ASL

Autonomous Systems Lab

Spring 2021

Locomotion Concepts
Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 1
 ASL
Autonomous Systems Lab

Autonomous mobile robot | the see-think-act cycle

knowledge, mission
data base commands
Localization “position“ Cognition
Map Building global map Path Planning

environment model
path
local map

Information Path

Motion Control
Extraction Execution
Perception

see-think-act actuator
raw data
commands

Sensing Acting

Real World
Environment
Autonomous Mobile Robots
Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 2
 ASL
Autonomous Systems Lab

Locomotion Concepts: Principles Found in Nature

 On ground

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 3
 ASL
Autonomous Systems Lab

Locomotion Concepts

 Nature evolved a multitude of locomotion concepts

 Adaptation to environmental characteristics
 Adaptation to the perceived environment (e.g. size)

 Concepts found in nature

 Difficult to imitate technically
 Do not employ wheels
 Sometimes imitate wheels (bipedal walking)
 The smaller living creatures are, the more likely they fly

 Most technical systems today use wheels or caterpillars

 Legged locomotion is still mostly a research topic

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 4
 ASL
Autonomous Systems Lab

Walking or rolling?

 number of actuators
 structural complexity
 control expense
 energy efficient
 terrain (flat ground, soft ground, climbing..)
 Cost of Transportation
 𝐶𝑜𝑇 𝐶
· · · ·
 𝐸 𝐸𝑛𝑒𝑟𝑔𝑦 𝑃 𝑃𝑜𝑤𝑒𝑟 𝑚 𝑚𝑎𝑠𝑠
 𝑑 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑣 𝑠𝑝𝑒𝑒𝑑
 movement of the involved masses
 walking / running includes up and down movement
of COG
 some extra losses
Autonomous Mobile Robots
Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 5
 ASL
Autonomous Systems Lab

Characterization of locomotion concept

 Locomotion
 physical interaction between the vehicle and its environment.
 Locomotion is concerned with interaction forces, and the mechanisms and
actuators that generate them.
 The most important issues in locomotion are:
 stability  characteristics of contact
 number of contact points  contact point or contact area
 center of gravity  angle of contact
 static/dynamic stabilization  friction
 inclination of terrain  type of environment
 structure
 medium (water, air, soft or hard
ground)
Autonomous Mobile Robots
Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 6
 ASL
Autonomous Systems Lab

Walking Robots

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 7
 ASL
Autonomous Systems Lab

Why legged robots?

 Legged systems can overcome many

obstacles, that are not reachable by
wheeled systems!

 But it is quite hard to achieve this since

 many DOFs must be controlled in a
coordinated way
 the robot must see detailed elements of the
terrain

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 8
 ASL
Autonomous Systems Lab

Number of Joints of Each Leg (DOF: degrees of freedom)

 A minimum of two DOF is required to move a leg forward
 a lift and a swing motion.
 Sliding-free motion in more than one direction not possible
 Three DOF for each leg in most cases (e.g. pictured below)
 4th DOF for the ankle joint
 might improve walking and stability
 additional joint (DOF) increases the complexity of the design and especially of the locomotion control.

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 9
 ASL
Autonomous Systems Lab

Legged Robotics | Impressive Mobility

ALoF – The Quadruped Robot I (ETH)
 Highly integrated leg design
 large range of leg motion allowing for complex maneuvers
 “static” walking

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 10
 ASL
Autonomous Systems Lab

The number of distinct event sequences (gaits)

 The gait is characterized as the distinct sequence of lift and release events of
the individual legs
 it depends on the number of legs.
 the number of possible events N for a walking machine with k legs is:

𝑁 2𝑘 1 ! factorial
 For a biped walker (k=2) the number of possible events N is:

𝑁 2𝑘 1 ! 3! 3⋅2⋅1 6

 For a robot with 6 legs (hexapod) N is already

𝑁 11 ! 39′916′800

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 11
 ASL
Autonomous Systems Lab

Most Obvious Natural Gaits with 4 Legs are Dynamic

 Changeover
Walking

 Galloping

free fly

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 12
 ASL
Autonomous Systems Lab

Most Obvious Gait with 6 Legs is Static

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 13
 ASL
Autonomous Systems Lab

Dynamic Walking vs. Static Walking

 Statically stable  Dynamic walking

CoG
CoG

 body weight supported by at least three legs  the robot will fall if not continuously moving
 even if all joints ‘freeze’ instantaneously, the  less than three legs can be in ground contact
robot will not fall  fast, efficient and demanding for actuation and
 safe, slow and inefficient control

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 14
 ASL
Autonomous Systems Lab

Dynamic Locomotion | SLIP Principles in Robotics

 Pneumatic piston
 Hydraulic leg “angle” orientation

Early Raibert hoppers (MIT Leg Lab), 1983

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 15
 ASL
Autonomous Systems Lab

Biped Walking

 Biped walking mechanism

 not too far from real rolling when moving on flat ground
 rolling of a polygon with side length equal to the length of the
step
 the smaller the step gets, the more the polygon tends to a
circle (wheel)

 But…
 rotating joint was not invented by nature
 work against gravity is required

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 16
 ASL
Autonomous Systems Lab

Modeling Legged Locomotion

Inverted Pendulum
 Walking can be represented by an inverted pendulum

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 17
 ASL
Autonomous Systems Lab

Efficiency Comparison | Cost of Transportation (CoT)

 𝐶 𝐶𝑜𝑇
· · · ·
 𝐸 𝐸𝑛𝑒𝑟𝑔𝑦 𝑃 𝑃𝑜𝑤𝑒𝑟 𝑚 𝑚𝑎𝑠𝑠 𝑑 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑣 𝑠𝑝𝑒𝑒𝑑

