Particle filter based landmark mapping for SLAM of mobile robot based on RFID system

Abstract

This paper proposes a novel Simultaneous Localization and Mapping (SLAM) based on distributed particle updates for landmark mapping and validates it with an HFband RFID-based mobile robot. Multiple RFID readers are embedded at the bottom of an omni-directional vehicle and tags are installed on the floor. The IC tags are used as landmarks of the environment. FastSLAM[1] uses particles to estimate the position and orientation of the robot and Kalman filter to update the positions of IC tags. However, an update of the detected IC tags with Kalman filter is not appropriate because the probability of the IC tag detection cannot be modeled with a Gaussian distribution. We use two separate particle filters to estimate both the position and orientation of the robot and positions of IC tags simultaneously. The proposed method has been tested on the simulated and real environments. Experimental results show the validity and computational efficiency of the proposed method.

Research on 3D semantic map based on SLAM and YOLOv4

“…In this paper, the realization of simultaneous localization and map building of mobile robots is based on SLAM [5] . There are five steps in the process of classic SLAM, as follows:…”