Collective and Coope= rative Robotics

 

Instructor: Dr. Lynne E. Parker, Oak Ridge N= ational Laboratory, USA

Language of instruction: English

Course overview:

This course presents the key principles and desi= gn approaches in the development and implementation of autonomous multipl= e robot systems. The course begins by reviewing the behavior-based approa= ch to robot control, upon which most multi-robot systems are based. The = behavior-based approach is compared and contrasted with other types of ro= bot control approaches, including planner-based control, reactive control= , and hybrid deliberative/reactive architectures. We then review the bac= kground of the field of collective and cooperative robotics and the motiv= ation for building teams of robots. We examine the origins of cooperatio= n and taxonomies of multiple robot systems. Issues in group architecture= s and representative examples of multi-robot control strategies are studi= ed. Geometric multi-robot motion problems are examined, as well as issue= s of multi-robot learning. A variety of applications of multi-robot team= s are described and illustrated. Throughout the course, many references, = descriptions, and videos of actual implemented multi-robot teams will be = shown to illustrate the current state of the art in collective and cooper= ative robotics.

 

1. Introduction to behavior-based robotics

  • Cybernetics
  • Artificial intelligence planners
  • Early robots

2. Robot control architectures

  • Subsumption
  • Motor schemas
  • Internalized plans
  • Planner-based
  • Reactive
  • Hybrid

3. Overview of collective and cooperative robotics<= /P>

  • Motivation

  • Taxonomy of multi-robot teams
  • Collective vs. cooperative

4. Origin of cooperation

  • Game-theoretic approaches
  • Economic approaches
  • Animal societies

5. Group architectures

  • Centralization/decentralization

  • Differentiation

  • Communication structures

  • Agent modeling

  • Representative architectures

 

6. Geometric problems

  • Multi-robot path planning
  • Moving in formation
  • Pattern generation
  • Multi-target tracking

7. Multi-robot learning

  • Reinforcement learning
  • L-ALLIANCE
  • Tropism-based cognition
  • Imitation
  • Evolution of communication

8. Applications

  • Flocking
  • Herding
  • Box pushing
  • Foraging
  • Cooperative manipulation
  • Traffic control

 

Prerequisites: General background in artific= ial intelligence. This course is suitable for advanced undergraduates, gr= aduate students, scientists, engineers, and professionals.