Presentation is loading. Please wait.

Presentation is loading. Please wait.

3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations1 Biological Foundations of the Reactive Paradigm Review Why? -comp. theory IRM.

Similar presentations


Presentation on theme: "3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations1 Biological Foundations of the Reactive Paradigm Review Why? -comp. theory IRM."— Presentation transcript:

1 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations1 Biological Foundations of the Reactive Paradigm Review Why? -comp. theory IRM Perception -Summary Describe the three levels in a Computational Theory. Explain in one or two sentences each of the following terms: reflexes, taxes, fixed-action patterns, schema, affordance. Be able to write pseudo-code of an animal’s behaviors in terms of innate releasing mechanisms, identifying the releasers for the behavior. Given a description of an animal's sensing abilities, its task, and environment, identify an affordance for each behavior. Given a description of an animal's sensing abilities, its task, and environment, define a set of behaviors using schema theory to accomplish the task.

2 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations2 Robots In the Hierarchical Paradigm

3 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations3 Timeline of Influences 19801990 1970 Braitenberg’s Vehicles Arbib’s Schemas Tinbergen & Lorenz & von Frisch Marr’s Computational Theory J.J. Gibson Neisser Middlestat Arkin’s Schemas Brook’s Insects Payton

4 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations4 Marr’s Computational Theory Level 1: What is the phenomena we’re trying to represent? for (i=nCol.. Level 2: How it be represented as a process with inputs/outputs? Level 3: How is it implemented?

5 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations5 Level 1: Existence Proof Level 1: What is the phenomena we’re trying to represent? Goal: how to make line drawings of objects? people can do this by age 10, computers should

6 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations6 Level 2: Inputs, Outputs, Transforms for (i=nCol.. Level 2: How it be represented as a process with inputs/outputs? light drawing retina (gradient) light lines (edges)drawing

7 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations7 Level 3: Implementation Level 3: How is it implemented? + - - - - - - - - Center Surround Cell in retinal ganglion Sobel Edge Detector in computer vision 0 +2 +10 -2 0+1 0 0 -2 0 +1 +2+1

8 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations8 Class Discussion Give three examples of how biology has influenced modern technology? –ex. Wright Brothers- control flaps on airplane wings

9 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations9 Behavior Definition (graphical) BEHAVIOR Sensory Input Pattern of Motor Actions

10 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations10 Types of Behaviors Reflexive –stimulus-response, often abbreviated S-R Reactive –learned or “muscle memory” Conscious –deliberately stringing together WARNING Overloaded terms: Roboticists often use “reactive behavior” to mean purely reflexive, And refer to reactive behaviors as “skills” WARNING Overloaded terms: Roboticists often use “reactive behavior” to mean purely reflexive, And refer to reactive behaviors as “skills”

11 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations11 Reflexive behaviors Reflexes - lasts as long as the stimulus only, Taxes - moves in a particular direction (tropotaxis in baby turtles, chemotaxis in ants), Fixed-action patterns - continues for a longer duration than the stimulus.

12 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations12 Ethology: Study of Animal Behaviors Nobel 1973 in physiology or medicine von Frisch Lorenz Tinbergen www.nobel.se INNATE RELEASING MECHANISMS

13 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations13 Ways to acquire a behavior To be born with a behavior (innate) To be born with a sequence of innate behaviors (sequence of innate behaviors) To be born with behaviors that need some initialization (innate with memory) To learn a set of behaviors

14 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations14 Arctic Terns Arctic terns live in Arctic (black, white, gray environment, some grass) but adults have a red spot on beak When hungry, baby pecks at parent’s beak, who regurgitates food for baby to eat How does it know its parent? –It doesn’t, it just goes for the largest red spot in its field of view (e.g., ethology grad student with construction paper) –Only red thing should be an adult tern –Closer = large red

15 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations15 behavior template BEHAVIOR Sensory Input Pattern of Motor Actions

16 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations16 “the feeding behavior” Feeding BEHAVIOR Sensory Input Pattern of Motor Actions REDPECK AT RED

17 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations17 the releaser template Releaser present? N Y /dev/null Sensory input and/or internal state

18 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations18 Example IRM in pseudo-code (assume synchronous processing, update occurs every N ms or “1 step”) enum Releaser={PRESENT, NOT_PRESENT}; Releaser predator; while (TRUE) { predator = sensePredators(); //perception of releaser if (predator == PRESENT) //releaser flee(); //behavior } Perception of predator “persists” even if not directly in sight leading to a Fixed-Pattern Action style of behavior

19 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations19 “the feeding releaser” Feeding BEHAVIOR REDPECK AT RED Releaser present? N Y /dev/null internal state RED & HUNGRY sensory input

