Browse wiki

In this paper, we consider the problem of active sensing using mobile sensor nodes that are jointly estimating the state of a dynamic target as a sensor network. We propose a gradient search-based decentralized algorithm that demonstrates the benefits of distributed sensing. We then examine the task of tracking multiple targets. We use a greedy algorithm for associating sensors with targets along with the gradientsearch strategy for sensor vehicle motion planning. Simulation results show that these simple decentralized strategies perform quite well and the sensor nodes exhibit interesting cooperative behavior.