C J. Braun, University of Edinburgh, UK

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 18
 ASL
Autonomous Systems Lab

Static walking principles

Inverted Pendulum
 Static walking can be represented by inverted pendulum
 Exploit this in so-called passive dynamic walkers
 Add small actuation to walk on flat ground

Cornell Ranger
Total distance: 65.24 km
Total time: 30:49:02
Power: 16.0 W
CoT: 0.28

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 19
 ASL
Autonomous Systems Lab

Towards Efficient Dynamic Walking: Optimizing Gaits

C Structure and motion laboratory

University of London

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 20
 ASL
Autonomous Systems Lab

Efficient Walking and Running | serial elastic actuation

https://www.youtube.com/watch?v=6igNZiVtbxU

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 21
 ASL
Autonomous Systems Lab

Towards Efficient Dynamic Walking: Optimizing Gaits

 Nature optimizes its gaits

 Storage of “elastic” energy
 To allow locomotion at varying frequencies
and speeds, different gaits must utilize Trotting Galloping
these elements differently
Walking

3 2 1 2
1 4 1 2
 The energetically most
Crawl Bound economic gait is a function of
2 1 1 1 desired speed.
1 2 1 1
(Figure [Minetti et al. 2002])
Trot Pronk
1 1 2 4
2 2 1 3

Autonomous Mobile Robots Pace Gallop

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 22
 ASL
Autonomous Systems Lab

History of Legged Robotics

Humanoid Robots After 2000

HONDA TOYOTA

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 23
 ASL
Autonomous Systems Lab

Fukushima – where no human should go

Fukushima, 2011 DARPA Robotics Challenge, 2012

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 24
 ASL
Autonomous Systems Lab

DARPA Robotics Challenge

… and a thing we learned after the DRC Finals
 Walking is still difficult

12 x speed

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 25
 ASL
Autonomous Systems Lab

Legged Robotics | Impressive Mobility

Atlas from Boston Dynamics

https://www.youtube.com/watch?v=knoOXBLFQ-s
Autonomous Mobile Robots
Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 26
 ASL
Autonomous Systems Lab

Legged Robotics Today | Finally, a breakthrough?

 Example: Boston Dynamics

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 27
 ASL
Autonomous Systems Lab

ANYmal (ETH) learning robot locomotion in rough terrain

https://youtu.be/knIzDj1Ocoo
Autonomous Mobile Robots
Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 28
 ASL
Autonomous Systems Lab

Flying Robots

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 29
 ASL
Autonomous Systems Lab

Flying « Locomotion »

 Forces acting on a profile with relative wind speed

 Lift force Fl

 Drag force Leading edge Chord length

 Moment Angle of attack

M Trailing edge
Fd
Chord
Relative
Wind 25 % Chord Thickness

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 30
 ASL
Autonomous Systems Lab

Scaling| Structural Load Wing loading [N/m2]

 All flying objects follows

the square-cube law of scaling:

W  L  mg  b3 W: weight
S  b2 S: wing area

From Insects to Jumbo Jets, MIT Press, Cambridge

Weight [N]

Tennekes H (1996) The Simple Science of Flight,

W /S  b
1
W / S  k1  W 3

 Possible Wing loading

as function of weight

W / S  47  W 1/ 3
Tennekes H (1996) The Simple Science of Flight, From
Insects to Jumbo Jets, MIT Press, Cambridge

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 31
 ASL
Autonomous Systems Lab

How to Fly (Unmanned Aerial Vehicles) | flight concepts

 Helicopters: (video Prof. D’Andrea ETH)

 < 20 minutes
 Highly dynamic and agility
 Fixed Wing Airplanes:
 > some hours; continuous flights possible
 Non-holonomic constraints
 Blimp: lighter-than-air
 > some hours (dependent on wind conditions);
 Sensitive to wind
 Large size (dependent on payload)
 Flapping wings: (video Festo BionicOpter)

 < 20 minutes; gliding mode possible

 Non-holonomic constraints
 Very complex mechanics Festo BionicOpter

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 32
 ASL
Autonomous Systems Lab

Service Robots – flying robots for challenging tasks

wingtra – developed by students

| the VTOL UAV
https://www.youtube.com/watch?v=QADvPDWtgFU

| 81 hours non-stop in summer 2015

| 5.64 m, 6.2 kg
https://www.youtube.com/watch?v=8m4_NpTQn0E
https://www.youtube.com/watch?v=wyS6W1t_ryQ

Voliro – developed by students

| the omni-directional multi-copter
https://www.youtube.com/watch?v=9FJn_t-YCwM

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 33
 ASL
Autonomous Systems Lab

Swimming Robots

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 34
 ASL
Autonomous Systems Lab

Underwater Robots
– designed by ETH students
Naro (2009)
| the tuna robot
https://www.youtube.com/watch?v=L61O2CmZCc4

Taratuga (2012)
| the turtle robot
https://www.youtube.com/watch?v=pqy_NSHcGLs

Sepios (2014, with Disney)

| the Kalmar robot
https://www.youtube.com/watch?v=GeCLL2RWV1c

Scubo (2016, with Disney)

| the agile underwater robot
https://www.youtube.com/watch?v=-g2O8e1j3fw

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 35
 ASL
Autonomous Systems Lab

The end

Autonomous Mobile Robots

Roland Siegwart, Margarita Chli, Nick Lawrance 2021 | Locomotion Concepts - add ons | 36