Self-Supervised Steering Angle Prediction for Vehicle Control Using Visual Odometry
This addresses the data cost issue for self-driving car development, though it is incremental as it matches rather than surpasses existing supervised approaches.
The paper tackles the problem of expensive labeled data for vision-based vehicle control by training a model to predict steering angles using visual odometry in a self-supervised manner, achieving performance comparable to supervised methods in the CARLA simulator.
Vision-based learning methods for self-driving cars have primarily used supervised approaches that require a large number of labels for training. However, those labels are usually difficult and expensive to obtain. In this paper, we demonstrate how a model can be trained to control a vehicle's trajectory using camera poses estimated through visual odometry methods in an entirely self-supervised fashion. We propose a scalable framework that leverages trajectory information from several different runs using a camera setup placed at the front of a car. Experimental results on the CARLA simulator demonstrate that our proposed approach performs at par with the model trained with supervision.