Dynamic object tracking on UGV

Abstract

The techniques for obtaining 3D profiles of the environment had advanced very far, but interpreting this data and using it for robot navigation have not been widely explored yet.The LIDAR sensor is able to take 1 million readings per second, scan 360° horizontally and 26.8° vertically, and sense objects 120m away. With this high resolution data points, objects can be detected accurately further away compared to conventional sensors. We can potentially use it for navigating vehicles at high speed safely, paving the way for autonomous, road-worthy vehicles.With this objective in sight, we will explore how to turn the massive amount of data collected into usable information for navigating. More specifically, we will condense the data into objects of which locations can be tracked and predicted. [3rd Award]