ASSIGNMENT

BIOCYBERNETICS
“Control Engineering In Biomedical Application”

Disusun guna memenuhi salah satu tugas mata kuliah Biocybernetics kelas A semester
genap tahun akademik 2020/2021.

Oleh :
Hanif Prayoga Haditama
07311840000031

Dosen Pengampu :
Dr. Achmad Arifin, S.T., M.Eng.

DEPARTEMEN TEKNIK BIOMEDIK

FAKULTAS TEKNOLOGI ELEKTRO DAN INFORMATIKA CERDAS
INSTITUT TEKNOLOGI SEPULUH NOPEMBER
2021
 1.  In development of FES control method for swing phase gait, describe
approaches of authors in the point of view Bio-cybernetics. 

Answer :
1.
a) Advance control strategies
The author mentioned a lot of control strategies that is used in a similar research
regarding FES control of swing phase gait. These includes PID Controller,
Adaptive Controller, artificial neural network, neuro-PID controller, and neuro-
Fuzzy controller. In this research, the author uses a combination of two control
system, the usage of a framework using fuzzy controller to realize a cycle-to-cycle
control rule is used in order to simulate a multi-joint movement and gait restoration.
The author claimed using these advanced control strategies proved to be easy.

b) Involvement of Artificial Intelligence

The author’s preference on using Artificial intelligence in FES control method
for swing phase gait is with the usage of a Fuzzy Control System. The usage of
artificial intelligence may be implemented on a later stage of development, but for
now, the author only mentions the involvement of Fuzzy control system,
specifically for the feedback control system of the gait swing phase.

c) Non-invasive approach and utilizing model

The model in which the author uses was never intended to be invasive, the
system uses a surface electrode (F-150, Nihon Kohden) as the driving force for the
system. Other than that, a consent from the test subject was already achieved, so
experimentation with the human subject is already clear.
The model and experiment goes by having the target maximum value of knee
extension angle as 30 degrees, with a pulse width and pulse frequency fixed at 200
us and 20Hz respectively, and having the electric goniometer as the measuring tool
for the knee joint angle.

d) Experimental with human subject after several testing with clear performance
The result after several testing gave a result that the Fuzzy controller is able to
stimulate burst duration of the muscle of each subject in all trials in hopes to restore
a gait phase. On the first trial, the maximum knee extension angle is reached on the
third cycle, and managed to maintain to be close to the target until the end of the
trial. Utilizing the knowledge on cycle-to-cycle control on the fuzzy control method
was easy.
 2. Describe how discrete PID control algorithm has been studied in realizing cycle-
to-cycle control for swing-phase FES gait.

Answer :
2.
a). Discrete control algorithm determination by author.

Figure 1 PID controller of the cycle-to-cycle control

Structure of the PID controller modified is shown in Figure 1. Transfer function

of the discrete PID controller is expressed in the z domain in Equation 1.

(1)
where G is the controller gain, and z1 and z2 are zeros of the controller. The control
algorithm of the PID controller for the cycle-to-cycle control was derived from the
transfer function.

b). How dynamics of system to be controlled was studied to find controller parameter
determination.
The PID controller has three parameters that the values of parameters have to be
determined. The parameters are controller zeros (z1 and z2), and controller gain (G). For
the single-joint and two-joint control, the parameter z1 was set by the parameter
identified from the characteristics of the joint model. Passive test of the knee joint was
performed by release the shank and the foot segment from a certain release angle in
seated position. The characteristics of the ankle joint model was obtained by releasing
the joint from a certain plantarflexion angle.

c). Method of derivation of adaptive PID control algorithm.

 Figure 3 Adaptive PID controller of the cycle-to-cycle control.

The adaptive PID controller was designed by adding an online parameter adjustment.
The structure of the adaptive PID controller is shown in Figure 3. First of all, an
objective function in Equation (1) was defined.
1
V = (θtarget −θobtained [n−1])2 (3)
2

The parameter adjustment works to modify parameter by minimizing the objective

function. The parameter adaptation of the each controller parameter are described as
followings.

(i) Adjustment of the controller gain

The gradient of the objective function respect to G was given by:
∂V ∂V ∂θ obtained ∂TB
=
∂ G θ obtained ∂ TB ∂ G

where,
δV
= θobtained −θtarget
δθobtained
δ θobtained δ θobtained [ n ]−θobtained [n−1]
δTB
=
TB [ n ] −TB [n−1]

The adjustment of the G is shown in Equation below :

δV
G[n] = G[n - 1] - ղ
δG
where ղ is the adaptation constant. In this study ´ of the adaptive PID controller was
0.001.

(ii) Adjustment of the controller zeros

 δV δV δ θobtained δTB
=¿
δ z1 δθobtained δTB δ z1
δV δV δ θobtained δTB
=¿
δ z2 δθobtained δTB δ z2
Where,

The adjustemnts of the z1 and z2 are as shown in Equation (2).

δV
𝚣1[n] = 𝚣1[n - 1] - ղ
δ z1
δV
𝚣2[n] = 𝚣2[n - 1] - ղ (4)
δ z2

d). Performance evaluation method of the comptrollers.

In the design of control systems, analysis is used to determine the limits of
system parameters in stable conditions. In this study, the PID controller parameters were
selected based on the accuracy analysis technique of the closed loop control system. A
system with characteristic equations is stable if and only if all the roots of the
characteristic equation lie on the unit circle. The equation of the loop transfer function
of the PID controller is shown in the following equation :
z−z 2
1+G =0
z −1
 3. Describe complete steps of Knowledge Engineering approach in realizing FES
cycle-to-cycle control.

