An Indoor Wi-Fi Localization Algorithm Using BP Neural Network

This chapter presents a Wi-Fi localization algorithm based on BP neural network. This localization algorithm first transforms the received signal strength indicator (RSSI) data by translation and scaling. A BP neural network is utilized to develop a ranging model based on the transformed RSSI data, which estimates the distance between the target point and each reference point. To improve the accuracy of the ranging model, the initial weights and biases of the BP neural network are optimized using a genetic algorithm (GA). Subsequently, localization is achieved using the ranging model alongside Sequential Quadratic Programming (SQP), an iterative nonlinear optimization technique. For brevity, the ranging model is referred to as GTBPD, and the localization method is referred to as GTBPD-LSQP. The performance of the ranging and localization algorithms are evaluated through conducting experiments in three areas of two academic office buildings.