probabilistic approaches to reasoning and control: towards autonomous interactive mobile robots...

46
Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar March 28, 2003

Upload: dwain-wheeler

Post on 18-Jan-2016

218 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control:

Towards autonomous interactive mobile robots

Joelle PineauCarnegie Mellon University

TAMALE Seminar

March 28, 2003

Page 2: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Our vision of robotic-assisted health-care

Moving thingsaround

Moving thingsaround

Enabling use of remote

health services

Enabling use of remote

health services

Supportinginter-personal

communication

Supportinginter-personal

communication

Calling for helpin emergencies

Calling for helpin emergencies

Monitoring Rx adherence

& safety

Monitoring Rx adherence

& safety

Providinginformation

(TV, weather)

Providinginformation

(TV, weather)

Management support of

ADLs

Management support of

ADLsReminding

to eat, drink, & take meds

Reminding to eat, drink, & take meds

Providing physical

assistance

Providing physical

assistance

Linking the caregiver to resources

Linking the caregiver to resources

Page 3: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Introducing Pearl: A mobile robotic assistant for elderly people and nurses

cameras

sonars

handle bars

mobile base

carrying tray

LCD mouth

touchscreen

microphone& speakers

laser

Page 4: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

What are the challenges?

• Interaction with the environment:

– navigating robustly

– handling dynamic obstacles

• Interaction with individuals:

– communicating by speech

– providing cognitive reminders

– interpreting and satisfying user requests

Page 5: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

System Overview

Cognitive supportNavigation Communication

High-level controller

Page 6: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

System Overview

Cognitive support Communication

High-level controller

• Localization and map building(Burgard et al., 1999)

• People detection and tracking(Montemerlo et al., 2002)

Navigation

Page 7: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Navigation and people tracking

Page 8: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

System Overview

Navigation Communication

High-level controller

• Autominder system (Pollack et al., 2002)

Cognitive support

Page 9: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

• Speech recognition: Sphinx system(Ravishankar, 1996)

• Speech synthesis: Festival system(Black et al., 1999)

System Overview

Cognitive supportNavigation

High-level controller

Communication

Page 10: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Speech recognition with Sphinx

Page 11: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

The role of the top-level controller

Cognitive supportNavigation Communication

ACTION SELECTION - based on the trade-off between:

- goals from different modules;

- goals with varying costs / rewards;

- reducing uncertainty versus accomplishing goals.

High-level controller

Page 12: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Types of uncertainty in robotics

• Cause #1: Non-deterministic effects of actions

• Cause #2: Partial and noisy sensor information

• Cause #3: Inaccurate model of the world and the user

Page 13: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Robot control under uncertainty using Partially Observable Markov Decision Processes

State

User + Environment + Robot

Action={ say-weather,update-appointment,clarify-query}

Speech=“today”Belief State

e.g. request-weather-today

e.g. P(st=weather-today)=0.5 P(st=appointment-today )=0.5

Page 14: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Existing applications of POMDPs

– Maintenance scheduling

» Puterman, 1994

– Robot navigation

» Koenig & Simmons, 1995;

Roy & Thrun, 1999

– Helicopter control

» Bagnell & Schneider, 2001;

Ng et al., 2002

– Dialogue modeling

» Roy, Pineau & Thrun, 2000;

Peak&Horvitz, 2000

– Preference elicitation

» Boutilier, 2002

Page 15: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Graphical Model Representation

POMDP is n-tuple { S, A, , T, O, R }:

What goes on: st-1 st

at-1 at

T(s,a,s’) = state-to-state transition probabilitiesO(s,a,o) = observation generation probabilitiesR(s,a) = Reward function

S = state setA = action set = observation set

What we see: ot-1 ot-1

Belief update:

Ss

jijii

j

tt sbsasToasOsb 1,,,,

Page 16: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Understanding the belief state

• A belief is a probability distribution over states

Where Dim(B) = |S|-1

– E.g. Let S={s1, s2}

P(s1)

0

1

Page 17: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Understanding the belief state

• A belief is a probability distribution over states

Where Dim(B) = |S|-1

– E.g. Let S={s1, s2, s3}

P(s1)

P(s2)

