Research on a Multimodal Fusion-Based Localization State Evaluation Algorithm for an Intelligent RF Power Supply

This paper proposes a multimodal fusion-based localization state evaluation algorithm for a mobile RF power supply. The algorithm computes the information entropy between point cloud data acquired by the laser sensor on the mobile RF power supply platform and a preconstructed prior grid map; a higher entropy score indicates more accurate localization. Meanwhile, IMU/odometry data fusion is performed by estimating the platform pose from the localization system at consecutive timestamps, and then calculating the differences in Euclidean distance and rotational attitude between this estimated pose and the poses output by the IMU and odometer at the same timestamps. By jointly considering static and dynamic conditions, the laser point cloud matching scores are integrated with the IMU/odometry fusion results to obtain the final localization state evaluation. Experimental results demonstrate that the proposed algorithm can accurately characterize the localization state of the intelligent RF power supply positioning system, achieving an evaluation accuracy of 99%.