CBBA + GCS: New algorithm for multi-agent task planning in cluttered dynamic environments
Combining consensus-based allocation with convex optimization for safe, time-efficient trajectories in 3D+time.
Deep Dive
A method integrating Graphs of Convex Sets (GCS) for trajectory optimization with Consensus-Based Bundle Algorithm (CBBA) for distributed task allocation uses a 3D+time configuration space to handle dynamic tasks and obstacles. In simulated cluttered environments, agents avoid collisions and provide accurate task completion time estimates for static and dynamic tasks.
Key Points
- Integrates Graphs of Convex Sets (GCS) trajectory optimization with Consensus-Based Bundle Algorithm (CBBA) task allocation for multi-agent systems.
- Operates in a 3D+time configuration space to handle dynamic obstacles and moving tasks simultaneously.
- Provides collision-free trajectories and accurate time estimates for task completion, validated in simulated cluttered environments.
Why It Matters
Enables safe, efficient coordination of robot swarms in unpredictable environments—critical for autonomous logistics, search-and-rescue, and defense applications.