Exogenous fault detection and recovery solutions in swarm robotics

A robotic swarm needs to ensure continuous operation even in the event of a failure of one or more individual robots. If one robot breaks down, another robot can take steps to repair the failed robot, or take over the failed robot’s task. Even with a small number of faulty robots, the expected time to achieve the swarm task will be affected. Observing failure detection techniques requires an investigation of similar techniques in insects. The synchronisation approach of fireflies is an exogenous failure detection technique. This approach requires all the robots in the swarm to be initially synchronised together in order to announce a healthy status for each individual robot. Another exogenous failure detection approach is the Robot Internal Simulator. The concept of this approach is to have robots that are capable of detecting partial failures by possessing a copy of every other robot’s controller, which they then instantiate within an internal simulator on-board to be run for a short period of time to predict the future state of the other robots. The work in this research draws inspiration from both approaches, which both still have several issues when they are implemented in swarm robotics. The enhanced technique developed in this research will depend on the input and output values in the robot’s controller to diagnose other robots within the swarm during the entire swarm operation. In this research, communication plays an important part of the diagnosis procedure. While robots retain possession of their own controller values including their co-ordination, the receiver computes the distance between them based on the signal strength. A fault suspicion is generated if the computed distances do not match and an acknowledgement of the failure will be broadcast to the robotic swarm. This research explores the performance of the simulation experimental results. It has shown that failed robots are rapidly detected failures using the proposed technique. A mitigation procedure takes place after the faulty robot is shut down, either by pushing it away or allowing it to work as a communication bridge to operational robots.