I. Introduction
5G simultaneous localization and mapping (5G SLAM) utilizes the underlying geometric information of 5G signals to estimate the user location and to create maps of the environment [1]–[5]. 5G cellular systems bring significant advantages to multipath-assisted SLAM due to their large bandwidth and beamforming capability [6], [7]. This means a better resolution in the delay and angular domains, thereby efficiently resolving and identifying multipath components (MPCs) to achieve better positioning and mapping accuracy. An end-to-end 5G SLAM framework was described in [5], which contains downlink data transmission, channel estimation, clustering, and SLAM filtering. It is obvious that the accuracy of the channel estimations directly affects the localization and mapping performance, and the complexity of the channel estimator greatly affects the real-time performance. Therefore, the channel estimator plays a significant role in the 5G SLAM framework, and having a low-complexity and high-accuracy channel estimator is an important problem.