Fast Data Collection in Linear Duty-Cycled Wireless Sensor Networks

We study fast data collection in linear duty-cycled wireless sensor networks (WSNs). We first present a benchmark algorithm that can achieve optimality in a general case (i.e., in non-duty-cycled case). Then, based on the insights obtained in the general case, we propose an optimal algorithm and a distributed algorithm for the case when each sensor only works at one slot in a cycle in duty-cycled mode. It is proven that the two latter algorithms with cycle length equal to 3 have bounded performance gap to the benchmark algorithm in the non-duty-cycled case. Simulation results are used to demonstrate the effectiveness of the proposed algorithms.