Improving tracking error by dead reckoning and RSSI technologies with a fuzzy fusion scheme in indoor location

Abstract

Studies on indoor positioning systems have typically used received signal strength indicators (RSSIs) for positioning; RSSI offers the advantage of low calculation complexity, but indoor environments frequently cause RSSI to produce large errors. Improving the RSSI positioning accuracy requires deploying a large number of reference nodes, which increases both the cost and the computational complexity of RSSI. Because of the progress of technology, some sensors now feature high precision and low manufacturing cost. Therefore, a variety of inertial sensor components are commonly applied, and numerous studies on positioning systems have avoided RSSIs and have used inertial sensors to achieve indoor positioning by constructing various dead reckoning (DR) systems. Inertial sensors typically suffer from cumulative errors. This study proposes a fuzzy control method to combine DR and RSSI systems, to produce a positioning system with low positioning error. DR positioning can produce superior positioning accuracy, and the RSSI method does not generate cumulative errors; thus the proposed system solves the cumulative error problem that plagues DR systems. Compared with other positioning technologies, the method proposed by this study can be applied prevalently; one need not consider the environment for deployment of reference nodes. This reduces the complexity of the deployment of a positioning system. In addition, when the angular deviation for the direction of motion of the DR system was large, the proposed method had superior correction performance. With the proposed fuzzy control method, the average positioning error was 0.625 m, which was 73.8% more accurate than the average positioning error obtained using the conventional DR method.