Research On Indoor Localization Method Based On Fusion Of Wi-Fi/PDR

In recent years,with the rapid development of mobile networks and smart phones,the demand for Indoor Positioning is becoming more intense.Considering the cost,stability and other factors,the relatively mature positioning technology currently include Wi-Fi Fingerprint Positioning and PDR(Pedestrian Dead Reckoning,PDR)Positioning.Wi-Fi Fingerprint Positioning base on the wireless signal receiving intensity RSSI(Received Signal Strength Indication,RSSI).Although there is no accumulated error,the complex indoor environment causes the RSSI to be volatile and seriously affects the positioning accuracy.The PDR Positioning relies on the real-time detection data of mobile sensor,which has a high positioning accuracy in a short time.It is easy to cause accumulated errors in the positioning process.Each of these two positioning technologies has its advantages and disadvantages,but it is difficult for a single positioning technology to achieve precise positioning.Thus,Multi-sensor information fusion positioning has become one of the most important and essential issues in Indoor Positioning.For Wi-Fi fingerprint positioning system:(1)In order to overcome the problem of RSSI volatility,we preprocessed the RSSI samples collected in off-line phase,and proposed a RSF(RSSI-Sample-filter,RSF)algorithm.In addition,we rebuilt the fingerprint database based on RSF algorithm,avoid the impact of RSSI volatility.(2)In the on-line phase,we proposed a RWF(Rank-based Wi-Fi Fingerprinting,RWF)algorithm based on the relative stability of RSSI,which fully uses the more valuable information of RSSI and significantly improves the efficiency and accuracy of the positioning system.For PDR positioning system:(1)In the traditional step counting method,a single threshold is used to judge the peaks.In this paper,we proposed a FP(Filter peak,FP)algorithm based on acceleration peaks filtering.This method can effectively reduce the error of step counting.(2)The step length estimation has a crucial influence on the PDR positioning result.We proposed a novel ASLE(Adaptive Step Length Estimation,ASLE) algorithm.The algorithm uses Wi-Fi fingerprint positioning technology and PDR positioning technology comprehensively,generates the linear relation model of step frequency and step length by least square.Experiments show that ASLE can effectively reduce the error caused by the walking randomness.In this paper,we optimized the problems of Wi-Fi fingerprinting and PDR positioning respectively,proposed a Wi-Fi/PDR fusion algorithm bases on UKF(Unscented Kalman Filter,UKF)to optimize the result of indoor positioning system.The performance of the fusion algorithm is significantly improved.