 13 depicts the variations in battery current caused
by a variety of energy management controllers and road conditions. The graph shows that
the average battery current of PID, intelligent, hybrid, and supervisory controllers is 196,
162, 113, and 74 A, respectively. Due to the high nonlinear behavior of the vehicle, the
 Vehicles 2024, 6 530

7.3. Battery Current

In electric vehicles, the accelerator and brake pedal positions affect the magnitude and
Vehicles 2024, 6, FOR PEER REVIEW direction of battery current. Figure 13 depicts the variations in battery current caused by 23
a variety of energy management controllers and road conditions. The graph shows that
the average battery current of PID, intelligent, hybrid, and supervisory controllers is 196,
162, 113, and 74 A, respectively. Due to the high nonlinear behavior of the vehicle, the
traditional controllers (PID, intelligent, and hybrid) show higher battery current than the
traditional controllers (PID, intelligent, and hybrid) show higher battery current than the
recommended controller under real-time driving conditions. Also, due to the absence of
recommended controller under real-time driving conditions. Also, due to the absence of
real-time parameter tunings across various travelling conditions, conventional controllers
real-time parameter tunings across various travelling conditions, conventional controllers
are incapable of achieving the highest possible vehicle performance. As a result, the pro-
are incapable of achieving the highest possible vehicle performance. As a result, the
posed supervisory
proposed controller
supervisory effectively
controller reduces
effectively reduces battery
batterycurrent
currentfluctuations
fluctuationsunder
underreal-
real-
time operating conditions by utilizing its adaptive self-learning capabilities.
time operating conditions by utilizing its adaptive self-learning capabilities. Moreover, the Moreover,
the maximum
maximum battery
battery current
current of various
of various EMCsEMCs is 383.1,
is 383.1, 319.6,
319.6, 232.5,
232.5, and and
176.8176.8 A, respec-
A, respectively.
tively. The absence of real-time parameter correction in conventional
The absence of real-time parameter correction in conventional controllers will result controllers will re-
in
sult in a decline
a decline in theinelectric
the electric vehicle’s
vehicle’s performance,
performance, rapidrapid depletion
depletion of theofbattery
the battery
SOC,SOC,
and
and
an an increase
increase in EC
in EC under
under different
different conditions.
conditions. That That being
being said,
said, the the proposed
proposed supervi-
supervisory
sory controller shows very little fluctuation in battery current under rural,
controller shows very little fluctuation in battery current under rural, urban, and highway urban, and
highway road conditions.
road conditions. Therefore,
Therefore, it is usedit to
is used
lowertoEC lower
and EC and operating
extend extend operating
range, asrange,
well as
well as to improve
to improve batterybattery discharge
discharge rate inrate in different
different driving driving situations.
situations. In comparison
In comparison to
to other
other traditional
traditional controllers,
controllers, the proposed
the proposed controller
controller improves
improves battery
battery power,power, current,
current, and
and SOC,
EC, EC,
SOC, driving range,
driving and regenerative
range, and regenerativeefficiency in real-time
efficiency driving
in real-time scenarios.
driving scenarios.

Figure 13.13.Variations
Figure Variationsininbattery
batterycurrent
currentwith
withdifferent
differentenergy
energymanagement
management controllers
controllers under
under varied
varied
driving conditions.
driving conditions.

7.4. Battery C-Rate

The battery discharge rate has an impact on the C-rate in a variety of real-world driv-
ing conditions. The energy management controllers such as intelligent, ASSC, PID, and
hybrid were used under different conditions such as urban, rural, and highway, and the
resulting variations in C-rate are illustrated in Figure 14. From the figure, the maximum
 Vehicles 2024, 6 531

