Adaptive local quantizer design for tracking in a wireless sensor network

We investigate the problem of tracking a target using a wireless sensor network, where quantized sensor measurements are utilized because of inherent communication and energy constraints. Due to the severe nonlinearity of the measurement model, we resort to sequential Monte Carlo methods for tracking, i.e., particle filters. The tracking performance is a function of local sensor quantizer thresholds. We propose a new dynamic adaptive local quantizer design approach along with some practical implementation considerations. Simulation results are presented to demonstrate the significant performance improvement achieved by our quantizer design technique.