When finding the moving target in the acoustic sensor network, a dynamic cluster is formed to triangulate the target's position. While the target is moving, the old cluster is revoked and a new cluster is formed to follow the target. An election strategy based on the probability is used to select the cluster head and thus improve the cluster formation probabilities. Given the random function distribution, the paper also presents how to design the random function. The time window which is associated with the data association process is a very important parameter. The paper also presents the domain of the time window. The simulation shows that choosing right random function and right value of the time window is very important to the tracking precision.
Zang CZ,Yu HB,Liang W,et al. Target tracking based on the dynamic cluster method in the acoustic sensor network[C]. 见:5th International Conference on Computer and Information Technology. Shanghai, China. September 21-23, 2005.