Fayette Shaw, Agreement on Stochastic Multi-Robot Systems with Communication Failures

Agreement algorithms allow individual agents in a population to estimate a global quality by sharing information in a distributed fashion. A common example is computing the global mean of a sensor measurement from each agent. We study agreement algorithms in stochastic systems; multi-agent systems in which interaction patterns are random and memoryless. We present a practical agreement algorithm, input-based consensus (IBC), that produces bounded error and recovery in the face of significant communications failures. We compare our algorithm to linear average consensus, which produces an exact result under ideal conditions, but is not robust to message loss. For each algorithm, we measure performance with respect to a varying percentage of dropped messages. The algorithms are analyzed numerically, simulated with a stochastic agent simulation, and demonstrated experimentally on a testbed of 20 robots. In all cases, the IBC algorithm produced reasonable estimates, even when tested with up to 90% message loss.