Hazard control algorithms for heterogenous multi-agent cloud-enabled robotic network

Abstract

Robots exploring unknown terrains autonomously always face the threat of getting isolated or lost from the rest. Geo-coordination technologies such as GPS cannot be used in extra terrestrial situations. Therefore, an inbuilt mechanism is needed in the robot that helps to track the robot's position and prevents it from getting off the network's coverage. We define hazard-control as the task of executing a series of decisions for recovering lost robots and bringing them back to the robotic swarm network. For a swarm of robots, this task involves communicating with other robots in the system to make sure that all the robots are within the range of the multi-hop network formed by the robots. If one robot goes away of the robotic network, hazard-control algorithms help the robots to recover the lost robot back in the network. We propose three robot-centric hazard control algorithms that help to get the lost robot in a robotic swarm back in the network: (i) Random Motion algorithm, (ii) Variable Spiral algorithm, and (iii) Motion Reversal algorithm. We study the performance of the proposed algorithms, compared their efficiency and identify the scenarios where each algorithm is most suited.