Research On Indoor Location Algorithm Of Optimized RBF Neural Network Based On Bluetooth Low Energy

The rapid development of artificial intelligence and the Internet of Things has promoted the close integration of indoor positioning technology with wireless signals and intelligent algorithms.The demand for location services is increasing day by day,especially in complex indoor environments where it is often necessary to obtain the location information of objects.Therefore,a low-cost,high-performance indoor positioning algorithm has become the research goal of this paper.In this paper,the optimization research of indoor positioning algorithm is mainly carried out on the basis of Bluetooth low energy,and simulation experiment verification and experiment analysis are carried out respectively.The article first analyzes the advantages and disadvantages of several commonly used indoor positioning algorithms,chooses the position fingerprint algorithm for position estimation,and uses the Radial Basis Function(RBF)neural network as the matching algorithm in its online positioning stage.Through the excellent self-learning ability of RBF neural network,complex nonlinear problems are quickly transformed into simple linear problems,and finally the functional relationship model corresponding to "fingerprint" is fitted.According to the principle formula of RBF neural network,the parameters that need to be optimized are derived—spread.In the next article,we choose to introduce Particle Swarm Optimization(PSO).Using its simple structure and optimization mechanism to maintain the advantages of group diversity,combined with the excellent global optimization ability of Genetic Algorithm(GA),we propose After a limited number of iterations using the hybrid particle swarm optimization algorithm(PSO-GA),the optimal spread value is searched out.In this way,the local optimization of the particle optimization process is avoided,and the estimation and generalization capabilities of the RBF neural network model are improved.Secondly,in order to obtain the sample data of the experiment,the article selects TI's CC2640 chip as the core on the basis of the Bluetooth low energy protocol stack,and builds the iBeacon base station as a module for transmitting Bluetooth signals.And use Android to develop smart phone terminal application program to achieve the function of receiving and displaying Bluetooth signal strength value(Received Signal Strength Indication,RSSI).At the same time,establish a coordinate system in the indoor positioning space,divide the grid points,test the RSSI value reception on the spot and select a relatively strong location to deploy the iBeacon base station,and build a data collection platform to obtain sample data.In view of the interference caused by environmental factors to the signal,the singular value in the RSSI value is removed,and then the remaining data is processed by Kalman filtering to achieve the purpose of reducing signal noise and smoothing the RSSI value.The article finally carries on the experiment simulation analysis to the involved algorithm on the MATLAB simulation platform.The PSO and PSO-GA algorithms are used to search the optimal value of the spread factor to establish the corresponding positioning model.Starting from the performance indicators such as the positioning accuracy and stability of the positioning model,the focus is on analyzing the positioning performance of the RBF neural network model,PSO-RBF neural network model and PSO-GA-RBF neural network model,and the PSO-GA-RBF neural network model is obtained It has the advantages of higher positioning accuracy and smaller error fluctuation.