Constraint-Based Motion Planning for Multiple Agents - PowerPoint PPT Presentation

About This Presentation
Title:

Constraint-Based Motion Planning for Multiple Agents

Description:

Team rescue operations compromised senses. Game scenarios (e.g., sports) Military operations continuous line of sight. What's this all for, anyway? ... – PowerPoint PPT presentation

Number of Views:44
Avg rating:3.0/5.0
Slides: 11
Provided by: luv
Learn more at: http://www.cs.unc.edu
Category:

less

Transcript and Presenter's Notes

Title: Constraint-Based Motion Planning for Multiple Agents


1
Constraint-Based Motion Planning for Multiple
Agents
  • Luv Kohli
  • COMP290-058
  • September 29, 2003

2
Constraint-based?
  • Garber Lin formulated the motion planning
    problem as a dynamical system simulation
  • Each robot is a rigid body or a collection of
    rigid bodies influenced by constraint forces in
    the environment

3
Constraints
  • Hard constraints
  • Absolutely must be satisfied (e.g.
    non-penetration, articulated robot joint
    connectivity)
  • Soft constraints
  • Encourage objects to follow certain behaviors
    (e.g. moving towards a goal, obstacle avoidance)

4
Multiple agents
  • I would like to extend the constraint-based
    framework to study scenarios involving multiple
    interacting agents
  • Possible scenarios
  • Team rescue operations compromised senses
  • Game scenarios (e.g., sports)
  • Military operations continuous line of sight

5
Whats this all for, anyway?
  • If the constraints of a real multiple-agent
    system can be identified and modeled, then the
    feasibility of the goal can be studied
  • Virtual environments
  • Games

6
Tasks
  • Minimally I would like to get a constraint-based
    system working with multiple agents
  • The multiple agents will be acting either against
    each other or with one another towards some
    global goal, but influenced by local behavior

7
Behavior and Intelligence
  • It would be interesting to add higher levels of
    behavior and intelligence
  • Flocking-style algorithms
  • Agents that learn skills that can be applied to
    multiple scenarios

8
Simple example
Field of view
communication
9
Steps
  • Create a working constraint-based system
  • Create a constraint for line of sight, and get
    multiple agents to interact under the influence
    of this constraint
  • Determine how behaviors (e.g. information
    gathering) relate to constraint model
  • Implement new constraints

10
References
  • Garber, M. and Lin, M. Constraint-Based Motion
    Planning using Voronoi Diagrams. Proc. Fifth
    International Workshop on Algorithmic Foundations
    of Robotics (WAFR), 2002.
  • Garber, M. and Lin, M. Constraint-Based Motion
    Planning for Virtual Prototyping. Proc. ACM
    Symposium on Solid Modeling and Applications,
    2002.
  • Reynolds, C. W.. Flocks, Herds, and Schools A
    Distributed Behavioral Model. Computer Graphics,
    21(4) 25-34, 1987.
  • Goldenstein, S., Large, E., and Metaxas, D.
    Dynamic Autonomous Agents Game Applications.
    Computer Animation, 1998.
Write a Comment
User Comments (0)
About PowerShow.com