Underground adaptive positioning algorithm based on SAPSO-BP neural network

In view of problems of slow convergence, easy to form local extremum and large positioning error in strong time-varying electromagnetic environment of underground positioning algorithms based on traditional BP neural network, an underground adaptive positioning algorithm based on simulated annealing and particle swarm optimization and BP neural network (SAPSO-BP) was proposed. SAPSO algorithm is used to optimize the initial weight and threshold of BP neural network to accelerate training convergence speed and make it reach the global optimum. The target point RSSI value is collected by wireless calibrator installed in underground roadway and real-time calibrated by adaptive dynamic calibration method, in order to reduce influence of time-varying electromagnetic environment on positioning accuracy. Finally, the SAPSO-BP neural network is used to estimate position coordinates of target point. The experimental results show that confidence probability of positioning error within 2 m of the proposed algorithm is 77.54%, average error is 1.53 m, the positioning performance is better than uncalibrated SAPSO-BP neural network algorithm, PSO-BP neural network algorithm and BP neural network algorithm.