Answer :
3. Knowledge engineering is a method of construction of a knowledge-base system
through investigation of a particular field, learns what concepts are important in the
field, and creates formal representations of the objects and the relation in the field. The
cycle-to-cycle control method can be realized by using such knowledge about
mechanism of the joint movements during swing phase of gait and functions of the lowe
limb muscles. Therefore, a cycle-to-cycle control framework was developed using a
knowledge engineering approach. The process approach in developing the framework is
described in the following subsections :
a) Task Definition
The main task is to establish a concrete framework to realize the cycle-to-cycle
control method. The task can be accomplished by the following sub tasks:
 To extract the knowledge about mechanisms of joint movements of the lower
limb during level gait through gait analysis of normal level gait.
 To extract the relationship of joint movements and the schedule of muscle
stimulation through assessment of functions of the muscles around the joint of
human lower limb. The functions of the lower limb muscles are in the human
anatomy literatures.
 To capture parameter values of the joint movements in level gait that will be
useful in realization of the cycle-to-cycle control.

b) Knowledge Acquisition
The knowledge acquisition process contains of two process of knowledge
acquisition : the first is knowledge acquisition through measurement and
analysis of normal gait data, and the second is functional assessment of muscle
of human lower limb. This assessment aimed to make a stimulation schedulre
for each muscle during swing phase of gait. Control of the joint movement by
FES during swing phase gait was aimed to generate the joint movements
mimicking the joint movements of the measured normal gait.

c) Knowledge Representation
The objective of the control of swing phase by using the cycle-to-cycle control
method is to generate the joint movements as in the normal gait. During the
swing phase, the joints move in a number of motion arcs. The peak of each
motion arc was assigned as the target joint angle,i.e, the maximum knee flexion,
 the maximum ankle dorsiflexion,etc. Next, it is necessary to make the ontology
of joint movements of the swing phase, which is contains the mechanism of the
joint movements during swing phase captured in the knowledge acquisition and
the vocabularies assigned to be used in the domain of the cycle-to-cycle control
are summarized. Stimulation of each muscle is aimed to induce related joint
movement reaching a target joint angle are represent in a data table.

d) Implementation and Feasibility Test

Fuzzy controller is a model-free approach and non-linear nature. It does not
require a mathematical of the system under control,, whereas most
conventional control techniques either analytical or experimental model. In this
point of view, the fuzzy controller is supposed to work in controlling system
with a latge uncertainty or unknown variation of system parameters, suach as is
controlling multi-joint movements of FES gait. Seven fuzzy controllers are
designed and placed in various desired muscles. Stimulation at the beginning
and at the middle of swing phase are controlled using control algorithm
equation as below .

Design of fuzzy FES Controller tested through computer stimulation using

musculo-skeletal model for hemiplegic gait. Several different subject models
were used in the computer simulation test to explore the capability of the
designed fuzzy controllers in compensating inter-subject variability responses.

4). Why Fuzzy controller recommended for further development?

a) Superiority of performance to other controllers.

Fuzzy controller shows several advantages compared to other conventional
Controller, such as PID ( Proportional-Integral-Derivative) Controller. First, Fuzzy
controller can be implemented based on the cycle-to-cycle control for controlling
complicated multi joint (hip, knee, and ankle) movements in the swing phase of FES-
induced gait, while the PID Controller showed usefulness and effectiveness, were only
for restoring single joint movements. Fuzzy controller showed better performance than
PID controller in its control capabilitiy, including automatic generation of stimulation
burst duration and compensation of muscle fatigue through computer simulations usinga
musculo-skeletal model. Third, the controller parameter adjustment based on the fuzzy
model was also shown to become effective when oscillating response was caused due to
inter-subject variability. Fuzzy controller is considered to be more practical to
implement the cycle-to-cycle control method.
b). Readiness in design and maintenance.
Fuzzy control technology was shown to be effective in developing nonlinear
controllers applied in numerous applications. Main advantage of the fuzzy control is the
capability of capturing and utilizing qualitative human expert knowledge in quantitative
way to calculate appropriate control action. Since the characteristics of the electrically
stimulated musculo-skeletal system is known to be non-linear, the control law was
realized using a fuzzy control scheme. Fuzzy controller is also very ready in terms of
design and maintenance, as an example the modified fuzzy controller with OAF;s
performed knee joint control apppropriately.
c). Adaptive algorithm in human subject experimental test.
The fuzzy FES controller was examined in controlling maximum knee extension
angle stimulating the vastus muscles with neurogically intact subjects. In most of
control trials, the joint angle was controlled well compensating changes in muscle
responses to electrical stimulation (muscle fatigue and force enhancement). The
experimental results show that the fuzzy controller would be practical in clinical
applications. In addition, it was considered that increased TB of electrical stimulation
pulses enhanced muscle force. For practical clinic applications, it’s desirable to design
an automatic parameter adjustment of the fuzzy controller for quickly reaching the
target and for quick compensation of changes in muscle response without causing
oscillationg response. Sequence of the muscle stimulation is arranged in a stimulation
schedule. The stimulation schedule to generate the joint movements of the swing phase
is represented in the bar chart.