Research On Human Localization Algorithm Based On WSN For Indoor Environments

For the particularity of indoor environment, such as space limited, more walls and barriers, signal transmission decaying severely, multi-path fading seriously, performance of the WSN localization algorithm decreases significantly. Compared with the general localization algorithms, the research on indoor localization algorithms technology seems more difficult, with its own distinct characteristics.Firstly, indoor people localization algorithm based on RSSI is studied in this paper. Considering the large errors of indoor RSSI localization, the author put forward a solution that selecting the best beacon nodes depending on unreliable beacon node tables previously established, then averaging the measurement of signal, updating the environmental parameters factor with beacon node value from real-time measurement, finally locating the objective node with maximum likelihood estimation (MLE). Simulation experiment results show that better precision of localization is got for this algorithm.Secondly, TDOA localization algorithm of indoor environment is investigated. Variation relationship between ranging corners and elevation of cricket nodes is identified by actual measurement. Also, the impact of temperature on ranging error, which can be effectively reduced by adopting temperature sensor, is analyzed. MLE centroid algorithm and improved triangular algorithm are proposed. Simulation results have shown that this algorithm is superior in reducing error than others like MLE and triangular algorithm as the author known.Thirdly, multiple sound sources localization based on energy is explored. According to the problem that multi-source localization tending to trap into local optimal value in WSN, an advanced multi-source localization algorithm based on PSO is proposed. By establishing sound source energy model according to energy measurements from sound sources, parameters such as source energy and position can be estimated. Together with reasonably selecting parameters such as learning factor and inertia weight, the situation of trap into local optimal value can be avoid effectively. Simulation experiment results reveale that PSO can avoid local optimum, get better localization precision than other methods the article mentioned.