Description
We approach this problem by studying some models and simulations of sensor networks and by building testbeds and running experiments. First, we simulate the use of a network connectivity parameter in an optimal path planning controller for pursuit of a target traversing through a sensor network. Then, we look at an example of bounding the performance of a sensor network control application -- finding a probabilistic barrier solution to a single pursuer/single evader pursuit-evasion game. From the implementation side, we discuss the design of an indoor and and outdoor sensor network testbed for multiple-target tracking, a piece of the estimator in the control loop for multiple-evader pursuit-evasion games. Finally, we discuss some preliminary experimental results of multiple-target tracking on these sensor network testbeds.