Fault Consensus in a Network of Unmanned Vehicles

Abstract

This paper investigates development of a distributed fault consensus algorithm for a network of unmanned vehicles. In our proposed approach, it is assumed that the dynamics of each vehicle is governed by a stochastic hybrid system with different operational modes corresponding to the occurrence of different faults in the vehicles. By utilizing the interactive multiple model (IMM) algorithm, each vehicle generates the probabilities of operational mode of the vehicles that it has measurements with respect to them. A new consensus algorithm for a network of multi-agent systems with limited local input signal is proposed for sharing the information among the vehicles in the network in order to reach the consensus on the current operational mode of each vehicle. In contrast to the previous work in the literature where consensus is achieved when all the agents have certain consensus signal (here represented as a health status) to agree upon, the methodology developed in this work does not required that all agents have this health status signal. Our proposed fault consensus algorithm is applied to the formation flight of satellites and simulation results demonstrate the effectiveness of our proposed methodology.