Angle Domain Channel-Based Camera Pose Correction for Vision-Aided ISAC Systems

Visual image has been explored to aid various communication tasks in the integrated sensing and communication (ISAC) systems. However, vision-aided ISAC system requires accurate measurement of camera pose, i.e., camera’s location and orientation, to achieve precise object detection and effective assistance capability. In this letter, we propose a novel channel angle information based camera pose correction method (CPCM) to reduce the measurement error of the camera pose. With the aid of 3D object detection, the deployed camera is used to obtain a biased estimate of the orientation angle of the user relative to BS (OUB). Then, we obtain the approximate estimate of the OUB from the path angle of wireless uplink channel. Then, we calculate the measurement error of the camera pose by the deviation between the biased and approximate estimate of the OUB. Simulation results show that the proposed CPCM can effectively rectify the camera’s location and orientation under different measurement errors.