Distributed Water Pollution Source Localization with Mobile UV-Visible Spectrometer Probes in Wireless Sensor Networks

• Main Authors: Junjie Ma, Fansheng Meng, Yuexi Zhou, Yeyao Wang, Ping Shi
• Format: Article
• Language: English
• Published: MDPI AG 2018-02-01
• Series: Sensors
• Subjects: pollution source localization; wireless sensor networks; mobile nodes; UV-visible spectroscopy; water quality multi-parameter; distributed algorithm; particle swarm optimization
• Online Access: http://www.mdpi.com/1424-8220/18/2/606

Pollution accidents that occur in surface waters, especially in drinking water source areas, greatly threaten the urban water supply system. During water pollution source localization, there are complicated pollutant spreading conditions and pollutant concentrations vary in a wide range. This paper provides a scalable total solution, investigating a distributed localization method in wireless sensor networks equipped with mobile ultraviolet-visible (UV-visible) spectrometer probes. A wireless sensor network is defined for water quality monitoring, where unmanned surface vehicles and buoys serve as mobile and stationary nodes, respectively. Both types of nodes carry UV-visible spectrometer probes to acquire in-situ multiple water quality parameter measurements, in which a self-adaptive optical path mechanism is designed to flexibly adjust the measurement range. A novel distributed algorithm, called Dual-PSO, is proposed to search for the water pollution source, where one particle swarm optimization (PSO) procedure computes the water quality multi-parameter measurements on each node, utilizing UV-visible absorption spectra, and another one finds the global solution of the pollution source position, regarding mobile nodes as particles. Besides, this algorithm uses entropy to dynamically recognize the most sensitive parameter during searching. Experimental results demonstrate that online multi-parameter monitoring of a drinking water source area with a wide dynamic range is achieved by this wireless sensor network and water pollution sources are localized efficiently with low-cost mobile node paths.