An Indoor Localization Algorithm based on IMM in Mixed LOS/NLOS Environments

Wireless Sensor Networks (WSN) composed of micro-distributed sensors with sensory, computing and communication capabilities have attracted more and more attention due to their potential wide application. It is one of the most important applications to use WSN to locate the target. The positioning of the target is of great significance, which plays an important role in detecting the invasion of enemy targets, the movement direction of various objects, and the location of the event, providing a strong guarantee for the subsequent targeted processing. In this paper, an algorithm mixed IMM-KF and UKF with parameter correction is proposed, which is named as MKF-PC. In a LOS/NLOS mixed environment, a relatively good result can be achieved. After judging LOS and NLOS, KF and UKF are respectively used for filtering calculation. Finally, the error caused by initial assignment of noise parameters is alleviated by modifying noise parameters after filtering. According to MATLAB simulation results, the accuracy of MKF-PC is superior to classical KF, EKF, PF, UKF methods.