7.4. Battery C-Rate

The battery discharge rate has an impact on the C-rate in a variety of real-world
Vehicles 2024, 6, FOR PEER REVIEW 24
driving conditions. The energy management controllers such as intelligent, ASSC, PID, and
hybrid were used under different conditions such as urban, rural, and highway, and the
resulting variations in C-rate are illustrated in Figure 14. From the figure, the maximum
controllers
and average exhibit greater
C-rates C-rate EMCs
of various variations under4.8,
are 5.36, different real-time
3.25, and road
2.48, and conditions.
2.7, So,
1.6, 0.8, and
0.4, the
with respectively.
traditionalIncontrollers,
comparisonthe to battery
the proposed supervisory
discharge controller,
characteristics the traditional
will degrade under
controllers
real-time exhibit
driving greater C-rate
conditions. variations
In urban, rural, under different
and highway real-time road
conditions, conditions.
a higher So,
C-rate also
with thebattery
shortens traditional controllers,
life and decreasesthebattery
batterydischarge
dischargeefficiency.
characteristics will degrade
Furthermore, under
an increase
inreal-time driving
C-rate will conditions.
result in elevated Inbattery
urban, rural, and highway
temperatures conditions,
and greater a higher
energy C-rate dur-
dissipation also
shortens battery life and decreases battery discharge efficiency. Furthermore,
ing mobility at varying velocities. Therefore, it is not feasible to attain the optimal perfor-an increase in
C-rate will result in elevated battery temperatures and greater energy
mance of EVs under various real-time driving conditions using conventional controllers. dissipation during
mobility at
However, varying
the velocities. Therefore,
implementation it is not supervisory
of the proposed feasible to attain the optimal
controller performance
will contribute to
enhanced performance through the reduction in C-rate variations across variousHowever,
of EVs under various real-time driving conditions using conventional controllers. real-time
the implementation
driving conditions. The of proposed
the proposed supervisory
controller controller
has the ability towill contribute
adjust to enhanced
the different control
performance through the reduction in C-rate variations across various real-time driving
parameters instantly because of its adaptive self-learning capabilities. Consequently, it
conditions. The proposed controller has the ability to adjust the different control parameters
can effortlessly manage nonlinear behavior systems of EVs in a variety of dynamic scenar-
instantly because of its adaptive self-learning capabilities. Consequently, it can effortlessly
ios. Therefore, in this situation, under real-time operating conditions, the supervisory con-
manage nonlinear behavior systems of EVs in a variety of dynamic scenarios. Therefore, in
troller will improve battery life and discharge rate. Additionally, it will reduce the energy
this situation, under real-time operating conditions, the supervisory controller will improve
consumption of EVs and increase their operating range in highway, rural, and urban sit-
battery life and discharge rate. Additionally, it will reduce the energy consumption of EVs
uations. Ultimately, it is clearrange
and increase their operating that inin various
highway, driving circumstances,
rural, and the supervisory
urban situations. Ultimately, con-
it is
troller performs better than other EMCs in terms of C-rate, power, EC,
clear that in various driving circumstances, the supervisory controller performs better than performance,
range,
other recovery energy,
EMCs in terms ofetc.
C-rate, power, EC, performance, range, recovery energy, etc.

Figure
Figure14.
14.Variations
VariationsininC-rate
C-ratewith
withdifferent
differentenergy
energy management
management controllers under varied
varied driving
driving
conditions.
conditions.

7.5. State of Charge

One important metric for assessing an EV’s remaining driving range is battery SOC,
which serves as a direct measure of the entire amount of energy in the battery during a
trip. Figure 15a illustrates the variations in battery state of charge (SOC) corresponding to
distinct EMCs across a range of real-time driving conditions. The PID controller end-of-
Vehicles 2024, 6 532

