1. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm1 Part 1: Overview & Managerial Architectures Part 2: State Hierarchy Architectures

2. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm2 Objectives Describe the hybrid paradigm in terms of 1) SPA and 2) sensing organization Given a list of responsibilities, be able to say whether it belongs in the deliberative layer or in the reactive layer List the five basic components of a Hybrid architecture: sequencer agent, resource manager, cartographer, mission planner, performance monitoring and problem solving agent. Be able to describe the difference between managerial, state hierarchy, and model-oriented styles of Hybrid architectures. Be able to describe the use of state to define behaviors and deliberative responsibilities in state hierarchy styles of Hybrid architectures

3. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm3 Motivating Example for Deliberation: USAR Worker places robot at entrance to unstable building, loads in the floor plan, contextual knowledge and tells robot to look for survivors efficiently and map out safe routes for workers to pass through contextual knowledge includes probability of where people are more likely to be What can a reactive architecture do? What can’t it do? Path planning, handling detours due to blockage, map making, learn from past rescues

4. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm4 Organization: Plan, Sense-Act

5. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm5 Sensing Organization Deliberative functions *Can “eavesdrop” *Can have their own Sensors *Have output which Looks like a sensor Output to a behavior (virtual sensor)

6. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm6 Deliberation v. Reaction as a function of TIME Past, Present, Future Reactive –exists in the PRESENT (will a bit of duration) Deliberative –can reason about the PAST –can project into the FUTURE

7. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm7 Architectures:Key Questions How does the architecture distinguish between reaction and deliberation? How does it organize responsibilities in the deliberative portion? How does overall behavior emerge?

8. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm8 Architectures: Common Functionality Mission planner – interacts with human, puts commands into robot terms, constructs mission Cartographer – creates, stores, maintains map or world model, knowledge; Sequencer agent – generates set of behaviors to use to accomplish task Behavioral/resource manager – allocates resources (schemas, etc.) to behaviors Performance monitor/problem solving agent (fairly rare)

9. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm9 Architectures: 3 Styles Managerial (division of responsibilities looks like in business) –AuRA, SFX State Hierarchies (strictly by time scope) –3T Model-Oriented (models serve as virtual sensors) –Saphira, TCA

10. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm10 Mgr Architecture 1: AuRA (Autonomous Robot Arch.) Ron Arkin, Georgia Institute of Technology

11. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm11 AuRA Architectural Layout reactive deliberative Interfaces with human Computes path, breaks into subtasks Generates behaviors for subtasks Uses potential fields Changes gains of behaviors’ potential fields

12. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm12 Architectures: Common Functionality Mission planner Cartographer Sequencer agent Behavioral manager Performance monitor/problem solving agent

13. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm13 AuRA Architectural Layout Cartographer Sequencer Mission Planner Behavioral manager (mgr+schemas) Performance Monitoring Emergent behavior

14. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm14 HOW WOULD THIS DO USAR TASK?

15. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm15 Motivating Example for Deliberation: USAR Worker places robot at entrance to unstable building, loads in the floor plan, contextual knowledge and tells robot to look for survivors efficiently and map out safe routes for workers to pass through contextual knowledge includes probability of where people are more likely to be What can a reactive architecture do? What can’t it do? Path planning, handling detours due to blockage, map making, learn from past rescues

16. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm16 Motivating Example for Deliberation: USAR Worker places robot at entrance to unstable building, loads in the floor plan, contextual knowledge and tells robot to look for survivors efficiently and map out safe routes for workers to pass through contextual knowledge includes probability of where people are more likely to be

17. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm17 Example USAR (overlay) Cartographer accepts the map Navigator uses path planning algorithm to visit nodes in order of likelihood of survivors Pilot determines the list of behaviors, Motor Schema Manager instantiates them (MS & PS) and waits for termination Homeostatic might notice that robot is running out of power, so opportunistically picks up low probability room on way back to home

18. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm18 Mgr. Architecture 2: SFX (Sensor Fusion Effects) Focus on sensing Biomimetic organization deliberative layer consists of managerial agents reactive layer has tactical behaviors

19. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm19 SFX (Sensor Fusion Effects)

20. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm20 SFX (Sensor Fusion Effects) Behaviors (using direct perception, fusion) Sense Muscle Actuators Deliberative Layer Managers Sense Sensor Sense Receptive Field Choice of behaviors, resource allocation, motivation, context Focus of attention, recalibration Sensor Whiteboard Behavioral Whiteboard Deliberative Layer Reactive Layer Parameters to behaviors, sensor failures, task progress actions Superior Colliculus-like functions Cerebral Cortex-like functions Cartographer (model/map making) Recognition perception

21. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm21 Cartographer SFX Implementation Sensor Mgr Task Planner Interface Mgr Behaviors Sensors Acuators Effector Mgr Lisp C++ Behaviors Use, Fuse Information directs sensing Interacts with human, specifies constraints Managers are independent agents; negotiate; use AI planning, scheduling, problem solving to allaocate effector and sensor resources to accomplish task monitors task performanc e and whether habitat has changed Map making, path planning Layered into strategic and tactical behaviors

22. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm22 Ability to Substitute Components

23. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm23 Motivating Example for Deliberation: USAR Worker places robot at entrance to unstable building, loads in the floor plan, contextual knowledge and tells robot to look for survivors efficiently and map out safe routes for workers to pass through contextual knowledge includes probability of where people are more likely to be

24. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm24 Example USAR (overlay) Cartographer accepts the map Task Planner agent asks for path, requests behaviors, passes to managerial layer Sensing and Effector Mgrs negotiate allocation Behaviors run until terminate or encounter exception (either preset condition by mgrs or through monitoring) Mgrs can see “below” but not above--cannot relax constraint of Planner/Boss

25. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm25 Tactical Behaviors Tactical behaviors subsume strategic behaviors; interaction of stratregic and tactical behaviors create emergent behavior Task typically has 1 strategic, several tactical behaviors

26. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm26 Summary: Managerial Architectures How does the architecture distinguish between reaction and deliberation? –Deliberation: global knowledge or world models, projection forward or backward in time –Reaction: behaviors which have some past/persistence of perception and external state How does it organize responsibilities in the deliberative portion? –hierarchy of managerial responsibility, managers may be peer software agents How does overall behavior emerge? –From interactions of a set of behaviors dynamically instantiated and modified by the deliberative layer –assemblages of behaviors

27. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm27 Chapter 7: Hybrid Deliberative/Reactive Paradigm Part 1: Overview & Managerial Architectures Part 2: State Hierarchy & Model-Based Architectures

28. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm28 Architectures: 3 Styles Managerial (division of responsibilities looks like in business) –AuRA, SFX State Hierarchies (strictly by time scope or “state”) –3T Model-Oriented (models serve as virtual sensors) –Saphira, TCA

29. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm29 State Hierarchy Architectures How does the architecture distinguish between reaction and deliberation? –Deliberation: requires PAST or FUTURE knowledge –Reaction: behaviors are purely reflexive and have only local, behavior specific; require only PRESENT How does it organize responsibilities in the deliberative portion? –By internal temporal state PRESENT (controller) PAST (sequencer) FUTURE (planner) –By speed of execution How does overall behavior emerge? –From generation and monitoring of a sequence of behaviors –assemblages of behaviors called skills –subsumption

30. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm30 3T Architecture Used extensively at NASA Merging of subsumption variation (Gat, Bonasso), RAPs (Firby), and vision (Kortenkamp) Has 3 layers –reactive –deliberative –in-between (reactive planning) Arranges by time Arranges by execution rate –ex. vision in deliberation Dave Kortenkamp, TRAC Labs (NASA JSC) Used for planetary rovers, robot assistants for astronauts, underwater vehicles Reactive action packages – a reactive planner

31. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm31 Uses RAPS to select behaviors (skills), develop network of tasks Skills have events that indicate than an action has had the correct effect present Past and present Uses past and present to predict future

32. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm32 cartographer Mission planner sequencer Behavior mgr Performance monitor Emergent behavior

33. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm33 Model-Oriented Architectures How does the architecture distinguish between reaction and deliberation? –Deliberation: anything relating a behavior to a goal or objective –Reaction: behaviors are “small control units” operating in present, but may use global knowledge as if it were a sensor (virtual sensor) How does it organize responsibilities in the deliberative portion? –Behavioral component –Model of the world and state of the robot –throwback to Hierarchical Paradigm with global world model but virtual sensors –Deliberative functions How does overall behavior emerge? –From generation and monitoring of a sequence of behaviors –voting or fuzzy logic for combination

34. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm34 Saphira Architecture Developed at SRI by Konolige, Myers, Saffioti Comes with Pioneer robots Behaviors produce fuzzy outputs, fuzzy logic combines them Has a global rep called a Local Perceptual Structure to filter noise Instead of RAPs, uses PRS-Lite Procedural Reasoning System

35. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm35 Saphira and LPS Local Perceptual Space (the world model)

36. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm36 Sequencer agent, Mission Planning, Performance mon. Cartographer Behavior mgr Emergent behavior Fuses behaviors with fuzzy logic

37. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm37 Symbol-Grounding Problem Computers (and AI) reasons using symbols –Ex. “room”, “box,” “corner,” “door” Robots perceive raw data How to convert sensor readings to these labels?

38. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm38 Spatial World Knowledge What do you see? How could a robot reliably extract the same labels?

39. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm39 Types of Knowledge (Arkin) Spatial World knowledge Object knowledge Perceptual knowledge Behavioral knowledge Ego knowledge Intentional knowledge

40. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm40 Task Control Architecture Developed by Reid Simmons Used extensively by CMU Field Robotics Projects –NASA’s Nomad, Ambler, Dante Closest to Hierarchical in philosophy, but strong reactive theme showing up in implementation

41. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm41 TCA Task Control Architecture – operating system flavor low-level tasks are like behaviors Evidence grids (Ch 11) – sensor info percolates up through the global model

42. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm42 Evaluation of Hybrids Support of Modularity: high Niche targetability: high (ex. Lower levels of AuRA, SFX, 2 1/2 T is just reactive) Robustness: SFX and 3T explicitly monitor Think in closed world, act in open world

43. 7 Introduction to AI Robotics (MIT Press), copyright Robin Murphy 2000 Chapter 7: Hybrid Deliberative/Reactive Paradigm43 Hybrid Summary P,S-A, deliberation uses global world models, reactive uses behavior-specific or virtual sensors Architectures generally have modules for mission planner, sequencer, behavioral mgr, cartographer, and performance monitoring Deliberative component is often divided into sub-layers (sequencer/mission planner or managers/mission planner) Reactive component tends to use assemblages of behaviors