A topological approach with MEMS in smartphones based on routing-graph

Abstract

GNSS (Global Navigation Satellite System) supported navigation with smartphones is an established technique. Solutions for position estimation for GNSS-shaded areas are developed more and more in the research field of the indoor navigation. The MEMS (MicroElectroMechanical Systems) sensors integrated into smartphones will be used increasingly for this. The aim of the research group at the HCU is the autonomous navigation only with MEMS sensors, favors are accelerometer, gyroscope and barometer. The quality of MEMS inertial sensors based position estimate decreases with time. Additional information should be used as support which is a prerequisite to realize a navigation application. A support is therefore possible with the available map data or the routing graph. In most approaches, the low-cost sensors are fused with the help of particle filter and Kalman filter at present. This serves the simple integration of external support, such as maps. In this work an approach without these filters is presented. A position estimate based on the routing nodes and edges shall be realized, only with the integrated MEMS inertial sensors in the smartphone. A fusion with further supports is not provided. The position estimate is calculated on the path network which is normally the basis for the routing to calculate the path. The position on a routing edge is derived from the acceleration sensors and the gyroscopes with step detection and orientation comparison. At a routing node, the sensor data is used to choose the probably nearest routing edge. For comparison purposes, approaches with Kalman filter and particle filter are applied to the same data set.