Research On UWB Indoor Positioning Algorithm Based On Genetic Neural Network

With the human society entering a new era of mobile internet,related services based on geographical location information have also developed rapidly.At present,people are not just satisfied with the location information service in the outdoor enviroument,but have great demands for the service such as personnel and equipment positioning in the indoor environment.However,the indoor environment is different from the open outdoor's because there are many obstacles,which affect the location effect of wireless signal.Therefore,how to improve the related technology research of indoor wireless positioning accuracy has become a hot spot.Ultra-wideband(UWB)technology is a new wireless communication technology and this paper expands the research of how to improve the accuracy of UWB indoor positioning:Firstly,the channel simulation is carried out under UWB channel model provided by IEEE 802.15.4a standard,and it is determined that non-line-of-sight propagation(NLOS)is the main factor that causes the propagation delay and signal intensity attenuation of UWB signal in indoor environment.For the problem of NLOS,this paper analyzes the advantages and disadvantages of four commonly used positioning methods based on received signal strength(RSS),angle of arrival(AOA),time of arrival(TOA),and time difference of arrival(TDOA).Among this methods,TDOA can give full play to the advantages of UWB signal with high time resolution,thaf5 why this paper chooses the TDOA method.After determining the positioning method,this paper introduces the typical UWB positioning algorithm based on TDOA:Chan algorithm,Taylor algorithm and Chan-Taylor cooperative algorithm combining the advantages of both.The simulation of three algorithms shows that the positioning accuracy will be greatly affected when there is a high NLOS error in the environment.Therefore,a genetic neural network algorithm is proposed to fit the mapping relationship between the TDOA measurement data and coordinate values of the target points,and the strong nonlinear mapping ability of genetic neural network is used to reduce the influence of NLOS error in TDOA measurement data on positioning accuracy.The simulation results show that the positioning accuracy of genetic neural network algorithm is obviously higher than the previous three algorithms in the presence of higher NLOS error,and the algorithm has better stability with the improvement of NLOS error.At the same time,in order to obtain higher positioning accuracy,this paper proposes to combine the genetic neural network algorithm with the Taylor algorithm by referring to the Chan-Taylor collaborative algorithm.Simulation results show that the combined algorithm has better positioning performance than the simple genetic neural network algorithm.