An Optimal Algorithm Design of RSSI Indoor Location based on Neural Network

The RSSI positioning method is affected by complex environment, multipath effect and other factors, resulting in poor positioning accuracy and artificial interference in the selection of positioning model parameters. An optimal RSSI indoor location algorithm based on neural network is designed. Data elimination and Kalman filter are used to preprocess the RSSI value, and the RSSI value between the unknown node and the reference node is used as the input of the neural network, and the position coordinate of the unknown node is used as the output of the network to build the model. In this paper, BBPSO algorithm is used to optimize the smooth parameters of the network, and the root mean square error of the predicted coordinates and the actual coordinates of the positioning nodes are selected to build the fitness function, and GRNN neural network is used to build the precise positioning model of the network to achieve the prediction of the coordinates of the unknown nodes. The experimental results show that the optimized GRNN algorithm has higher positioning accuracy in node positioning. The positioning error of the optimized GRNN algorithm is less than 1 meter, and the average positioning error of 12 positioning nodes is 0.4436m. RSSI indoor positioning algorithm based on optimized neural network has high positioning accuracy and good stability, which has certain application value.