Research On Wireless Location Algorithm Based On Indoor Environment Characteristics

Indoor positioning services are related to national development,economic rise and public safety.With the development of Wireless Sensor Networks(WSN),Location Based Service(LBS)is receiving more and more attention,and the demand for indoor high precision positioning requirements are also increasing.Due to the complicated distribution of obstacles in the indoor environment,there are serious multipath fading and complicated time-varying characteristics of wireless positioning signals.High-quality wireless location needs to further improve the robustness of the system to changes in indoor environment characteristics.Under normal circumstances,such as warehouses,libraries and other places,the location of indoor obstacles,tables and chairs and other fixed equipment will not change significantly,so these places are considered to be in a static environment.However,in places such as parking,offices and large-scale supermarkets,due to the fact that many pedestrians and vehicles move at any time,the path fading effect of indoor wireless signal propagation is more serious,and the indoor place can be considered as a dynamic environment.Therefore,this paper mainly studies the key technologies of high-precision wireless location in static and dynamic indoor environment,and it is mainly divided into the following aspects:In a static indoor environment,the accuracy of the indoor positioning system based on Time of Arrival(TOA)is limited by the accuracy of pseudorange and the number and spatial layout of effective anchor points(AP).The realization of high-precision wireless location not only needs to further reduce the pseudorange error of ranging signals,but also needs the optimal deployment of anchor points,without increasing the number of anchor points,the best deployment of the anchor points can obviously improve the positioning accuracy.In the third part of this paper,firstly,the problem of high precision positioning signal model in static indoor environment is studied,and the Sinusoidal Frequency Modulation(SFM)signal model is proposed.Then,it is proved theoretically that the SFM signal has a higher range resolution than the traditional indoor positioning signal from the Wideband Ambiguity Function(WAF),and can achieve high-precision indoor positioning.Then the ranging performance and positioning performance of the new model is simulated.Compared with the traditional indoor positioning signal,the superiority of the proposed SFM signal model is verified.In the fourth part,aiming at the problem of multi-satellite positioning when the number of active anchor points is large,a criterion using Weighted Horizontal Dilution of Precision(WHDOP)as the anchor point selection is proposed.The WHDOP selection criteria can give greater weight toanchor points with smaller errors and can eliminate the adverse effect of coarse errors on the positioning system.Therefore,the positioning accuracy can be further improved.The robust performance of WHDOP positioning system under different indoor noise environments(Gauss noise and Rayleigh noise)is simulated.With the increase of noise environment,the robust performance of WHDOP method is stronger.The WHDOP is more capable of reflecting the trend of actual positioning error than the conventional HDOP.Then a high precision indoor wireless location based on SFM-WHDOP method is proposed.Compared with the UWB-HDOP location method,the experimental results show that the SFM-WHDOP method can realize the indoor wireless location with high accuracy when there are more effective anchor points.In the fifth part,aiming at the problem of single-satellite positioning in static indoor environment,an indoor single-satellite positioning method based on IMU assistant is proposed,while can gradually eliminate the cumulative error effect of IMU in the indoor environment through the constraint of scene analysis,and further optimize the position of positioning target by single-satellite positioning.Simultaneously,a weighted fusion scheme of various information data is given according to the criterion of minimizing error variance.Through the weighted fusion of multi-information,the accuracy of the single-satellite positioning system not only makes full use of the estimated value of each subsystem but also effectively restrains the impact of large noise on the overall accuracy.It is theoretically proved that the weighted fusion method can further improve the system's positioning performance.And the theoretical analysis of the system's positioning error is analyzed on the calculus.Finally,the algorithm is tested by real test in a garage,which verifies the feasibility of the algorithm and realizes the positioning requirement of sub-meter precision in static indoor underdetermined circumstances.Finally,in order to mitigate the adverse effect of the indoor dynamic environment on the accuracy of the positioning system,an indoor positioning algorithm based on transfer learning is proposed.By random sampling in the space after the environment changes,the algorithm finds the potential feature space for the calibration of the position,so as to avoid the problem of poor positioning accuracy caused by the failure of the original fingerprint database.The experimental results show that the proposed algorithm has obvious advantages in positioning accuracy,positioning error range and small error positioning confidence probability.