🦾CBS-RRT*: Scalable Multi-Robot Planning in 3D Environments 

🎯 Objective 

Develop a hybrid Conflict-Based Search (CBS) + RRT* framework for multi-robot path planning in continuous 3D spaces, combining CBS’s optimal conflict resolution with RRT*’s adaptability to high-dimensional environments and enable coordinated drone–arm systems to navigate overlapping workspaces efficiently and collision-free. 

💡 Motivation

Coordinating multiple robots in shared 3D workspaces remains one of the hardest challenges in autonomous systems. Traditional planners either ensure optimality but struggle with scalability (like CBS) or handle continuous motion but lack global coordination (like RRT*).

This project was motivated by the need for a hybrid framework that merges discrete conflict resolution with continuous sampling-based planning, enabling drones and manipulators to operate collaboratively in overlapping environments. By combining CBS for inter-agent coordination and RRT* for smooth, kinodynamically feasible motion, the system achieves scalable, collision-free multi-robot planning suited for real-world applications such as warehouse automation, exploration, and cooperative manipulation.

⚙️ Approach 

📈 Key Findings

• Achieved scalable coordination for up to four UR5e arms and multiple drones in 3D.

• CBS-RRT* improved collision-free planning success by > 90% compared to individual RRT* runs.

• Hybrid A* + RRT* reduced drone planning time by 2–3 orders of magnitude.

• Maintained asymptotic optimality and improved path efficiency in overlapping joint spaces.

🌍 Impact

CBS-RRT* bridges discrete and continuous planning, enabling real-time multi-robot coordination for drones and manipulators in shared 3D spaces.
It lays the foundation for warehouse automation, aerial-ground collaboration, and space robotics, supporting scalable conflict-free planning in dynamic environments.