Probabilistic Robotics

Introduction

Probabilities
Bayes rule
Bayes filters
Probabilistic Robotics
Key idea:
Explicit representation of uncertainty
using the calculus of probability theory

• Perception = state estimation

• Action = utility
optimization

2
Axioms of Probability Theory
Pr(A) denotes probability that proposition A is true.

• 0 Pr( A) 1

• Pr(True ) 1 Pr( False) 0

• Pr( A  B ) Pr( A)  Pr( B )  Pr( A  B )

3
A Closer Look at Axiom 3

Pr( A  B ) Pr( A)  Pr( B )  Pr( A  B )

True
A A B B

4
Using the Axioms

Pr( A   A)  Pr( A)  Pr( A)  Pr( A   A)

Pr(True )  Pr( A)  Pr( A)  Pr( False)
1  Pr( A)  Pr( A)  0
Pr( A)  1  Pr( A)

5
Discrete Random Variables

• X denotes a random variable.

• X can take on a countable number of values
in {x1, x2, …, xn}.

• P(X=xi), or P(xi), is the probability that the

random variable X takes on value xi.
.
• P( ) is called probability mass function.

• E.g. P ( Room)  0.7,0.2,0.08,0.02

6
Continuous Random Variables

• X takes on values in the continuum.

• p(X=x), or p(x), is a probability density
function. b
Pr( x  (a, b)) p ( x)dx
a

p(x)

• E.g.

7
Joint and Conditional Probability
• P(X=x and Y=y) = P(x,y)

• If X and Y are independent then

P(x,y) = P(x) P(y)
• P(x | y) is the probability of x given y
P(x | y) = P(x,y) / P(y)
P(x,y) = P(x | y) P(y)
• If X and Y are independent then
P(x | y) = P(x)
8
Law of Total Probability, Marginals
Discrete case Continuous case

 P( x) 1
x
p( x) dx 1
P ( x)  P ( x, y ) p ( x) p ( x, y ) dy
y

P ( x)  P ( x | y ) P ( y ) p ( x) p ( x | y ) p ( y ) dy
y

9
Bayes Formula

P ( x, y )  P ( x | y ) P ( y )  P ( y | x ) P ( x )


P ( y | x) P ( x) likelihood prior
P( x y )  
P( y ) evidence

10
Normalization
P( y | x) P( x)
P( x y )   P ( y | x) P ( x)
P( y )
1
 P ( y ) 1

 P( y | x)P( x)
x

Algorithm:
x : aux x| y P ( y | x) P ( x)

1

 aux x| y
x

x : P ( x | y )  aux x| y

11
Conditioning
• Law of total probability:

P ( x) P ( x, z )dz

P ( x) P ( x | z ) P ( z )dz

P ( x y ) P ( x | y, z ) P ( z | y ) dz

12
Bayes Rule
with Background Knowledge

P ( y | x, z ) P ( x | z )
P( x | y, z ) 
P( y | z )

13
Conditioning
• Total probability:

P ( x) P ( x, z )dz
P ( x) P ( x | z ) P ( z )dz
P ( x y ) P ( x | y, z ) P ( z ) dz

14
Conditional Independence

P ( x, y z ) P ( x | z ) P ( y | z )

equivalent to
P ( x z ) P ( x | z , y )
and
P ( y z ) P ( y | z , x)

15
Simple Example of State Estimation

• Suppose a robot obtains measurement z

• What is P(open|z)?

16
Causal vs. Diagnostic Reasoning

• P(open|z) is diagnostic.
• P(z|open) is causal.
• Often causal knowledge is easier to
obtain. count frequencies!
• Bayes rule allows us to use causal
knowledge:
P ( z | open) P (open)
P (open | z ) 
P( z )

17
Example
• P(z|open) = 0.6 P(z|open) = 0.3
• P(open) = P(open) = 0.5
P ( z | open) P (open)
P (open | z ) 
P ( z | open) p (open)  P ( z |  open) p ( open)
0.6 0.5 2
P (open | z )   0.67
0.6 0.5  0.3 0.5 3

• z raises the probability that the door is open.

18
Combining Evidence
• Suppose our robot obtains another
observation z2.

• How can we integrate this new

information?

• More generally, how can we estimate

P(x| z1...zn )?

19
 Recursive Bayesian Updating
P ( zn | x, z1,  , zn  1) P ( x | z1,  , zn  1)
P ( x | z1,  , zn) 
P ( zn | z1,  , zn  1)

Markov assumption: zn is independent of z1,...,zn-1 if

we know x.
P( zn | x) P( x | z1,  , zn  1)
P( x | z1,  , zn) 
P( zn | z1,  , zn  1)
 P( zn | x) P( x | z1,  , zn  1)
1...n  P( z | x) P( x)
i 1...n
i

20
Example: Second Measurement

• P(z2|open) = 0.5 P(z2|open) = 0.6

• P(open|z1)=2/3
P ( z 2 | open) P (open | z1 )
P (open | z 2 , z1 ) 
P ( z 2 | open) P (open | z1 )  P ( z 2 |  open) P ( open | z1 )
1 2

2 3 5
   0.625
1 2 3 1 8
  
2 3 5 3

• z2 lowers the probability that the door is open.

