Humanoid Robots 3: Balance recap sufficient condition for balance: - - PowerPoint PPT Presentation

humanoid robots 3
SMART_READER_LITE
LIVE PREVIEW

Humanoid Robots 3: Balance recap sufficient condition for balance: - - PowerPoint PPT Presentation

Autonomous and Mobile Robotics Prof. Giuseppe Oriolo Humanoid Robots 3: Balance recap sufficient condition for balance: ZMP inside the support polygon ZMP dynamics modeled from Newton-Euler equations approximate model: Linear


slide-1
SLIDE 1

Autonomous and Mobile Robotics

  • Prof. Giuseppe Oriolo

Humanoid Robots 3:

Balance

slide-2
SLIDE 2

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

2

recap

  • sufficient condition for balance: ZMP inside the support polygon
  • ZMP dynamics modeled from Newton-Euler equations
  • approximate model: Linear Inverted Pendulum (LIP)
slide-3
SLIDE 3

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

3

Linear Inverted Pendulum: basic scope

Although extremely simplified, the LIP equation describes in first approximation the time evolution of the CoM trajectory. Moreover

  • it defines a differential relationship between the CoM trajectory

and the ZMP (or CMP) time evolution

  • it is easier to design a controller which makes the actual CoM

follow a desired behaviour

  • dynamic balancing will be characterized in terms of the ZMP
  • the problem will then be to understand which CoM trajectory,

solution of the LIP equation, guarantees that dynamic balancing is achieved

slide-4
SLIDE 4

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

4

ZMP

Support Polygons during Double Support vertical component

  • f the GRF
slide-5
SLIDE 5

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

5

ZMP - 2D case

CoP/ZMP equivalent force/torque generic px specific px

components of the GRF

slide-6
SLIDE 6

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

6

ZMP - 3D case

vertical component of the GRF

slide-7
SLIDE 7

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

7

ZMP - 3D case

horizontal component of the GRF

if robot moves, the z component will be different from 0

slide-8
SLIDE 8

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

8

ZMP as long as the ZMP is in the Support Polygon, the support foot will not rotate z z

slide-9
SLIDE 9

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

9

ZMP

slide-10
SLIDE 10

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

10

static balance

humanoid motionless: statically balanced robots keep the center of mass within the polygon of support in order to maintain postural stability (sufficient when the robot moves slow enough so all the inertial forces are negligible) statically balanced statically unbalanced equivalent representation: mass M on a massless table with finite length base the table starts tipping over if the CoM stays within these boundaries no tipping over occurs M M M

slide-11
SLIDE 11

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

11

GRF

gravity force

dynamic balance

Question: how do you keep a pendulum in a non-vertical position? Answer: by continuously accelerating it acceleration

M

non-inertial frame (pendulum stands still in an accelerating frame) inertial force (fictitious force) Hyp: no torque around the CoM

slide-12
SLIDE 12

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

12

GRF

inertial force gravity force acceleration humanoid walking: the GRF will also have a component parallel to the ground; the motion requires the exchange of horizontal frictional force with the ground hyp: no torque around the CoM

dynamic balance

slide-13
SLIDE 13

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

13

Ground Reaction Force (GRF): 2D components (x,z) GRF

inertial force gravity force acceleration hyp: no torque around the CoM from the previous derivation: GRF GRF

x

= GRF

z

= Mg

dynamic balance

slide-14
SLIDE 14

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

14

which robot falls down?

slide-15
SLIDE 15

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

15

statically balanced dynamically balanced

inertial force

falls

which robot falls down?

slide-16
SLIDE 16

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

16

where is the ZMP?

use the dynamics equation on horizontal flat ground and neglect zx (ZMP): point on the ground where the GRF is applied

slide-17
SLIDE 17

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

17

statically balanced dynamically balanced

inertial force

falls

where is the ZMP?

hyp CoM at constant height LIP equation in the sagittal plane

slide-18
SLIDE 18

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

18

vertical acceleration is generated

what if CoM acceleration is increased?

inertial force inertial force gravity from the previous analysis one could think that the ZMP, increasing the CoM acceleration, would leave the support foot support, but it doesn’t

  • once the ZMP has reached the foot border, a rotation starts

around that point

  • with the rotation of the foot, the center of mass starts

accelerating vertically

  • with a vertical acceleration of the CoM, its height does not

remain constant

  • model changes, ZMP remains constant

the vertical CoM acceleration generates a vertical inertial force

slide-19
SLIDE 19

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

19

dynamically balanced

sum of moments around zx Hyp: no torque around the CoM

zx cx cz

+ torque ¿y around the CoM

zx cx cz ¿y

Centroidal Moment Point CMP

slide-20
SLIDE 20

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

20

dynamically balanced

+ torque ¿y around the CoM (or equivalently an ankle torque ¿y)

zx cx cz ¿y

positive torque ¿y (counter-clockwise) moves the Center of Pressure CoP to the right

zx is not the CoP anymore

CoP

¿y ¿y ¿y ¿y

CoP CoP

slide-21
SLIDE 21

Oriolo: Autonomous and Mobile Robotics - Humanoid Robots 3

21

dynamically balanced locomotion

generate a gait for walking while maintaining balance maintaining balance ZMP-based criterion assumed to be equivalent to not tipping over the support foot ZMP needs to stay inside the Support Polygon