Enabling Plug-and-Play SLAM in Wireless Communication Systems

Abstract

Simultaneous localization and mapping (SLAM) during communication is emerging, which promises to provide information on propagation environments and the position of transmitters and receivers and create new services and applications for environment-aware communication. However, when fusing multi-domain measurements collected from RF signals and sensors, the unknown measurement biases may generate serious errors. In this study, we consider the practical measurement bias and develop a robust plug-and-play SLAM method. Specifically, we classify measurements into three categories according to their unknown biases. Next, we establish a Bayesian mechanism to fuse different categories of biased measurements, called the measurement plug-and-play mechanism. Finally, the corresponding unknown biases, such as clock and orientation biases, and RSS model parameters can be estimated during SLAM. Numerical results show that the proposed method can flexibly fuse different categories of measurements. Moreover, compared with the state-of-the-art method, under large bias levels, the proposed method can achieve 68% and 76% performance gain in localization and mapping, respectively.