All Projects
Academic
CANADARM Motion Planning
Probabilistic Roadmap-based motion planning for the ISS Canadarm2 robotic arm, navigating obstacles in constrained 2D workspace.
Motion Planning Probabilistic Roadmap Robotics Python
View Source
Overview
Developed motion planning algorithms for a simplified 2D version of the International Space Station’s Canadarm2 robotic arm. Given initial and goal configurations, the system finds valid paths through obstacle-filled environments while satisfying multiple kinematic constraints.
Constraints Handled
- Primitive step limits: joint movement restricted to 0.001 units per step
- Obstacle avoidance for all arm segments
- Self-collision prevention between arm links
- Workspace boundary enforcement within [0,1] x [0,1] space
- Joint angle constraints (15-165 degrees)
- Segment length bounds
Approach
- Probabilistic Roadmap (PRM): Sampled random configurations in C-space, connected valid neighbors, and searched the resulting graph
- Collision Detection: Implemented efficient line-segment intersection tests for arm-obstacle and arm-self collision checking
- Graph Search: Applied A* search on the PRM graph with configuration-space distance heuristics
Key Learnings
- High-dimensional configuration space planning for multi-joint systems
- Sampling-based motion planning algorithms and their probabilistic completeness guarantees
- Efficient collision detection in 2D environments
- Trade-offs between roadmap density and planning speed