Enhanced Indoor Navigation Using Fusion of IMU and RGB-D Camera

Abstract

Accurate indoor navigation, especially precise attitude estimation is a challenge topic. Unlike the rate gyroscope in an IMU, the camera based visual-gyro does not suffer from drift errors. In order to overcome the drawbacks of the standalone systems, an INS/visual-gyro integration using direction cosine matrix (DCM) models is presented. Compared to the conventional Euler angle models, the usage of DCM can provide linear system models and avoided singularity problems. Using iterative closest point (ICP) algorithm, the depth measurement from a RGB-D camera can be converted to positioning information and further added to the integrated navigation system. To show the performance of the presented system, one field experiment is carried out with a mobile robot and the numerical results are shown.