An Iterative Algorithm for Autonomous Tasking in Sensor Networks

Modern environments are saturated with sensors. These sensors may be active or passive, stationary or mobile, based on land, sea, air or space. They may differ in size, complexity, modes of operation, power, costs and a number of other features. The challenge in a sensor-rich environment is to coordinate the sensing of all the sensors in order to obtain the best total group performance. The complexity of the problem suggests that traditional methods of optimization will be prohibitively costly in terms of computation and communication. Additionally, there is some ambiguity over how to aggregate individual sensor utilities into a "best total group performance." Under these circumstances, game theoretic methods can be employed to both limit the computation and communication necessary for group coordination and avoid the problem of aggregating utilities. Specifically, an economic paradigm is adopted from which an efficient method for sensor coordination can be derived. The resultant coordination algorithm is shown to have bounded Pareto suboptimality, demonstrating the tradeoff between inter-sensor communication and utility optimization