21
A Typical Pitfall

• Two possible locations x1 and x2

• P(x1)=0.99
• P(z|x2)=0.09 P(z|x1)=0.07
1
p(x2 | d)
0.9 p(x1 | d)

0.8

0.7

0.6
p( x | d)

0.5

0.4

0.3

0.2

0.1

0
5 10 15 20 25 30 35 40 45 50
Number of integrations

22
Actions

• Often the world is dynamic since

• actions carried out by the robot,
• actions carried out by other agents,
• or just the time passing by
change the world.

• How can we incorporate such

actions?

23
Typical Actions

• The robot turns its wheels to move

• The robot uses its manipulator to grasp
an object
• Plants grow over time…

• Actions are never carried out with

absolute certainty.
• In contrast to measurements, actions
generally increase the uncertainty.

24
Modeling Actions

• To incorporate the outcome of an

action u into the current “belief”, we
use the conditional pdf

P(x|u,x’)

• This term specifies the pdf that

executing u changes the state
from x’ to x.

25
Example: Closing the door

26
State Transitions
P(x|u,x’) for u = “close door”:

0.9
0.1 open closed 1
0

If the door is open, the action “close

door” succeeds in 90% of all cases.
27
Integrating the Outcome of Actions
Continuous case:

P ( x | u ) P ( x | u , x' ) P ( x' )dx'

Discrete case:

P ( x | u )  P ( x | u , x' ) P ( x' )

28
Example: The Resulting Belief
P(closed | u )  P (closed | u , x' ) P( x' )
P (closed | u , open) P(open)
 P(closed | u , closed ) P (closed )
9 5 1 3 15
    
10 8 1 8 16
P (open | u )  P (open | u , x' ) P ( x' )
P (open | u , open) P (open)
 P(open | u , closed ) P (closed )
1 5 0 3 1
    
10 8 1 8 16
1  P(closed | u )
29
Bayes Filters: Framework
• Given:
• Stream of observations z and action data u:
d t {u1 , z1  , ut , zt }
• Sensor model P(z|x).
• Action model P(x|u,x’).
• Prior probability of the system state P(x).
• Wanted:
• Estimate of the state X of a dynamical system.
• The posterior of the state is also called Belief:
Bel ( xt ) P ( xt | u1 , z1  , ut , zt )

30
Markov Assumption

p( zt | x0:t , z1:t , u1:t )  p( zt | xt )

p( xt | x1:t  1 , z1:t , u1:t )  p( xt | xt  1 , ut )
Underlying Assumptions
• Static world
• Independent noise
• Perfect model, no approximation errors
31
 z = observation
u = action
Bayes Filters x = state

Bel ( xt ) P( xt | u1 , z1  , ut , zt )
Bayes  P( zt | xt , u1 , z1 ,  , ut ) P( xt | u1 , z1 ,  , ut )
Markov  P ( zt | xt ) P ( xt | u1 , z1 ,  , ut )
Total prob.  P( zt | xt ) P( xt | u1 , z1 ,  , ut , xt  1 )
P( xt  1 | u1 , z1 ,  , ut ) dxt  1
Markov  P( zt | xt ) P( xt | ut , xt  1 ) P( xt  1 | u1 , z1 ,  , ut ) dxt  1
Markov  P ( zt | xt ) P ( xt | ut , xt  1 ) P( xt  1 | u1 , z1 ,  , zt  1 ) dxt  1

 P( zt | xt ) P( xt | ut , xt  1 ) Bel ( xt  1 ) dxt  1

32
Bayes
Bel ( xt ) Filter Algorithm
P( zt | xt ) P( xt | ut , xt  1 ) Bel ( xt  1 ) dxt  1

1. Algorithm Bayes_filter( Bel(x),d ):

2. 0
3. If d is a perceptual data item z then
4. For all x do
5. Bel ' ( x) P ( z | x) Bel ( x)
6.    Bel ' ( x)
7. For all x do
8. Bel ' ( x)   1 Bel ' ( x)
9. Else if d is an action data item u then
10. For all x do
11. Bel ' ( x) P ( x | u , x' ) Bel ( x' ) dx'
12. Return Bel’(x)

33
Bayes Filters are Familiar!

Bel ( xt )  P( zt | xt ) P( xt | ut , xt  1 ) Bel ( xt  1 ) dxt  1

• Kalman filters
• Particle filters
• Hidden Markov models
• Dynamic Bayesian networks
• Partially Observable Markov Decision
Processes (POMDPs)

34
Summary
• Bayes rule allows us to compute
probabilities that are hard to assess
otherwise.
• Under the Markov assumption,
recursive Bayesian updating can be
used to efficiently combine evidence.
• Bayes filters are a probabilistic tool
for estimating the state of dynamic
systems.
35