K. Madhava Krishna Henry Hexmoor
CSCE Dept., University of Arkansas, Fayetteville AR
This paper studies the effect of endowing the social parameter, autonomy, on a distributed sensor network. The sensor network is used for surveillance of multiple targets that crisscross a rectangular surveillance zone. Crossing targets are ascribed priorities by a priority ascription scheme based on fuzzy inference methods. The scheme fuses local and global priorities of a target through autonomy modeled as a parameter that take values in [ ] 0,1 . Sensors coordinate to allocate themselves to a particular target. Autonomy affects the manner in which sensors allocate themselves to a target. Lower autonomy biases a sensor to allocate itself to a target that need not be tracked or tracked for shorter duration. Higher autonomy biases a sensor to allocate itself to a target that needs to be tracked for longer duration. Test cases are presented regarding how changes in autonomy affect the tracking performance of the system. Conclusions are derived that suggest when high or low values of autonomy are beneficial to the system.