Onboard Perception

Onboard perception focuses on equipping robots and vehicles with the ability to understand their surroundings using only locally available sensors, eliminating reliance on external infrastructure. Current research emphasizes developing robust and efficient algorithms, often employing convolutional neural networks and transformers, for tasks like object detection, pose estimation, and map creation from onboard camera, LiDAR, and other sensor data. This field is crucial for enabling autonomous operation in diverse and challenging environments, ranging from aerial manipulation and autonomous driving to space exploration and railway transportation, ultimately improving safety and efficiency in various applications.