20 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations20 Compound Releaser enum Releaser={PRESENT, NOT_PRESENT}; Releaser food; while (TRUE) { food = senseFood(); //perception of releaser Part 1 hungry = checkState();//perception of internal state if (food == PRESENT && hungry==PRESENT) //releaser feed(); } Must be hungry AND sense food to feed

21 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations21 Implicit Sequence enum Releaser={PRESENT, NOT_PRESENT}; Releaser food, hungry, nursed; while (TRUE) { food = sense(); hungry = checkStateHunger(); child = checkStateChild(); if (hungry==PRESENT) searchForFood(); //sets food = PRESENT when done if (hungry==PRESENT && food==PRESENT) feed(); // sets hungry = NOT_PRESENT when done if (hungry== NOT_PRESENT && parent==PRESENT) nurse(); // set nursed = PRESENT when done if (nursed ==PRESENT) sleep(); } Why is this implicit, not explicit? If asynchronous, could have conflicts

22 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations22 Innate Releasing Mechanisms BEHAVIOR Sensory Input Pattern of Motor Actions Releaser present? N Y /dev/null Sensory input and/or internal state

23 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations23 Example: Hide Behavior shows –taxis (oriented relative to light, wall, niche) –fixed action pattern (persisted after light was off) –reflexive (stimulus, response) –implicit sequencing –use of internal state

24 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations24 Example: Cockroach Hide light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out even if the lights are turned back off earlier

25 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations25 Reflexive Behaviors S-R light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out even if the lights are turned back off earlier

26 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations26 Fixed Pattern Actions light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out even if the lights are turned back off earlier

27 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations27 Exhibits Taxis light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out even if the lights are turned back off earlier to light to wall to niche

28 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations28 Class Exercise Draw flowchart of how this works light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out even if the lights are turned back off earlier light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out even if the lights are turned back off earlier

29 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations29 Break into Behaviors light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out Flee Follow- wall hide

30 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations30 Find Releasers light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out Flee Follow- wall hide LIGHT present? N Y SCARED & SURROUNDED present? N LIGHT present? N Ooops, need internal state: Scared Ooops, need internal state: Scared

31 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations31 Internal State Set light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out Flee Follow- wall hide LIGHT present? N Y SCARED & SURROUNDED present? N BLOCKED & SCARED present? N SCARED

32 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations32 Action light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out Flee Follow- wall hide LIGHT present? N Y SCARED & SURROUNDED present? N BLOCKED & SCARED present? N SCARED steer 180, drive forward steer =F(dist to wall) drive forward const. steer =F(dist to wall) drive forward const. stop

33 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations33 Sensory Input light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out Flee Follow- wall hide LIGHT present? N Y SCARED & SURROUNDED present? N BLOCKED & SCARED present? N SCARED steer 180, drive forward steer =F(dist to wall) drive forward const. steer =F(dist to wall) drive forward const. stop encoders IR

34 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations34 How Do You Link Them? light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out Flee Follow- wall hide LIGHT present? N Y SCARED & SURROUNDED present? N BLOCKED & SCARED present? N SCARED steer 360, drive forward steer =F(dist to wall) drive forward const. steer =F(dist to wall) drive forward const. stop encoders IR

35 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations35 Analogy: IRMs work on THREADS, not sequential processing! Very simple modules Nice building blocks since not directly linked If one module (part of brain) fails, what happens?

36 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations36 What happens when there’s a conflict from concurrent behaviors? Equilibrium –Feeding squirrels: feed or flee - hesitate in-between Dominance –Sleepy, hungry: either sleep or eat Cancellation –Sticklebacks : defend or attack - build a nest instead ?

37 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations37 Inhibition while (TRUE) { predator = sensePredator(); //has a time delay if (predator==PRESENT) //as long as predator persists flee(); else { food = senseFood(); hungry = checkStateHunger();... } } Could also be done as an interrupt

38 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations38 Perception Two uses of perception (can be the same percept) –Release a behavior –Guide a behavior Action-oriented perception (Neisser) –Planning is not needed to act –Perception is selective Cognitive Activity World Perception of Environment Samples, Finds Potential Actions Acts & Modifies World Directs what to look for

39 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations39 Gibson’s Ecological Approach Acting and sensing co-evolved as agent survived in a particular environment. The environment affords the agent what it needs to survive. The perception needed to release or guide the “right action” is directly in the environment, not inferred or memorized –Ex. Red on Artic Terns== food –Ex. Sound of filling container==full Percepts are called affordances or said to be obtained through direct perception

40 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations40 Gibsonian Affordances How do you know you’re going fast in a car? Or in a space movie? How do animals know when to mate? How do mosquitoes know to bite in the most tender areas? What should you do when you think you’re being stalked by a mountain lion? What’s your favorite fishing lure?

