Skip to Main Content
Intensive research has focused on redundance reduction in wireless sensor networks among sensory data due to the spatial and temporal correlation embedded therein. In this paper, we propose a novel approach termed asynchronous sampling that complements existing study. The key idea of asynchronous sampling is to spread the sampling times of the sensor nodes over the time line instead of performing them in a synchronous manner. Compared with existing strategies, asynchronous sampling introduces another dimension for optimization, without additional computation or communication overhead on sensor nodes. Theoretically, we show that asynchronous sampling benefits sensor networks through increased entropy of the sensory data or reduced reconstruction distortion. Furthermore, we formulate the optimal asynchronous sampling problem for determining the time shifts among the nodes. A heuristic solution, termed O-ASYN, is presented that uses local optimum search to approximate the global optimal solution. Simulation results based on simulated data and real experimental data both demonstrate the entropy increases.