Autonomous Multi Robot Disinfection Mission

This project demonstrates the development and implementation of decentralized controllers for a team of autonomous robots tasked with disinfecting a high-traffic clinical environment. Conducted as part of the ENME808T Network Control Systems course, the system was implemented using MATLAB simulations and validated on the Robotarium platform.

In the context of a simulated pandemic, a team of six robots collaboratively performs disinfection tasks while adhering to strict operational constraints:

• Decentralized Operation: No direct communication between robots; decisions are made using localized information.

• Collision Avoidance: Energy-based control laws minimize collision risks using proximity sensors.

• Connectivity: Maintenance of a $\Delta$-disk communication graph to ensure network stability.

• Sequential Mission Goals: Waypoint navigation, room disinfection (grid-based coverage), and refueling formation.

Key Algorithms & System Design

• Decentralized Controllers: Robots operate independently to ensure robust and scalable operations.

• Energy-Based Control: Gradient-based control laws utilizing energy functions to maintain formation while avoiding collisions.

• Formation Control: Adaptive strategies allowing robots to navigate narrow spaces and efficiently cover target areas.

• Consensus Protocols: Used for leader-following waypoint navigation and dynamic task assignment.

Implementation Results

• Simulation: MATLAB simulations achieved full mission completion with 0% collision rate.

• Hardware Validation: Robotarium experiments successfully completed 3 out of 4 sub-missions. The final sub-mission (room disinfection) demonstrated minor oscillations but maintained safety with zero collisions.

• Efficiency: Trajectory efficiency improved by 25% through the use of optimized formation graphs.

Technical Stack

• Language: MATLAB (Mapping Toolbox required)

• Platform: Robotarium (Georgia Tech)