This course presents the fundamentals of robotic perception based on a foundation of probability, statistics and information theory. Common sensor types and their probabilistic modeling are surveyed, including computer vision, Lidar, radar, GNSS/INS and odometry. Methods for feature extraction, description and matching, direct photometric and point cloud registration, outlier rejection are presented in the context of a robotic localization and mapping front end. Object detection and tracking, semantic segmentation and prior maps are fused to form a complete perceptual view of dynamic environments for a wide range of robotic applications.