| Due to the complex indoor environment and more uncertain interference factors,the indoor location sensing method based on field strength has poor robustness and low positioning accuracy,which cannot meet the demand of indoor high-precision location sensing.To address the above problems,the thesis jointly applies Elman neural network and bio-inspired algorithm for indoor location sensing method,and conducts research and design of high precision indoor location sensing system positioning method to improve location sensing system positioning accuracy and stability.The main tasks are as follows:1.Based on the analysis of the implementation principle and localization method of fingerprint indoor location sensing,the overall design scheme of indoor location sensing based on Elman neural network is proposed.Two indoor positioning technologies,Bluetooth and visible light,are used as the implementation platform of the location-aware system.The location fingerprint database is established based on the method of signal reception strength,and the database is pre-processed with filtering algorithm and interpolation algorithm to improve the location-aware accuracy and stability.2.Analyze the visible light communication principle and the propagation characteristics of LED light source,establish the visible light channel propagation model,reasonably deploy LED light source,and build the visible light indoor positioning environment;analyze the propagation law of Bluetooth signal,establish the signal attenuation model,reasonably deploy i Beacon beacons,and build the Bluetooth indoor positioning environment.Different filtering algorithms are studied and analyzed to compare their performance,and the Kalman filtering algorithm is selected to pre-process the Bluetooth data;the interpolation algorithm is used to pre-process the visible light data,reset the samples,optimize the fingerprint database,and increase the data volume while improving the data validity.3.The principle of Elman neural network is studied,and an indoor location sensing model based on Elman network is established to improve the accuracy of indoor location sensing.To further improve the system accuracy and stability,the model is optimized by using K-means clustering algorithm and SSA algorithm.The database is divided into subsets by K-means algorithm,and sub-training models are established to obtain the coordinates of prediction points by two model predictions;the initial parameters of Elman network are optimized by SSA algorithm to avoid the network from falling into local optimum.In order to improve the positioning accuracy and stability of the system.4.Based on the proposed indoor location awareness method combined with visible l ight indoor positioning technology and Bluetooth indoor positioning technology,the expe rimental system is designed and built.On this basis,indoor location-awareness experime nts are conducted,the experimental results are analyzed,and an indoor location-awarene ss system application is designed to achieve real-time acquisition and display of indoor location-awareness information.Theoretical analysis and experimental results show that the indoor position sensing system designed in the thesis has an average positioning error of 3.22 cm in 2D and 5.13 cm in 3D in visible light indoor position sensing experiments;In the Bluetooth indoor location awareness experiment,the average positioning error in 2D is 0.54 m,and the average error in 3D is 0.73 m.The positioning accuracy meets the needs of different positioning scenarios,and the system stability is also effectively improved. |
