An Information Fusion Positioning Algorithm Based on Extended Dempster-Shafer Evidence Theory

Abstract

Due to a variety of noise interference, received signal strength indicator (RSSI)-based fingerprint data are often accompanied by uncertain factors. In order to solve the problem of positioning with high precision and accuracy in a complex indoor environment, this study designs a fingerprint positioning algorithm based on extended Dempster-Shafer evidence inference. First, a recognition framework is built to design a basic probability distribution function. Then a new evidence combination rule is proposed to assign different trust levels to the signal strength messages received from multiple sources, and the final position is obtained by converging the RSSI values. Finally, simulation experiments are conducted to show that the proposed algorithm is more valuable for improving the accuracy and accuracy of indoor positioning.