Implementation of NLOS based FPGA for distance estimation of elderly using indoor wireless sensor networks

Abstract

Location or distance estimate data for the elderly is a major drawback in wireless sensor networks (WSNs). Previous research used an Artificial Neural Network (ANN) in the indoor spaces to calculate the distance between elderly people and a unique anchor node in the ZigBee WSN. The neural network was evaluated on a “Field Programmable Gate Array (FPGA)” because it was being used in a real-world application. To training, testing, and validating the ANN, the RSSI values of anchor nodes were extracted. As a result, the MAE-estimated distance error improved. In addition, one of the issues that might affect distance estimate accuracy is the mobility of persons inside the testing phase. As a result, the proposed work aims to concentrate on NLOS circumstances to examine their impact on estimated distance. NLOS situations are most frequent indoors when several rooms within a structure are involved. Difficulties can cause Non-Line-Of-Sight (NLOS) propagation when WSNs are implemented in indoor environments. However, the proposed method overcomes the issue of selecting the nodes by combining an optimization method including PSO with an FPGA to find the smallest number of nodes with a suitable “Distance Estimation Error”. To execute the indoor wireless sensor network (WSN) topology, the system includes ZigBee technology. As a result, the FPGA implementation's complexity can be decreased and it outperforms the other existing algorithms. As can be observed, the analysis of the proposed hybridizing method based on WSN is highly effective in enhancing the elderly's well-being.