A distributed control scheme for triangular formations with Markovian disturbances

Abstract

This work details a distributed control system for triangular formation control where the control actions applied by the agents can be chosen from a set of values during run-time (as opposed to a single value defined in a standard feedback controller). Such a control law is advantageous as it allows the individual agents freedom in choosing their individual headings and motion speeds (from the specified set) during execution. This offers greater flexibility and robustness than traditional distributed feedback formation control laws. An alternative interpretation of the results presented is that the control law presented is capable of handling measurements that may be corrupted by an unknown finite state Markov chain (which can model noisy sensors and/or communication channels). A strong convergence result is established which permits global exponential convergence of the formation to the desired shape.