Determining drivable free-space for autonomous vehicles
Published:
In this work, we use Deep neural networks to compute both a set of boundary points that correspond to a boundary dividing drivable free-space from non-drivable space in the physical environment and class labels for boundary points, given image data coming from the sensors of self-driving vehicle. This module helps the self-driving vehicle to distinguish drivable area from non-drivable area like pavements as well as identify the first obstacle in the drivable path. The class labels like pavement, pedestrian, vehicle etc. help in distinguishing the priority of obstacle avoidance for the self-driving vehicle in case of an accident.
Download here