7.5. State of Charge

One important metric for assessing an EV’s remaining driving range is battery SOC,
which serves as a direct measure of the entire amount of energy in the battery during a trip.
Figure 15a illustrates the variations in battery state of charge (SOC) corresponding to distinct
EMCs across a range of real-time driving conditions. The PID controller end-of-trip (EOT)
SOC is 87.2, 47.1, and 3.6% under different road (urban, rural, and highway) conditions. The
EOT values for intelligent and hybrid controllers are 89.3, 56.7, and 21.5, and 92.2, 69.3, and
44.9%, respectively. The proposed controller’s EOT SOC in urban, rural, and highway driving
scenarios is 94.1, 77.1, and 59.1%, respectively. The findings show that the EOT SOC for
different energy management controllers under urban driving conditions is 87.2, 89.3, 92.2,
and 94.1%. When compared to the other traditional controllers, the supervisory controller
exhibits the lowest SOC drop (95.3%) in urban driving conditions. Owing to the very nonlinear
real-time behavior of EVs, conventional controllers are unable to maintain the correct SOC
level in a variety of road conditions. Thus, due to the self-learning capabilities, the suggested
controller is suitable for reducing the variations in battery SOC corresponding to distinct
EMCs and for extending driving range under real-time driving conditions. Additionally, the
EOT SOC values are 47.1, 56.7, 69.3, and 77.1% for rural driving situations with different
energy management controllers. The supervisory controller records a lower SOC drop than
the other conventional controllers in rural driving circumstances because of its real-time
parameter-tuning capabilities. The battery discharge rate will increase due to the higher
SOC drop of the PID, intelligent, and hybrid controllers, which will also increase the EC and
reduce the vehicle’s operating range. Furthermore, Figure 15b displays the highway EOT
SOC of the different energy management controllers. The final SOC levels of the supervisory,
intelligent, hybrid, and PID controllers are shown in the figure as 3.6, 21.5, 44.9, and 59.1%,
respectively. The supervisory self-learning controller outperforms the traditional controllers
(PID, intelligent, and hybrid) in terms of trip SOC reductions at the EOT under different
real-time conditions. Thus, the proposed supervisory controller is used to improve the driving
experience and range under different real-time driving conditions.

7.6. Energy Consumption and Regenerative Efficiency

The amount of energy used and recovered during acceleration and braking, which is
mostly dependent on the driving conditions and features of road segments, vehicle physical
parameters, speed, and acceleration, is directly connected with the entire driving range of
electric vehicles. The regeneration efficiency and total energy consumption per kilometer for
different EMCs are shown in real time in Figure 16. In various driving situations, the overall
energy consumption of PID, intelligent, hybrid, and supervisory controllers is 97.21, 85.63,
60.14, and 44.67 Wh/km, respectively. Because of its self-tuning capabilities, the supervisory
controller can tune and optimize the control settings in real-time situations, resulting in a low
EC (44.67 Wh/km) compared to the other conventional controllers. The EC of an EV decreases
as a result of the nonlinear behavior of EVs under different conditions, which makes it difficult
for traditional controllers to modify and optimize control settings in real time. As a result,
under various dynamic road conditions, the driving range of an electric vehicle will increase
with the minimal EC (44.67 Wh/km) of the suggested controller. Furthermore, Figure 16
illustrates that the corresponding regenerative efficiencies of the PID, intelligent, hybrid, and
supervisory controllers are −21.81, −27.73, −41.64, and −58.28 Wh, respectively. According to
the data, compared to the other conventional controllers, the supervisory controller (−58–28 Wh)
recovers the highest amount of regenerative energy. In addition, the hybrid controller recovers a
substantially higher amount of energy than both the intelligent and PID controllers. However,
under different real-time driving conditions, the proposed controller recovers more regenerative
energy than the standard controllers. The maximal energy recovery rate will definitely increase
the EV’s operating range under different conditions. Lastly, an efficient energy management
controller enhances the vehicle’s power, efficiency, SOC, energy consumption, regenerative
efficiency, etc., according to an analysis of a variety of performance characteristics pertaining to
various EMCs (PID, intelligent, hybrid, and ASSC) under varied real-time driving conditions.
Vehicles
Vehicles 6
2024, 2024, 6, FOR PEER REVIEW 26 533

(a)

(b)

Figure 15. (a) Individual SOC variations of various energy management controllers under urban,
rural, and highway driving conditions. (b) Final SOC variations with different energy management
controllers.
Vehicles 2024, 6, FOR PEER REVIEW 27

Vehicles 2024, 6 performance characteristics pertaining to various EMCs (PID, intelligent, hybrid, and
534
ASSC) under varied real-time driving conditions.

