A Multi-Precision Indoor Localization Strategy Based on Hybrid Vive and Adaptive Monte Carlo Method

Abstract

The virtual reality (VR) motion tracking devices, such as the HTC Vive system, offer a low-cost, high-precision indoor positioning solution. However, Vive is applicable in a limited space, which does not fit the requirements of mobile robots. On the other hand, the simultaneous localization and mapping (SLAM) algorithms give a low-precision positioning in a wider area. In this paper, we propose a hybrid localization strategy that combines the advantages of Vive and SLAM such that a mobile robot can carry out delicate tasks around a workstation that requires high-precision positioning, and moves among workstations in a wide space with low-precision. To guarantee smooth and reliable transition from the low-precision to the high-precision area, we extend the workspace of Vive with additional lighthouses, and use the result of visual odometry to improve the robustness of Vive system after the tracker reconnection. Then we do experiments to verify the precision of the Vive system and the proposed hybrid localization strategy.