Cross-layer design of quantized-innovation-based target tracking in wireless sensor networks

In this paper, the problem of target tracking in wireless sensor networks (WSNs) using quantized innovations is investigated from the perspective of cross-layer design. Unlike previous works that design the tracking algorithm in a single layer (e.g. the application layer), we design the tracking algorithm incorporating the channel statistics in the application layer, and exporting to the link layer to schedule the sensors. The motivation for this study is twofold, incorporating the channel statistics to improve the tracking accuracy and scheduling the sensor nodes to reduce power consumption. Specifically, activated sensors first report their quantized innovations to a fusion center (FC) over noisy wireless channels that modeled as binary symmetric channels (BSCs). Using sequential importance resampling (SIR) particle filtering, the FC estimates the target state in terms of the noisy-corrupted quantized innovations from local sensors. The imperfection of the wireless channels between the sensors and the FC is incorporated in the tracking algorithm, for which the posterior Cramér-Rao lower bounds (CRLBs) are derived. Then, an energy-efficient sensor scheduling strategy is proposed based on an optimization problem that minimizing the derived posterior CRLBs with a constraint on the total energy consumption. Simulation results are presented to demonstrate that the proposed algorithm outperforms the traditional algorithm significantly with reduced energy consumption.