Figure16.
Figure Different energy
16. Different energy management
managementcontrollers’
controllers’energy consumption
energy andand
consumption regenerative efficiency.
regenerative effi-
ciency.
8. Conclusions
This study has been conducted to optimize the energy consumption of EVs by devel-
8. Conclusions
oping a robust energy management controller and testing it under various real-time driving
This study has been conducted to optimize the energy consumption of EVs by devel-
conditions. This study employs various methodologies to develop the efficiency maps, the
oping a robust energy management controller and testing it under various real-time driv-
EV model, and a real-time DC which are then combined with the developed EMCs, such as
ing conditions. This study employs various methodologies to develop the efficiency maps,
PID, intelligent, hybrid, and supervisory, to enhance the EV performance under real-time
the EV model, and a real-time DC which are then combined with the developed EMCs,
driving conditions. A MATLAB/Simulink-based mathematical model is formulated for
such as PID, intelligent, hybrid, and supervisory, to enhance the EV performance under
an EV with an SR motor. Through a novel experimental approach, the efficiency maps
real-time driving conditions. A MATLAB/Simulink-based mathematical model is formu-
of the motor and various EMCs are developed. These maps are then integrated into a
lated for an EV with an SR motor. Through a novel experimental approach, the efficiency
model-in-loop (MIL) EV test platform to assess the performance of different EMCs in EVs
maps
under ofreal-time
the motordrivingand various EMCs are
conditions. developed.
Further, for theThese maps are
validation then integrated
process, a real-timeintoDC
aismodel-in-loop (MIL) EV test platform to assess the performance
implemented for different types of road conditions, which include urban, rural, and of different EMCs in
EVs under roads.
highway real-time driving
This DC isconditions.
linked to the Further, for theEV
MIL-based validation process,
test platform a real-time
real-time DC
analysis
isofimplemented for different types of road conditions, which
energy consumption and battery discharge patterns. The present study concludes byinclude urban, rural, and
highway
simulating roads.
the This DC is linked
EV model using to the MIL-based
various EV test maps
EMC efficiency platformandreal-time analysis
the real-time DCofto
energy
analyze parameters like motor power, battery power, C-rate, energy consumption sim-
consumption and battery discharge patterns. The present study concludes by (EC),
ulating
state ofthe EV model
charge (SOC),using various EMC
regenerative efficiency
efficiency, maps The
and more. and validation
the real-timeandDC to analyze
interpretation
parameters
outcomes of like
thismotor power,
research paper battery power, C-rate,
are summarized below.energy consumption (EC), state of
charge (SOC), regenerative efficiency, and more. The validation and interpretation out-
• To
comes ofcarry out the EV
this research simulation,
paper the necessary
are summarized below.efficiency maps of the SR motor and the
controllers are developed under real-time conditions. For the validation of the model,
• To carry outdriving
a real-time the EVcycle
simulation, the necessary
is designed to encompass efficiency
diverse maps
roadof the SR motor
conditions, and
including
the controllers are developed
urban, rural, and highway scenarios. under real-time conditions. For the validation of the
model, a real-time driving cycle is designed to encompass diverse
• The supervisory controller performs better than conventional controllers as it exhibits road conditions,
including urban, based
less variations, rural, and
on thehighway
SR motorscenarios.
and battery performance characteristics in
• The supervisory controller performs
controllers in urban, rural, and highway driving better than conventional
conditions, controllers
as presented as in
it exhibits
Table 6.
• The proposed supervisory self-learning controller achieves significantly lessinenergy
less variations, based on the SR motor and battery performance characteristics con-
trollers in urban, rural, and highway driving conditions,
consumption (44.67 Wh/km) compared to PID (97.21 Wh/km), intelligent as presented in Table 6.
• The
(85.63proposed
Wh/km), supervisory
and hybrid self-learning
(60.14 Wh/km) controller achieves
controllers undersignificantly less energy
varying real-time oper-
consumption
ating conditions. (44.67 Wh/km)
This suggests compared to PID (97.21
an improvement Wh/km),
in battery intelligent
utilization behavior(85.63
and
Wh/km),
operatingand hybrid
range (60.14
for the Wh/km)
EV under controllers
dynamic under varying real-time operating
conditions.
Vehicles 2024, 6 535