0

1

1

Page 18: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

• A belief is a probability distribution over states

Where Dim(B) = |S|-1

– E.g. Let S={s1, s2, s3 , s4}

Understanding the belief state

P(s1)

P(s2)

0

1

1

P(s3)

Page 19: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Exact planning for POMDPs

• Simple problem: |S|=2, |A|=3, ||=2 Iteration # hyper-planes 0 1

P(s1)

V0(b)

b

Page 20: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Exact planning for POMDPs

• Simple problem: |S|=2, |A|=3, ||=2 Iteration # hyper-planes 0 1 1 3

P(s1)

V1(b)

b

Page 21: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Exact planning for POMDPs

• Simple problem: |S|=2, |A|=3, ||=2 Iteration # hyper-planes 0 1 1 3

P(s1)

V1(b)

b

Page 22: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Exact planning for POMDPs

• Simple problem: |S|=2, |A|=3, ||=2 Iteration # hyper-planes 0 1 1 3 2 27

P(s1)

V2(b)

b

Page 23: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Exact planning for POMDPs

• Simple problem: |S|=2, |A|=3, ||=2 Iteration # hyper-planes 0 1 1 3 2 27 3 2187

P(s1)

V2(b)

b

Page 24: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Exact planning for POMDPs

• Simple problem: |S|=2, |A|=3, ||=2 Iteration # hyper-planes 0 1 1 3 2 27 3 2187 4 14,348,907

P(s1)

V2(b)

b

Page 25: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Properties of exact planning

• Value function is always piecewise-linear convex

• Many hyper-planes can be pruned away

P(s1)

V2(b)

b

|S|=2, |A|=3, ||=2 Iteration # hyper-planes

0 1 1 3 2 5 3 9 4 7 5 13 10 27 15 47 20 59

Page 26: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Is pruning sufficient?

|S|=20, |A|=6, ||=8

Iteration # hyper-planes0 11 5

2 213 3 ?????

Not for this problem!

Page 27: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Certainly not for this problem!

Physiotherapy

Patientroom

Robothome

|S|=576, |A|=19, |O|=17

State Features: {RobotLocation, ReminderGoal, UserLocation, UserMotionGoal,

UserStatus, UserSpeechGoal}

Page 28: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

The two curses of POMDP planning

• The curse of dimensionality:

– the dimension of each hyper-plane = # of states

• The curse of history:

– the number of hyper-planes grows

exponentially with the planning horizon

||1

2 |||||| nAS

|| n

Complexity of POMDP value iteration:

dimensionality history

Page 29: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Methods to solve POMDPs

Complexity

Performance

QMDP

MDP

FIB

Grid

O(S2A) O(S2AO )O(S2AO) O(S2AB) T

POMDP

New methods?

Objective: Find a policy, (b), which maximizes reward.

Page 30: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

New approach: A hierarchy of POMDPs

Idea: Exploit domain knowledge to divide one POMDP into many smaller ones.

Motivation: Smaller action sets help overcome the curse of history.

