Research On WiFi/PDR Fusion Indoor Positioning Technology Based On Particle Filter

With the rapid development of mobile communication technology and the expansion of intelligent terminal application Services,people's demand for Location Based Services(LBS)is increasing day by day,and the requirement for providing accurate Location Services in indoor environment is becoming more and more intense.WiFi as network infrastructure has been widely spread in a variety of indoor environment,WiFi positioning with high precision,strong anti-interference,low cost advantage,in addition,smart phone terminal common embedded inertial sensor element,make Pedestrian Dead Reckoning(PDR)based on inertial sensors plays an important role in indoor navigation and positioning.Therefore,the fusion location technology based on WiFi and PDR has become a research hotspot in recent years.In this thesis,the following studies are carried out on indoor WIFI and PDR fusion positioning technology.Firstly,a WiFi location method based on K-means clustering and WKNN fingerprint matching was proposed to solve the problems of abnormal fingerprint points in the offline collection process of indoor WiFi location fingerprint database and large computation amount in the online location fingerprint database matching.In the offline stage,the K-means clustering algorithm was used to divide the whole fingerprint database and determine the clustering center.In the online stage,the similarity of the points to be measured was calculated according to the signal strength RSSI and the clustering center,and the points to be measured were classified into the cluster where the clustering center was located according to the similarity,and then in the cluster,the adjacent points of the location fingerprint were screened by WKNN algorithm and the location was realized.The simulation results show that the WiFi positioning method based on K-means clustering and WKNN fingerprint matching has better positioning accuracy than the traditional WKNN algorithm under the Gaussian positioning noise from 1 to 5d Bm.Under the condition of Gaussian noise of 3d Bm,the positioning accuracy can reach less than 1.2m,and the computational complexity of positioning is reduced by 50% compared with the traditional WKNN algorithm.Secondly,in view of the problem of false peak misjudgment in PDR positioning,the traditional step frequency detection algorithm uses fixed threshold to process real-time step frequency data,a double adaptive threshold interval peak and trough step frequency detection method is proposed.The experimental results show that the accuracy of this algorithm can reach 96% in the speed change and acceleration,and it can adapt to the change of pedestrian's gait.An adaptive step size estimation algorithm based on Kalman filter is proposed to solve the problem of insufficient accuracy of traditional nonlinear step size models.Experimental data show that the accuracy of the proposed algorithm is about 10% higher than that of the traditional nonlinear step size model,and it is more suitable for step size estimation.The heading Angle is estimated by Kalman filter algorithm,and the average value of direction Angle of each step is calculated by the magnetometer low-pass filter and gyroscope integration,and then the turn is judged.The experimental data show that the improved heading Angle algorithm has higher accuracy and better stability.Finally,the indoor fusion location algorithm based on particle filter is studied to solve the problem of unstable WiFi location results and the accumulated error of sensors in PDR.Particle swarm was used to establish the system state equation,PDR location was used as the state input parameter,WiFi location was used as the observation vector to update the weight of particles,and then resampling the particles according to the weight,and the particle state was predicted by the system state equation for positioning.The experimental results show that the algorithm improves the robustness of the positioning system,eliminates the influence of accumulated errors of sensors on the PDR positioning results,and the positioning accuracy reaches 1.6m,which is more than 30% higher than that of the single technology.Moreover,the indoor positioning method system has low cost and higher universality.