Table 6. Output parameters of EV with different energy management controllers under real-time conditions with SR motor.

Maximum Average
SNO Parameters PID Controller Intelligent Controller Hybrid Controller Supervisory Controller
Urban Rural Highway Urban Rural Highway Urban Rural Highway Urban Rural Highway PID FLC Hybrid ASSC
1 Motor Power (kW) 7.9 12.2 8.6 7.3 11.1 7.6 6.4 9.2 6.2 5.6 7.8 5.2 5.8 4.3 3.1 2.3
2 Battery Power (kW) 12.8 19.6 13.7 10.1 16.3 11.2 8.1 10.7 7.8 6.3 9.1 5.8 9.3 6.4 4.6 2.7
3 Battery Current (A) 249.6 383.1 267.4 213.2 319.6 218.7 160.4 232.5 154.5 125.1 176.8 115.1 196 162 113 74
4 C-rate 3.49 5.36 3.74 2.9 4.8 3.1 2.25 3.25 2.16 1.75 2.48 1.61 2.7 1.6 0.8 0.4
5 SOC (%) 87.2 47.1 3.6 89.3 56.7 21.5 92.2 69.3 44.9 94.1 77.1 59.1 - - - -
6 E/Km (Wh) 97.21 85.63 60.14 44.67 - - - -
Regenerative
7 −21.81 −27.73 −41.64 −58.28 - - - -
Efficiency (%)
Vehicles 2024, 6 536

• The end-of-trip state-of-charge (SOC) drop for the proposed supervisory controller
(59.1%) is notably lower than that of the PID (3.6%), intelligent (21.5%), and hybrid (44.9%)
controllers across different road conditions. Consequently, the supervisory controller
enhances battery efficiency and overall EV performance in real-time conditions.
• The regenerative efficiency of the PID, intelligent, hybrid, and supervisory controllers
is recorded as −21.81, −27.73, −41.64, and −58.28 Wh under varying road condi-
tions. Notably, the proposed supervisory controller exhibits greater energy recovery
compared to the other conventional controllers. This improvement contributes to
enhanced battery consumption behavior and extended driving range under real-time
conditions.
The current study improves the performance of electric vehicles (EVs) under real-time
driving conditions through the implementation of the proposed adaptive supervisory
self-learning controller. It not only reduces energy consumption (EC) but also enhances
the driving range across various road conditions. Although the potential limitation of the
suggested controller is its heavy reliance on training information, impacting its overall
performance, it can be overcome by leveraging advanced controllers like the multi-adaptive
neuro fuzzy inference system (MANFIS). Furthermore, specialized optimization techniques
could be employed to train the data. Investigating these alternatives could contribute to
achieving a more stable and effective performance for the proposed controller.