Assumption: We are given POMDP M = {S,A,,b,T,O,R} and hierarchy H

Act

ExamineHealth Navigate

MoveVerifyPulse

ClarifyGoal

North South East West

VerifyMeds

subtask

abstract action

primitive action

Page 31: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

PolCA+: Planning with a hierarchy of POMDPs

Navigate

Move ClarifyGoal

South East WestNorth

AMove = {N,S,E,W}

ACTIONSNorthSouthEastWest

ClarifyGoalVerifyPulseVerifyMeds

ACTIONSNorthSouthEastWest

ClarifyGoalVerifyPulseVerifyMeds

Step 1: Select the action set

Page 32: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

PolCA+: Planning with a hierarchy of POMDPs

Navigate

Move ClarifyGoal

South East WestNorth

AMove = {N,S,E,W}

SMove = {X,Y}

STATE FEATURESX-positionY-position

X-goalY-goal

HealthStatus

STATE FEATURESX-positionY-position

X-goalY-goal

HealthStatus

ACTIONSNorthSouthEastWest

ClarifyGoalVerifyPulseVerifyMeds

ACTIONSNorthSouthEastWest

ClarifyGoalVerifyPulseVerifyMeds

Step 1: Select the action set

Step 2: Minimize the state set

Page 33: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

PolCA+: Planning with a hierarchy of POMDPs

Navigate

Move ClarifyGoal

South East WestNorth

AMove = {N,S,E,W}

SMove = {X,Y}

STATE FEATURESX-positionY-position

X-goalY-goal

HealthStatus

STATE FEATURESX-positionY-position

X-goalY-goal

HealthStatus

ACTIONSNorthSouthEastWest

ClarifyGoalVerifyPulseVerifyMeds

ACTIONSNorthSouthEastWest

ClarifyGoalVerifyPulseVerifyMeds

PARAMETERS

{bh,Th,Oh,Rh}

PARAMETERS

{bh,Th,Oh,Rh}

Step 1: Select the action set

Step 2: Minimize the state set

Step 3: Choose parameters

Page 34: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

PolCA+: Planning with a hierarchy of POMDPs

Navigate

Move ClarifyGoal

South East WestNorth

AMove = {N,S,E,W}

SMove = {X,Y}

STATE FEATURESX-positionY-position

X-goalY-goal

HealthStatus

STATE FEATURESX-positionY-position

X-goalY-goal

HealthStatus

ACTIONSNorthSouthEastWest

ClarifyGoalVerifyPulseVerifyMeds

ACTIONSNorthSouthEastWest

ClarifyGoalVerifyPulseVerifyMeds

PLAN

h

PLAN

h

PARAMETERS

{bh,Th,Oh,Rh}

PARAMETERS

{bh,Th,Oh,Rh}

Step 1: Select the action set

Step 2: Minimize the state set

Step 3: Choose parameters

Step 4: Plan task h

Page 35: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Results on small dialogue domain

-120

-100

-80

-60

-40

-20

0

0.01 0.1 1 10 100 1000 10000 100000 1000000

Time (secs)

R

POMDPPolCA-D1PolCA-D2FIBQMDP

|S|=12, |A|=20, |O|=3

Page 36: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Achieving a flexible trade-off

Planning time

Reward

QMDP

FIB

POMDP

PolCA+ D2

PolCA+ D1

Page 37: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

PolCA+ in the Nursebot domain

• Goal: A robot is deployed in a nursing home, where it provides reminders to elderly users and accompanies them to appointments.

Page 38: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Sample scenario

Page 39: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Comparing user performance

0.1 0.10.18

Page 40: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

The effects of confirmation actions

-2000

2000

6000

10000

14000

0 400 800 1200

Time Steps

Cu

mu

lativ

e R

ew

ard

PolCA+

PolCA

QMDP

Page 41: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

0

500

1000

1500

2000

2500

3000

3500

4000

4500

NoAbs PolCA PolCA+

# S

tate

ssubInform

subMove

subContact

subRest

subAssist

subRemind

act

Addressing the curse of dimensionality

Page 42: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Ongoing work

• New POMDP approximation techniques.

• Parameter estimation for adaptation to user-specific speech patterns and preferences.

• Exploration of emotion and personality types using a new head.

• Addition of an arm for object manipulation.

• Addition of weight-bearing bars for assisted walking.

Page 43: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Summary

• We have developed a first prototype robot able to serve as a mobile nursing assistant for elderly people.

• The top-level controller uses a hierarchical variant of POMDPs to select actions.

– PolCA+ addresses both the curse of dimensionality and the curse of history.

• Lessons learned during our experiments:

– Uncertainty is crucial when dealing with people

– Probabilistic techniques are necessary to reason about uncertainty.

– Real belief tracking and planning really matters!

Page 44: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Project information: www.cs.cmu.edu/~nursebotNavigation software: www.cs.cmu.edu/~carmenPapers and more: www.cs.cmu.edu/~jpineau

Joint work with: Michael Montemerlo, Martha Pollack, Nicholas Roy, Sebastian Thrun

Page 45: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

The Nursebot project in its early days

Page 46: Probabilistic approaches to reasoning and control: Towards autonomous interactive mobile robots Joelle Pineau Carnegie Mellon University TAMALE Seminar

Probabilistic approaches to reasoning and control for interactive mobile robots Joelle Pineau

Autominder System