Robustness of cooperative behaviour model on N robot-based multi-robot systems: Application to mine emergency and disaster management

Developing an efficient model for real-life management has been a rapidly growing robotic research area. Environments such as underground tunnels are one of many harsh areas that still need exploration and exploitation by autonomous systems in the field robotics. In this paper, a robust cooperative framework is presented for pre emergency and disaster management, in other words, safety prevention measures in the underground terrain. The system is designed for n-robots to understand the emergency and disaster behaviours of one another and cooperate while avoiding collision. The framework logically establishes a QLACS model based on Ant Colony System (ACS) and QLearning (QL) techniques. To provide a robust way of achieving pre-emergency and disaster management in the mine, the scalable QLACS was tested with 2-robots, 3-robots and 4-robots. The performance evaluation result shows that the QLACS is reliably robust in communication and search costs, and also scalable to n-based MRS.