Author Contributions: Methodology, B.A. and D.U.; investigation, D.U.; resources, B.A.; data
curation, B.A.; writing—original draft preparation, P.S.; writing—review and editing, P.S.; supervision,
B.A.; project administration, B.A. All authors have read and agreed to the published version of
the manuscript.
Funding: This study is supported and carried out as part of the funding from the European Union un-
der the Erasmus scheme of Jean Monnet Modules (Grant No: JMO-2023-HEITCH-RSCH—101127804),
and as part of the funding from the “British Council” for Going Global Partnerships Industry
Academia Grant 2023-24 (Application Unique ID-28).
Data Availability Statement: Data are contained within the article.
Acknowledgments: The authors of this research work wish to acknowledge the Management of the
Vellore Institute of Technology, Vellore, for the resources and amenities provided whilst conducting
the research.
Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations

ASSC Adaptive Supervisory Self-Learning Controller

BLDC Motor Brushless Direct Current Motor
BMP Battery Motoring Power
BRP Battery Regenerative Power
DC Driving Cycle
DOE Design of Experiments
DOF Degree of Freedom
EC Energy Consumption
EMC Energy Management Controller
EMs Electric Motors
EMS Energy Management System
EOT End of Trip
EV Electric Vehicle
FLC Fuzzy Logic Controller
GHG Green House Gas
GPS Global Positioning System
GVM Grass Vehicle Mass
GVW Grass Vehicle Weight
HEV Hybrid Electric Vehicle
Vehicles 2024, 6 537

HIL Hardware In Loop

IM Induction Motor
MANFIS Multi-Adaptive Neuro Fuzzy Inference System
MBC Model-Based Calibration
MEP Motor Electric Power
MIL Model In Loop
MMP Motor Mechanical Power
MRP Motor Regenerative Power
Ms Motor Speed
Mt Motor Torque
NN Neural Network
OBD On-Board Diagnosis
PID Proportional–Integral–Derivative
SOC State of Charge
SR Motor Switched Reluctance Motor
Ws Wheel Speed
Wt Wheel Torque

