Research On UWB Indoor Location Algorithm

With the development of wireless communication technology,people pay more and more attention to indoor high-precision positioning technology.However,in the complex indoor wireless environment,the signal is easily affected by various interference factors,which leads to the poor positioning performance of traditional indoor positioning algorithm.Ultra wide band is a new type of communication technology with strong ability of anti-multipath,and narrow band interference,It also has high transmission rate and strong penetration.And UWB positioning has higher positioning accuracy,which can well meet the needs of indoor positioning,so the research on UWB positioning technology has very important application value.Based on the theory of UWB positioning,this paper studies UWB ranging and positioning.Firstly,according to the literature,the current UWB ranging methods are analyzed,and the single sided two-way ranging method is used to carry out ranging experiments in LOS and NLOS environments,and the ranging error is modeled and analyzed.Based on the established ranging error model,the performance of common UWB positioning algorithms is simulated and compared.The results show that in the complex NLOS environment,only using a positioning algorithm,the positioning performance is poor,so this paper combined common UWB positioning algorithms with Kalman filter to optimize the results.Because the performance of Kalman filter becomes worse in non Gaussian and complex multipath conditions,this paper further discusses an improved Kalman filter fusion localization algorithm based on NLOS recognition.The simulation results show that: in LOS environment,the Kalman-Chan algorithm has better positioning performance than Chan Taylor colocalization algorithm and Chan algorithm,while in NLOS environment,the improved Kalman filter can also improve the positioning accuracy.Secondly,aiming at the problem of NLOS ranging error suppression,a UWB positioning algorithm based on convolutional neural network is studied.The algorithm uses CNN to train the impulse response of the original channel of UWB to realize the regression prediction of the ranging error under NLOS.The ranging value subtracts the predicted ranging error to get the modified ranging value,and then the modified ranging value is used by the UWB positioning algorithm to get position.Simulation results show that: compared with the existing Wylie algorithm,the proposed algorithm has better positioning performance in NLOS environment.Finally,the paper builds the UWB test platform and collects the data in the actual scene,and verifies the above positioning algorithms.The experimental results show that the proposed positioning algorithm has positive significance for improving the indoor positioning accuracy.