An Improved BPNN Method Based on Probability Density for Indoor Location

With the widespread use of indoor positioning technology, need for high-precision services is rising; nevertheless, there are several challenges, such as difficulty simulating distribution interior location data and enormous inaccuracy probability computation. As a result, this paper proposes three different neural network model comparisons based on WiFi fingerprint - algorithm improved back propagation model, RSSI angle change, depth change to raise accurately predict coordinates. Changing action range activation function in standard back-propagation achieves goal predicting The revised has strong stability enhances accuracy experimental loss rate (loss), (acc), cumulative (CDF).