Improvement And System Implementation Of Indoor Positioning Algorithm Based On RSSI

With the growing development of location based services(LBS),people's demand for LBS has expanded from outdoor to indoor.However,due to the reduction of electromagnetic signals by the walls,the outdoor positioning system cannot meet people's demands for precision of indoor positioning system.Accurate indoor positioning system plays an important role in prisons,hospitals,kindergartens,mining areas and many other places.However,at present,the existing indoor positioning systems are difficult to be popularized because of precision or cost.In view of the precision and cost of indoor positioning system,this thesis studies the indoor positioning system based on radio frequency identification(RFID),and improves the positioning algorithm of the core part of the system.The purpose is to improve the performance of the positioning system by improving the indoor positioning algorithm.It is found that in addition to the accuracy of the equipment itself,the multi-path effect caused by electromagnetic signals in the complex indoor environment are the factors that affect the positioning accuracy of the indoor positioning system based on RFID.Therefore,how to weaken or eliminate the multi-path effect becomes a key issue in this thesis.In this thesis,the measurement data is filtered by Kalman filtering algorithm to reduce the influence of the multi-path effect caused by complex environment and the measurement error caused by the equipment precision problem on the positioning system.Improving the k-Nearest Neighbor algorithm to avoid the use of defective tag,and using the moving average algorithm to modify the indoor positioning trajectory.Finally,the optimal parameters of the improved positioning algorithm are determined through experimental analysis,including the number of readers,the reference tag density and k values in the k-Nearest Neighbor algorithm.And in view of the difficulty of layout reference tags in mining areas,hospitals and other places,the BP neural network is introduced into the positioning process of the indoor positioning system.Through training and learning,the positioning measurement can be implemented in the environment without reference tags,which not only reduces the equipment cost,but also facilitates the system implementation.The simulation results show that the localization performance of the improved localization algorithm has improved enormously.And,the CC2430 chips are selected to complete the design and construction of the positioning system.Through the system testing,it is proved that the improved positioning algorithm and positioning system have better positioning performance and meet the basic positioning requirements.