Research And Development Of Adaptive High-precision Indoor Wireless Positioning System

With the advancement of wireless communication technology,people pay more and more attention to location information.Because wireless signals are interfered during propagation,traditional positioning methods are difficult to apply in real life scenarios.Location-based fingerprint positioning technology has the advantage of strong universality and has become a current research hotspot,but it has the disadvantage of difficulty in fingerprint data collection.Pedestrian dead reckoning(Pedestrian Dead Reckoning,PDR)algorithm can be used to collect pedestrian motion characteristic information,and can get the position information of pedestrians during movement.However,the use of PDR algorithm alone to complete the positioning will cause cumulative errors,which makes the positioning error larger.As the fingerprint positioning or PDR method is used alone,there will be a large positioning error.Adaptive positioning methods were used in this thesis in order to achieve high precision positioning.In this thesis,The main tasks include:(1)A fingerprint extension method based on Ultra Wide Band(UWB)and Generative Adversarial Network(GAN)is proposed.The number of fingerprint data has a greater impact on the positioning accuracy,but the labor cost of collecting a large amount of fingerprint data is relatively large.Therefore,how to use a small amount of fingerprint data to achieve high positioning accuracy has become a difficult point in fingerprint positioning technology.To solve this problem,an indoor wireless positioning method based on UWB and GAN is proposed.First,get fingerprint data densely at equal intervals indoors,construct an initial fingerprint data set,select part of the fingerprint data in the initial fingerprint data set,and use GAN to obtain a large amount of fingerprint data using part of the fingerprint data;then,based on these generated data,use k nearest neighbor(K-Nearest Neighbor,KNN)classification algorithm model and random forest model for positioning prediction.Experimental results show that this method can use 50% of fingerprint data to achieve better wireless positioning accuracy,and achieve positioning accuracy similar to a large amount of fingerprint data.(2)A PDR indoor positioning method based on adaptive positioning area selection is proposed.In this thesis,using the acceleration value and angular velocity obtained by the accelerometer and gyroscope on the smart terminal device,the step length,step and direction angle of the pedestrian during the movement can be calculated.Combining with the previous position during the movement,the current position information can be obtained.And divide the positioning area into fingerprint positioning area and PDR positioning area,and use fingerprint positioning method and PDR positioning technology to locate in different positioning areas,reducing the cumulative error of PDR positioning method.Experimental results show that the positioning accuracy obtained by this method exceeds the positioning accuracy obtained by using the fingerprint positioning method and the PDR positioning method alone.(3)Design and build an adaptive positioning platform.Completed the software and hardware development of each part of the system,and used the system for positioning test analysis.In an experimental environment,the positioning results and errors of different positioning algorithms are compared and analyzed,which proves that the positioning accuracy of the adaptive positioning scheme proposed in this thesis is higher than that of other methods and can achieve better positioning accuracy.