You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
C++17 reference implementation of camera map-matching localization: rasterize lane/road-boundary perception to distance transforms, score a 3-DOF (x, y, yaw) pose grid against a polyline map, temporally aggregate, and fuse into an SE(3) Kalman filter. Optional CUDA; KITTI eval, benchmarks, and RViz/PNG visualization.
Real-time multi-sensor fusion tracker: a hand-written EKF/IMM on a bare-metal STM32 fuses 24GHz radar + camera (YOLOv8) + IMU, with CAN 2.0B output and a live dashboard. 35% more accurate than radar alone; NEES/NIS-validated.
GPS denied navigation and control for an Ackermann vehicle in a confined corridor. ROS 2 Jazzy, Gazebo Harmonic, C++. EKF localisation and a three way lateral controller comparison. Work in progress.
Labs for the Robot Learning class, focusing on robotics and Reinforcement Learning. Each lab focuses on a different topic, had mandatory tasks and eventually extensions. All the results have been discussed in the reports.
EKF-based visual-inertial SLAM pipeline that fuses IMU and camera measurements for state estimation, landmark mapping, and trajectory tracking in autonomous robotics applications.
Low-cost ESP32-based differential drive ros2_control hardware interface — an affordable alternative to Raspberry Pi–based robots, with EKF sensor fusion and Nav2 support.