Data Collection Maximization in Renewable Sensor Networks via Time-Slot Scheduling

TL;DR: This paper first formulate a novel data collection maximization problem by adopting multi-rate data transmissions and performing transmission time slot scheduling, and shows that the problem is NP-hard, and devise an offline algorithm with a provable approximation ratio for the problem by exploiting the combinatorial property of the problem.

Abstract: In this paper we study data collection in an energy renewable sensor network for scenarios such as traffic monitoring on busy highways, where sensors are deployed along a predefined path (the highway) and a mobile sink travels along the path to collect data from one-hop sensors periodically. As sensors are powered by renewable energy sources, time-varying characteristics of ambient energy sources poses great challenges in the design of efficient routing protocols for data collection in such networks. In this paper we first formulate a novel data collection maximization problem by adopting multi-rate data transmissions and performing transmission time slot scheduling, and show that the problem is NP-hard. We then devise an offline algorithm with a provable approximation ratio for the problem by exploiting the combinatorial property of the problem, assuming that the harvested energy at each node is given and link communications in the network are reliable. We also extend the proposed algorithm by minor modifications to a general case of the problem where the harvested energy at each sensor is not known in advance and link communications are not reliable. We thirdly develop a fast, scalable online distributed algorithm for the problem in realistic sensor networks in which neither the global knowledge of the network topology nor sensor profiles such as sensor locations and their harvested energy profiles is given. Furthermore, we also consider a special case of the problem where each node has only a fixed transmission power, for which we propose an exact solution to the problem. We finally conduct extensive experiments by simulations to evaluate the performance of the proposed algorithms. Experimental results demonstrate that the proposed algorithms are efficient and the solutions obtained are fractional of the optimum.