41 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations41 Sittability

42 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations42 But does this really hold for everything? “my car” versus “your car”? –Difference is where I parked it (memory) Semantic meaning (cars aren’t generic like nuts to a squirrel) Neisser’s Two Systems –Direct Perception: older, behavioral –Recognition: evolved later, deliberative

43 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations43 Review Questions What are the levels of a computational theory? Existence proof, inputs-outputs-transformations, implementation What is a behavior? A behavior is a mapping of sensory inputs to a pattern of motor actions Is sequencing normally implicit or explicit in IRM? implicit What is an affordance? A potentiality in the environment for an action

44 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations44 Schema Theory schema- is used in cognitive science and ethology to refer to a particular organized way of perceiving cognitively and responding to a complex situation or set of stimuli is generic, equivalent to an object in OOP –schema specific knowledge (local data) –procedural knowledge (methods) schema instantation is specific to a situation, equivalent to an instance in OOP a behavior is a schema, consists of –perceptual schema –motor schema

45 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations45 Behavioral Schema Motor Schema (MS) Perceptual Schema (PS) percept, gain action, intensity alternative PS, MS sequencing logic for reactive skills (judgment value function) Reflexive behaviors usually just have “methods”, not data

46 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations46 Ex. Fly Snapping Behavior IRM snap(blob)track(blob) x,y,z, 100% snap, 100% Releaser: small moving dark blob present? N Y /dev/null

47 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations47 Schema Instantiation (SI) snap(blob)track(blob) x,y,z, 100% snap, 100% Releaser: small moving dark blob present? N Y /dev/null

48 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations48 Schema/Schema Instantiation behavior schema releaser present? N 1. snap(blob)track(blob) snap(blob) track(blob) Perceptual Schema Library Motor Schema Library 2. snap(blob)track(blob) x,y,z, 100% 3.

49 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations49 Advantages modular can assemble new behaviors from existing schemas –learning by experimentation can substitute alternatives –reroute nerves

50 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations50 Instantiation for each eye snap, 100% snap(blob)track(blob) x,y,z, 100% Releaser: small moving dark blob present? N Y /dev/null snap, 100% snap(blob)track(blob) x,y,z, 100% Releaser: small moving dark blob present? N Y /dev/null Left eye Right eye Snap at vector sum (middle)

51 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations51 Where’s the MS and PS? light goes on, the cockroach turns and runs when it gets to a wall, it follows it when it finds a hiding place (thigmotrophic), goes in and faces outward waits until not scared, then comes out Flee Follow- wall hide LIGHT present? N Y SCARED & SURROUNDED present? N BLOCKED & SCARED present? N SCARED steer 180, drive forward steer =F(dist to wall) drive forward const. steer =F(dist to wall) drive forward const. stop encoders IR

52 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations52 General Principles All animals possess a set of behaviors Releasers for these behaviors rely on both internal state and external stimulus Perception is filtered; perceive what is relevant to the task Some behaviors and associated perception do not require explicit knowledge representation (e.g., rely on affordances)

53 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations53 Silicon v. Carbon Individual robots must survive, not species –detection of non-productive behaviors –graceful degradation Must be able to predict emergent behaviors Not clear how to learn quickly Robots need more alternative perceptual schemas since poorer understanding of the environment

54 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations54 Unresolved Issues How to resolve conflicts? –behavioral arbitration/combination When is explicit representations, memory needed? How to set up or learn new sequences of behaviors What are the affordances for a particular ecology?

55 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations55 Take Home Thoughts… Ideas bubbling up for robotics –Maybe programming in terms of behaviors is better than STRIPS or trying to set up a complex hierarchy –Intelligence has something to do with agent’s ecological niche: its abilities, its tasks (survival), and environment –Perception is going to be critical because it releases and guides actions –IRMs, Schemas are nice ways to start thinking about the computational structure of programming a robot

56 3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations56 Review Questions Think about the robots at the WTC. What are affordances of victims? color, motion, sound, heat Can schema theory represent behaviors in both biological and computational systems? yes A behavior schema is composed of at least the following: motor schema and a perceptual schema What is an example of behavior-specific knowledge? sequencing in a skill, alternative PS or MS


Download ppt "3 Introduction to AI Robotics (MIT Press)Chapter 3: Biological Foundations1 Biological Foundations of the Reactive Paradigm Review Why? -comp. theory IRM."

Similar presentations


Ads by Google