Efficient robotic SLAM by fusion of RatSLAM and RGBD-SLAM

This paper presents a novel method to improve the robotic SLAM by fusing aspects from the RGBD-SLAM and RatSLAM methods. The proposed solution takes the 3D feature representations of the environment much like the RGBD-SLAM system but instead of creating a voxel map, it creates a grid map taking elements from the RatSLAM system. The resulting integration of SLAM methods eliminates the need for a continuous stream of images as required in RGBD-SLAM and thus overcomes the problem of ambiguous data normally generated by featureless surface such as repeating wall patterns. The proposed solution is implemented on a Pioneer 3-DX mobile robot using Kinect sensor. The results show that the proposed method can successfully conduct SLAM in an efficient manner.