Research On The Target Location And Tracking Algorithm Based On Kalman Filter

With the continuous development of wireless communication technology and sensor technology,wireless sensor nodes emerge as the times require.Its advantages are miniaturization,low power consumption and low cost.Wireless sensor network contains a large number of wireless sensor nodes.At the same time,with the development of 5 G technology,wireless sensor networks can be put into use in a large number of fields.In these application fields using wireless sensor networks,the location information of nodes is particularly important.We mainly study the ranging algorithm based on received signal strength indication(Received Signal Strength Indication,RSSI)which can be used in the process of vehicle driving,The advantages of RSSI based ranging and positioning method are low power consumption,low cost and no additional hardware equipment.However,due to the weak penetration of millimeter wave signal used in 5G,the high dynamic characteristics of vehicles and the complex and changeable road traffic environment,it will be affected by a variety of factors,resulting in inaccurate positioning in the case of large ranging error.In view of the above problems,the main contents of this paper include the following aspects:Firstly,according to the ranging and positioning model,the influence of distance and environmental factors on RSSI is studied,and the Kalman filter algorithm is used to smooth it.The experiment shows that the Kalman filter algorithm can achieve better filtering effect in a few iterations.Secondly,a pre filtering localization algorithm based on RSSI is proposed.By pre filtering the RSSI and combining with the trajectory information estimated by the motion model,the noise fluctuation and data mutation in the RSSI data are reduced and eliminated,and the target is accurately located.The simulation results show that,compared with the traditional algorithm,it has higher fault tolerance rate and stronger anti-interference ability in strong occlusion environment,and has certain practical value,which can greatly improve the vehicle positioning and tracking accuracy in complex infinite electromagnetic environment.Finally,the proposed method is improved by adding OSELM(Online Sequential Extreme Learning Machine,OSELM)neural network to predict the state transition matrix.The parameters of the state transition matrix are updated adaptively when the target is under strong occlusion.The training process of the network is optimized according to the singular value decomposition theory.The computational complexity of the fitting is reduced by reducing the number of hidden layer nodes of oselm network,So the improved prefilter localization algorithm can be used in more complex and changeable driving state.