References
1. Singirikonda, S.; Pedda, O.Y. Investigation on performance evaluation of electric vehicle batteries under different drive cycles. J.
Energy Storage 2023, 63, 106966. [CrossRef]
2. Itani, K.; De Bernardinis, A.; Khatir, Z.; Jammal, A. Comparison between two braking control methods integrating energy recovery
for a two-wheel front driven electric vehicle. Energy Convers. Manag. 2016, 122, 330–343. [CrossRef]
3. Li, Z.; Khajepour, A.; Song, J. A comprehensive review of the key technologies for pure electric vehicles. Energy 2019, 182, 824–839.
[CrossRef]
4. Vignesh, R.; Ashok, B. Intelligent energy management through neuro-fuzzy based adaptive ECMS approach for an optimal
battery utilization in plugin parallel hybrid electric vehicle. Energy Convers. Manag. 2023, 280, 116792. [CrossRef]
5. Ibrahim, M.A.; Alsammak, A.N. Dynamic Modelling and Current Control for Switched Reluctance Motor Drives in Electric
Vehicles. IOP Conf. Ser. Mater. Sci. Eng. 2023, 1295, 012006. [CrossRef]
6. Miranda, M.H.; Silva, F.L.; Lourenço, M.A.; Eckert, J.J.; Silva, L.C. Electric vehicle powertrain and fuzzy controller optimization
using a planar dynamics simulation based on a real-world driving cycle. Energy 2022, 238, 121979. [CrossRef]
7. Abou-ElSoud, A.M.; Nada, A.S.; Abdel-Aziz, A.A.; Sabry, W. Torque ripples reduction of electric vehicle synchronous reluctance
motor drive using the strong action controller. Ain Shams Eng. J. 2024, 15, 102428. [CrossRef]
8. Li, Y.; Deng, H.; Xu, X.; Wang, W. Modelling and testing of in-wheel motor drive intelligent electric vehicles based on co-simulation
with Carsim/Simulink. IET Intell. Transp. Syst. 2019, 13, 115–123. [CrossRef]
9. Saiteja, P.; Ashok, B.; Wagh, A.S.; Farrag, M.E. Critical review on optimal regenerative braking control system architecture,
calibration parameters and development challenges for EVs. Int. J. Energy Res. 2022, 46, 20146–20179. [CrossRef]
10. Hu, Z.; Su, R.; Wang, Y.; Wang, B.; Huang, L.; Lu, Y. Security Enhancement for Longitudinal Vehicle Platooning Under Denial-of-
Service Attacks: From Resilient Controller Design Perspective. IFAC-Pap. Online 2023, 56, 1088–1093. [CrossRef]
11. Bageshwar, V.L.; Garrard, W.L.; Rajamani, R. Model predictive control of transitional maneuvers for adaptive cruise control
vehicles. IEEE Trans. Veh. Technol. 2004, 53, 1573–1585. [CrossRef]
12. Hu, Z.; Liu, S.; Luo, W.; Wu, L. Intrusion-detector-dependent distributed economic model predictive control for load frequency
regulation with PEVs under cyber-attacks. IEEE Trans. Circuits Syst. I Regul. Pap. 2021, 68, 3857–3868. [CrossRef]
13. Hu, Z.; Su, R.; Zhang, K.; Xu, Z.; Ma, R. Resilient Event-Triggered Model Predictive Control for Adaptive Cruise Control Under
Sensor Attacks. IEEE/CAA J. Autom. Sin. 2023, 10, 807–809. [CrossRef]
14. Diao, K.; Sun, X.; Lei, G.; Bramerdorfer, G.; Guo, Y.; Zhu, J. System-level robust design optimization of a switched reluctance
motor drive system considering multiple driving cycles. IEEE Trans. Energy Convers. 2020, 36, 348–357. [CrossRef]
15. Zhu, Y.; Wu, H.; Zhen, C. Regenerative braking control under sliding braking condition of electric vehicles with switched
reluctance motor drive system. Energy 2021, 230, 120901. [CrossRef]
16. Vignesh, R.; Ashok, B.; Kumar, M.S.; Szpica, D.; Harikrishnan, A.; Josh, H. Adaptive neuro fuzzy inference system-based energy
management controller for optimal battery charge sustaining in biofuel powered non-plugin hybrid electric vehicle. Sustain.
Energy Technol. Assess. 2023, 59, 103379. [CrossRef]
17. Sivasamy, S.; Sundaramoorthy, P.; Beno, M. A comprehensive Investigation of Outer Rotor Permanent Magnet Switched
Reluctance Motor for Enhanced Performance in Electric Vehicles. IEEE Can. J. Electr. Comput. Eng. 2023, 46, 342–347. [CrossRef]
18. Fang, G.; Bauman, J. Optimized switching angle-based torque control of switched reluctance machines for electric vehicles.
In Proceedings of the 2020 IEEE Transportation Electrification Conference & Expo (ITEC), Chicago, IL, USA, 23–26 June 2020;
pp. 186–191.
Vehicles 2024, 6 538

