Distributed communication/navigation robot vehicle network

This paper describes algorithms for forming a communication/navigation network of robotic vehicles inside a building. All ultrasound ranging system on each vehicle is used to determine the distance between vehicles in the network. This distance information is transformed into x,y position information using a steepest descent algorithm that minimizes the error between measured distances and estimated distances (determined from estimated x,y position). This same optimization function is also used to guide the vehicles into a desired formation specified by the desired distance between vehicles. When used for guidance, the gradient vector is used as the commanded input to each vehicle. Most importantly, the algorithm does not require knowledge of the distance between every vehicle node, only those closest to it. Therefore, the algorithm can be distributed amongst the vehicles, providing a global control for the entire system using only local control laws implemented on each vehicle.