Implementation of Decentralized Multi-Robot Task Allocation (DC-MRTA)

Overview

Implemented the DC-MRTA paper's two-level framework for robotic task allocation using Reinforcement Learning. The framework coordinates decentralized multi-robot systems to autonomously distribute tasks while navigating complex environments. Decentralized collision avoidance is handled using Optimal Reciprocal Collision Avoidance (ORCA) to ensure collision-free paths for up to 10 robots completing tasks synchronously.

Key Highlights

• Implemented the DC-MRTA paper’s two-level framework for robotic task allocation using Reinforcement Learning
• Utilized ORCA (Optimal Reciprocal Collision Avoidance) for decentralized, collision-free robot navigation
• Simulated and validated the framework’s performance with up to 10 robots completing tasks autonomously

Resources