A Novel SLAM Method Using Wi-Fi Signal Strength and RGB-D Images

With the widespread deployment of Wi-Fi access points (APs), the Wi-Fi has become a promising means for Simultaneous Localization and Mapping (SLAM). Benefited from the wide coverage of Wi-Fi, the Wi-Fi signal can be used for large-scale SLAM. Therefore, a SLAM method using the Wi-Fi signal strength to estimate the location of the robot and the Wi-Fi APs based on Extended Karman Filter (EKF)is proposed. However, influenced by the noise of the measurement of Wi-Fi signal strength, it is hard to achieve an accurate result in SLAM which only relies on the Wi-Fi signal strength measurement. To improve the accuracy of the SLAM, a fusion of the Wi-Fi signal strength and the RGB-D images is needed. Therefore, we propose a novel SLAM method using both the Wi-Fi signal strength and the RGB-D images, which consists of the EKF and the graph-optimization. The EKF part employs the Wi-Fi signal strength information to estimate the pose of the robot and the locations of the APs, and the graph-optimization part employs the RGB-D images to estimate poses of the robot. The experiments prove that our method achieves a better localization and mapping result than traditional RGB-D SLAM methods, especially improves the robustness of the SLAM system.