Fast and accurate cooperative localization in wireless sensor networks

摘要

Cooperative localization capability is a highly desirable characteristic of wireless sensor networks. It has attracted considerable research attention in academia and industry. The sum-product algorithm over a wireless sensor network (SPAWN) is a powerful method to cooperatively estimate the positions of many sensors (agents) using knowledge of the absolute positions of a few sensors (anchors). Drawbacks of the SPAWN, however, are its high computational complexity and communication load. In this paper we address the complexity issue, reformulate it as convolution problem and utilize the fast Fourier transform (FFT), culminating in a fast and accurate localization algorithm, which we named SPAWN-FFT. Our simulation results show SPAWN-FFT's superiority over SPAWN regarding the computational effort, while maintaining its full flexibility and localization performance.