WCSE

A Multi-Agent Scheme for Energy-Efficient Coverage Control in Wireless Sensor Networks

Abstract— The underlying aim of this paper is to maximize the energy efficiency in the coverage control scheme in a wireless sensor network (WSN) by selecting the minimal number of working nodes while still maintaining network coverage area. The proposed algorithm is based on a self-adaptive multi-agent system (MAS) coverage control scheme whereby sensor nodes learn to adjust their own coverage to achieve the network-wide coverage. This paper proposes a variation of an existing MAS scheme called the distributed value function (DVF) which differs from the original scheme in the use of cost function which is a function of redundant coverage area. Performance evaluation were compared with a guaranteed complete coverage method, i.e., the optimal geographical density control (OGDC) scheme, and a partial area coverage scheme, i.e., the Probing Environment and Adaptive Sleeping (PEAS) scheme. Results show that modified DVF can achieve a nearly full coverage with only 13-64% of active sensor nodes whereas the OGDC and PEAS required 14-68% and 16-76% of active sensor nodes, respectively for high to low node densities. Results suggests that the MAS coverage control scheme can achieve efficient coverage control, is self-adaptive and therefore suitable for coverage control applications in WSNs such as lighting control in smart offices.

Index Terms— coverage control, multi-agent systems, wireless sensor networks, reinforcement learning.

Akkachai Phuphanin, Wipawee Usaha
School of Telecommunication Engineering, Suranaree University of Technology, THAILAND

[Download]

Cite: Akkachai Phuphanin, Wipawee Usaha, "A Multi-Agent Scheme for Energy-Efficient Coverage Control in Wireless Sensor Networks," Proceedings of 2016 6th International Workshop on Computer Science and Engineering, pp. 464-469, Tokyo, 17-19 June, 2016.