19. Bishop, J.D.; Doucette, R.T.; Robinson, D.; Mills, B.; McCulloch, M.D. Investigating the technical, economic and environmental
performance of electric vehicles in the real-world: A case study using electric scooters. J. Power Sources 2011, 196, 10094–10104.
[CrossRef]
20. Dhakal, R.; Parameswaran, S.; Muthukumar, R.; Moussa, H. Performance Analysis of Electrical Vehicle Battery Thermal Management
System; SAE Technical Paper; SAE International: Warrendale, PA, USA, 2022.
21. Saiteja, P.; Ashok, B. Critical review on structural architecture, energy control strategies and development process towards optimal
energy management in hybrid vehicles. Renew. Sustain. Energy Rev. 2022, 157, 112038. [CrossRef]
22. Li, W.; Xu, H.; Liu, X.; Wang, Y.; Zhu, Y.; Lin, X.; Wang, Z.; Zhang, Y. Regenerative braking control strategy for pure electric
vehicles based on fuzzy neural network. Ain Shams Eng. J. 2024, 15, 102430. [CrossRef]
23. Hamouda, M.; Abdel Menaem, A.; Rezk, H.; Ibrahim, M.N.; Számel, L. Numerical estimation of switched reluctance motor
excitation parameters based on a simplified structure average torque control strategy for electric vehicles. Mathematics 2020,
8, 1213. [CrossRef]
24. Ahmed, M.; Naiju, C.D. Modeling and Simulation for Hybrid Electric Vehicle with Parallel Hybrid Braking System for HEV; SAE
Technical Paper; SAE International: Warrendale, PA, USA, 2018.
25. Vinh, V.Q.; Ha, V.T. Improved Torque Ripple of Switched Reluctance Motors using Sliding Mode Control for Electric Vehicles.
Eng. Technol. Appl. Sci. Res. 2023, 13, 10140–10144. [CrossRef]
26. Hasanhendoei, G.R.; Afjei, E.; Naseri, M.; Azad, S. Automatic and real time phase advancing in BLDC motor by employing an
electronic governor for a desired speed-torque/angle profile. e-Prime-Adv. Electr. Eng. Electron. Energy 2023, 4, 100111. [CrossRef]
27. Ramesh, P.; Umavathi, M.; Bharatiraja, C.; Ramanathan, G.; Attika, S. Development of a PMSM motor field-oriented control
algorithm for electrical vehicles. Mater. Today Proc. 2022, 65, 176–187. [CrossRef]
28. Uddin, K.; Moore, A.D.; Barai, A.; Marco, J. The effects of high frequency current ripple on electric vehicle battery performance.
Appl. Energy 2016, 178, 142–154. [CrossRef]
29. Godfrey, A.J.; Sankaranarayanan, V. A new electric braking system with energy regeneration for a BLDC motor driven electric
vehicle. Eng. Sci. Technol. Int. J. 2018, 21, 704–713.
30. Hu, L.; Tian, Q.; Zou, C.; Huang, J.; Ye, Y.; Wu, X. A study on energy distribution strategy of electric vehicle hybrid energy storage
system considering driving style based on real urban driving data. Renew. Sustain. Energy Rev. 2022, 162, 112416. [CrossRef]
31. da Silva, S.F.; Eckert, J.J.; Corrêa, F.C.; Silva, F.L.; Silva, L.C.; Dedini, F.G. Dual HESS electric vehicle powertrain design and fuzzy
control based on multi-objective optimization to increase driving range and battery life cycle. Appl. Energy 2022, 324, 119723.
[CrossRef]
32. Al-Wreikat, Y.; Serrano, C.; Sodré, J.R. Driving behaviour and trip condition effects on the energy consumption of an electric
vehicle under real-world driving. Appl. Energy 2021, 297, 117096. [CrossRef]
33. Wager, G.; Whale, J.; Braunl, T. Driving electric vehicles at highway speeds: The effect of higher driving speeds on energy
consumption and driving range for electric vehicles in Australia. Renew. Sustain. Energy Rev. 2016, 63, 158–165. [CrossRef]
34. Ananda Padmanaban, L.; Saravanan, P. Design, analysis and comparison of switched reluctance motors for electric vehicle
application. Automatika 2023, 64, 239–247. [CrossRef]
35. Saiteja, P.; Ashok, B.; Mason, B.; Krishna, S. Development of efficient energy management strategy to mitigate speed and torque
ripples in SR motor through adaptive supervisory self-learning technique for electric vehicles. IEEE Access 2023, 11, 96460–96484.
[CrossRef]
36. Wang, G.; Makino, K.; Harmandayan, A.; Wu, X. Eco-driving behaviors of electric vehicle users: A survey study. Transp. Res. Part
D:Transp. Environ. 2020, 78, 102188. [CrossRef]
37. Donkers, A.; Yang, D.; Viktorović, M. Influence of driving style, infrastructure, weather and traffic on electric vehicle performance.
Transp. Res. Part D Transp. Environ. 2020, 88, 102569. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.