Research Of Gas Source Localization Based On Gas Diffusion In WSN

Because of its characteristics such as low cost of networking, nodes being deployed randomly, wide covering field and long period of work ability, wireless sensor networks(WSN) has a wide variety of applications including military and civilian fields in recent years. The research purpose of gas source localization based on WSN is to find an effective and accurate localization scheme, then to give out the early fore-warning about the monitoring field. The research results could be applied to many different fields such as fire early monitoring, poisonous gas spreading, flammable and explosive gas leakage and gaseous biochemical agent propagating.Focusing on the problem of gas source localization in wireless sensor networks, this paper mainly includes these research works as follows:Firstly,this paper discussed respectively the development and characteristic of WSN, sensor nodes composition, WSN architecture and its main applications; Then concluded the state and significance of gas source localization based on WSN, and analyzed the existing difficult problems.Secondly, the principle and classification of localization technology in wireless sensor network were stated. According to the standard classification of range-based and range-free technology, this paper discussed and analyzed some classical localization algorithms. Meanwhile, focusing on the characteristic of gas source localization, this paper introduced some common models of gas diffusion, and then derived the universal function of gas diffusion model under the different wind direction condition (including no wind).After that, based on the attenuation model of a gas source in the wind field, this thesis constructed a localization model, and then combined the intelligent optimization algorithm and distributed feature of sensor nodes, it presented two centralized gas source localization methods——gas source localization based on force-directed QPSO algorithm and based on weighted differential distance approach in the skew wind field. The simulation results demonstrate them feasibility under different situations.Finally, in view of some problems of centralized computation, we applied the idea of distributed calculation and analyzed its characteristics. Taking the nonconvex of objective function for localization into account, this paper proposed a distributed adaptive deflection projected subgradient method (DADPSM). Still based on the attenuation model of a gas source in the wind field, this method was able to process the distributed information from the sensors using DADPSM algorithm, replaced the original gradient direction with deflection subgradient direction, and utilized the deflected subgradient projection hyperplanes as the searching areas in the process of relaxed projection, so as to obtain the gas source position. The simulation within this paper illustrated its validity on localization accuracy and was superior to others.