Research On The Fusion Location Of PDR And WIFI Fingerprint In BP Neural Network Based On Artificial Fish Swarm

With the rapid development of MEMS technology in mobile communication and MEMS,the demand for location-based services(LBS)has increased dramatically year after year.LBS is widely used in military logistics,medical storage,disaster relief and other fields.In the indoor environment,GNSS is difficult to accurately locate due to the complex indoor environment,Therefore,the indoor positioning technology called "the last centimeter of location services" has been favored by major technology companies and scientific research institutions.WiFi,inertial navigation technology is the main positioning technology used in the existing relatively mature indoor positioning system.In recent years,WiFi has attracted much attention due to its advantages of wide deployment,low price,strong scalability and simple implementation.In indoor WiFi scene positioning,WiFi fingerprints are widely used.Although WiFi fingerprint indoor positioning technology is constantly being improved,it still has the defects of low single-point positioning accuracy and instability.The PDR(pedestrian dead reckoning)method has the characteristics of high accuracy,no equipment dependence and strong universality in short time,but it has the shortcoming of cumulative error over time.Aiming at the problems existing in the above two commonly used positioning technologies,this paper mainly carries out the following research work:(1)In the WiFi fingerprint positioning part,firstly,in order to ensure the accuracy of the constructed fingerprint library,the signal filtering and denoising method in this paper is analyzed and established for the RSSI signal processing part.Secondly,aiming at the problems of low positioning accuracy and slow convergence speed in the traditional indoor WiFi positioning algorithm based on BP neural network,and considering that the indoor location can't be accurately located by using the signal strength attenuation ranging model,an improved artificial fish swarm optimization algorithm for indoor WiFi fingerprint positioning in BP neural network(IAFSA-BPWF)is proposed.With the excellent search speed and capability of swarm intelligence optimization method,the weight and threshold value of indoor positioning BP neural network is optimized by IAFSA-BPWF,which effectively avoids the disadvantage that the predicted value in traditional BP neural network is easy to fall into local optimal,and the corresponding relationship between signal strength value(RSSI)and position coordinate is established.Experimental results show that the proposed method improves the reliability of WiFi positioning and the algorithm has better stability.and the average positioning error is 1.58 m.It can be used to improve the initial position accuracy in the PDR positioning phase.(2)Aiming at the cumulative error problem in PDR,in view of the cumulative error problem in PDR,in the PDR positioning module,several common gait detection algorithms,step size models,and heading angle solving methods were analyzed indepth,and the step frequency,step size and yaw angle algorithms used in this paper are determined through relevant tests and comparative experiments.Record the location of special indoor geographical locations,such as indoor corners,doors,and the location of wireless AP,and establish a landmark database.The correction methods reduces the positioning error of PDR positioning part.(3)Finally,for the problems of poor stability in WiFi fingerprint positioning,unsustainable positioning,and accumulated errors in PDR,the extended fingerprint Kalman filter model was used to achieve WiFi fingerprint-PDR fusion positioning.And we completed the establishment of the corresponding experimental environment,terminal positioning software development and positioning test.The experimental results show that the fusion positioning method proposed in this paper can effectively avoid the disadvantages of separate WiFi fingerprint indoor positioning and PDR positioning.The advantages and disadvantages are complementary,and finally,high-precision position coordinate information can be obtained.The positioning trajectory is more close to the predetermined trajectory.The overall positioning accuracy has been further improved.