Distributed Estimation in Distributed-Sensor Networks
Automatic assessment of a dynamic situation (e.g., tracking of airplanes in air-traffic control) is an important problem. The assessment can be done either by a single central processor or by several distributed processors. The basic distributed estimation technique has been applied together with the probabilistic data association scheme to handle the data association problem when measurement origins are uncertain. The increasing demands of modern surveillance tasks have made a distributed sensor system necessary. A distributed sensor network consists of many sensors or processors that can pool their information to achieve a better overall estimate. For simplicity, a one-dimensional single-target tracking problem will be considered. Three target dynamic models will be simulated, one with constant velocity, a second with (nearly) constant acceleration, and the third with acceleration driven by a large process noise to model the transient state between constant-velocity and constant-acceleration dynamics. The centralized algorithm performs exactly the same as the distributed algorithm.