Lecture 4-1 CS251: Intro to AI/Lisp II

Robots in Action

Lecture 4-1 CS251: Intro to AI/Lisp II

Announcements

• Feedback response– Late policy (Some credit, helps grading)– Structure of course project (Tyranny of the

majority, grading)– PowerPoint vs. chalk talk: doing the

reading

• Homework assigned today

• Course project descriptions

Lecture 4-1 CS251: Intro to AI/Lisp II

Asimov’s Three Laws

• A robot may not injure a human being, or, through inaction, allow a human being to come to harm.

• A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.

• A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Lecture 4-1 CS251: Intro to AI/Lisp II

What’s a Robot?

• Mobile?

• Autonomous

• Softbots

Lecture 4-1 CS251: Intro to AI/Lisp II

Lecture 4-1 CS251: Intro to AI/Lisp II

Lecture 4-1 CS251: Intro to AI/Lisp II

Lecture 4-1 CS251: Intro to AI/Lisp II

Lecture 4-1 CS251: Intro to AI/Lisp II

Snips and Snails and Puppy Dog Tails, that’s what robots are made of

• Effectors– Actuators– Degrees of freedom

• Sensors– Proprioception (Looking at your own hand)

Lecture 4-1 CS251: Intro to AI/Lisp II

Motion for Robots

• Degrees of freedom

Lecture 4-1 CS251: Intro to AI/Lisp II

Different Sensor, Different Task

• SONAR– Obstacle avoidance

• Lasers– Range-finding

• Vision– Obstacle avoidance– Proprioception

Lecture 4-1 CS251: Intro to AI/Lisp II

Robot Architecture

• Designing a robot– Common features of many different robots

• Classical

• Nouvelle AI (Situated automata)

Lecture 4-1 CS251: Intro to AI/Lisp II

Classical (aka SHAKEY)

• Theorem provers proved too general

• No execution monitoring

• Version 2– Specialized programs (LLAs, ILAs)

• Modeling uncertainty

– Learning with macro operators– PLANEX

Lecture 4-1 CS251: Intro to AI/Lisp II

SHAKEY

• Conclusions– Limited ability to handle unexpected

outcomes

– Each move took 1 hour of computing time• High probability of failure

– STRIPS produced good plans

– Sensory interpretation primitive

From http://hebb.cis.uoguelph.ca/~deb/Robotics/Notes/traditional/page5.html

Lecture 4-1 CS251: Intro to AI/Lisp II

Situated Automata

• Is classical robotics too difficult?

• Toss out the representation

• Embedded agents– Model the world as interacting automata– Physical environment + Agent– Local state of one = f(Signals from other)– Flakey

Lecture 4-1 CS251: Intro to AI/Lisp II

Elephants Don’t Play Chess

• What does this mean?

Lecture 4-1 CS251: Intro to AI/Lisp II

(Physical) Symbol Systems

• Biologically implausible

• Frame problem

• Planning is hard– NP-complete– Heuristics

Lecture 4-1 CS251: Intro to AI/Lisp II

Physical Grounding

• What’s the hypothesis?

• Evolution– What is Brooks’ argument?

Lecture 4-1 CS251: Intro to AI/Lisp II

Brooks’ Robots

• Allen

• Tom & Jerry

• Herbert

• Genghis

• Squirt

• Toto

• Seymour

• Gnats

• Ant farm

Lecture 4-1 CS251: Intro to AI/Lisp II

Subsumption, what is good for?

Lecture 4-1 CS251: Intro to AI/Lisp II