Research On Positioning Algorithms With Fusion Of Multiple Cellular Signal Features

At present,accurate indoor positioning service has become an indispensable part of people's life.With the development and application of the fifth generation communication technology,the cellular signals play a important role in indoor positioning.Indoor fingerprint positioning algorithms based on received signal strength,angle of arrive and channel state information had been widely studied among indoor positioning algorithms.In this paper,we mainly study the positioning method based on multi-fingerprint fusion using the cellular signals.Firstly,a new fingerprint fusion method is presented,and a weighted function with angle measurements as parameters is designed to constrain the signal fingerprint matching.The function is adaptive to the signal measurement statistics.Meanwhile,an algorithm for selecting the nearest neighboring nodes is developed by bounding the coordinates of nodes,and then WKNN algorithm is employed to implement the positioning.Secondly,to further improve the accuracy of the indoor positioning algorithms based on cellular signal,the method of obtaining CSI of the cellular signals are analyzed and then the method of improving the weight of WKNN algorithm based on Bayesian criteria is proposed.The probability of the online fingerprint vector appearing at the K points obtained by WKNN is calculated using the Bayesian fingerprint positioning algorithm based on the CSI and angle of the cellular signal is used to design the new weight of WKNN algorithm,thus the accuracy of indoor multi-fingerprint information fusion positioning algorithm is further improved.Moreover,considering that in some scenarios,efficient positioning can't be achieved without available cellular signals,the pedestrian dead reckoning trajectory correction algorithm based on multiple user information is proposed to improve the positioning availability.The distance of user pairs is first obtained by using the road sign and the Chip signal,then particle swarm optimization algorithm based on the distance between user pairs are used to find the optimal coordinate.Finally,the optimized coordinates are fused with the PDR positioning coordinates to correct the PDR positioning results.In this paper,multi-environment positioning simulation is conducted by changing the size of area and the dimension of fingerprint grid.The simulation results show that the positioning accuracy and reliability of the positioning method proposed in this paper can be improved significantly than other fingerprint-based positioning methods.Secondly,the method of designing the new weight of WKNN algorithm based on Bayesian criteria further improves the accuracy of indoor fingerprint information fusion positioning result.Lastly,the experiment shows that the PDR correction algorithm proposed in this paper can effectively correct the cumulative error of PDR positioning.