Research On Indoor Location Algorithm Based On RF RSS And Multi - Sensor Fusion

The demand for indoor positioning is increasing gradually, so does the demand on thepositioning accuracy. In terms of positioning accuracy, positioning ranges, equipment cost,and flexibility, current indoor positioning technology is not perfect yet, which hesitates thepopularity in applications. With the development of wireless communication technology,many indoor positioning methods based on wireless signals have emerged, such asultrasound, infrared, optical signals, radio frequency, etc. Particularly, the radio frequencyincludes radio frequency identification and wireless networks. In recent years, as a kind ofwireless local area network, Wi-Fi(Wireless Fidelity) is developing rapidly. Wi-Fi-basedindoor location positioning has become a hot research topic. The main theoreticalalgorithms based on Wi-Fi positioning technology are fingerprint algorithm and trilateralalgorithm. Meanwhile, not only wireless technology, but also some inertial sensors havebeen used in indoor positioning. The former provides convenient coverage, and requires lessequipment; while the latter is not susceptible to the surrounding environment. Taking intoaccount the strengths of these technologies, multi-sensor integration methods have beenwidely used in the positioning application.Our main work is as follows:1. Proposed RSS-based compensational trilateration localization algorithm. Taking intoaccount the different material structure and interior layout of buildings, the results of sameindoor positioning methods usually diverse because of different environmental factors.Different building material structure, interior decoration, indoor multipath phenomena andother factors will make positioning more difficult. In general, the positioning accuracy willnot improved if depending on the ideal state model for positioning. Therefore, this paperproposes RSS-based compensation model for trilateration positioning method. This methodconsiders the affect brought by different material of indoor barrier material, determines thewall attenuation factor, and then model will be more accurately. This method eliminates theimpact on signal propagation caused by indoor barrier material. Moreover, this paper selectsa common indoor room as experiment environment. We also choose fingerprint positioningto experiment, and then compare the results. The experimental results demonstrate that theproposed method can effectively improve the positioning accuracy.2. Proposed WLSE-KF-based multi-sensor integration location algorithm. In the process of using sensors for locating, we need to fuse the collected date. Particularly for the latter,Least Squares (LS) algorithms and Extended Kalman Flitering (EKF) have been introducedin the integration. Generally, people use the LS to obtain the locations based on RFID.However, it may lead to noticeable positioning errors. In this paper, we propose an accuratealgorithm by combing the weighted LS with EKF algorithms for RFID integration location.First, we use LS to estimate the preliminary positions of the mobile user. And then the EKFalgorithms are adopted to furtherize the accuracy of these positions. Simulation results showthat the proposed algorithm can significantly reduce positioning errors, while improve thepositioning accuracy.