| Based on Bluetooth RSSI positioning,due to low cost,simple operation and high positioning accuracy,it is widely used in various industries.In recent years,the demand for indoor positioning technology has grown rapidly,the indoor environment is complex,and the RSSI signal is seriously affected by the multipath effect,resulting in a reduction in indoor positioning and positioning accuracy of RSSI.This paper studies the key technologies based on RSSI neural network positioning,improves indoor positioning accuracy,and has important engineering application value.The main research work of the thesis is as follows:1.Considering the interference of the refraction and reflection of wireless signals in an indoor environment,this paper adopts the Threshold Kalman filter(Threshold Kalman filter,TKF)method,which can effectively filter the fluctuations of RSSI signals.2.The anchor node layout directly affects the positioning accuracy.This paper establishes a multi-objective optimization function for positioning accuracy and coverage area,and uses particle swarm optimization to optimize the layout of the smallest unit of the 3 anchor nodes.3.In order to reduce the labor cost of establishing neural network training samples,this paper proposes a multi-point multi-map probability sample enhancement method.According to the propagation characteristics and spatial correlation of wireless signals,through a small amount of measured point data,quickly construct multi-sample point data characteristics that reflect the environmental impact,and on this basis,use the probability method to generate multi-map normal distribution samples that characterize wireless signal fluctuation data.The experimental results show that the sample set generated by the multipoint multi-mapping probabilistic sample enhancement method has better characterization of signal characteristics,the positioning accuracy of the trained neural network has been significantly improved,and the workload of constructing the training sample set for manual collection can be reduced by 85.7%.4.Consider the fluctuation of the RSSI of the wireless signal.In this paper,the segmented Bezier curve is used to fit the trajectory of people in the indoor environment.On this basis,the coordinates of the discrete positioning points are mapped vertically to the Bezier fitting curve according to the direction of movement.In order to achieve secondary positioning,experiments have shown that the secondary positioning results can improve the positioning accuracy in the moving path in different situations. |
