Path Planning Systems

From RoboJackets Wiki
Jump to: navigation, search

The "Path Planning Systems" groups is responsible for find an optimal path through the environment, given a model of the environment from the data (provided by Data Filtering Systems group). This group will be developing the high-level code to control Candii during the autonomous and GPS waypoint navigation challenges--essentially, Candii's AI. Path planning algorithms, maps, and learning are the main focus points of this group.

Path Planning Systems Architecture

People

Meetings

  • IGVC Planning Day
    • Tuesday, 6-7pm, Tin Building
  • Path Planning Meeting (with Dr. Vela)
    • Friday, 1pm, Van Leer (VL) E368
  • IGVC Work Day
    • Saturday, 2-6pm, Tin Building

TODO List

Task List

  • [ ] Path Planning
    • [x] Dijkstra's algorithm
    • [ ] Gradient traversal
      • [x] manhatten distance
      • [x] vector addition
      • [ ] splines
    • [ ] A* algorithm
    • [ ] D* algorithm
  • [ ] Cost Map
    • [x] Sweeping Algorithm
      • [x] implement manhattan distance
      • [x] implement geometric approximation
      • [x] add obstacles
      • [x] add variable start/goal locations
    • [ ] Obstacle Avoidance
      • [ ] using sweeping algorithm
      • [ ] using depth-limited sweep
  • [ ] Implementation
    • [ ] Matlab Prototype
    • [ ] C/C++ Product

Resources

Books

  • "Planning Algorithms" by Steven M. LaValle
  • "Probabilistic Robotics" by Sebastian Thurn, Wolfram Burgard, Dieter Fox
    • JP has a hard copy.

Articles