Sensing & Estimation in Robotics: Course Projects
This set of projects delves into various aspects of sensing and estimation in robotics. In the first project, a projected gradient descent algorithm is utilized to estimate the orientation of a body, and cylindrical projection techniques are applied to construct detailed panoramas. The second project involves using the Iterative Closest Point (ICP) method and the Georgia Tech Smoothing and Mapping (GTSAM) library for precise trajectory estimation. LiDAR measurements and RGB-D camera images are used to create occupancy grid maps and texture maps respectively. The third project combines data from an IMU and a stereo camera to perform a comprehensive visual-inertial SLAM algorithm, employing an Extended Kalman Filter (EKF) to estimate the vehicle’s trajectory and the positions of surrounding landmarks.