Conditional Affordance Learning for Driving in Urban Environments

TL;DR

This paper introduces a direct perception method for autonomous urban driving that effectively navigates complex environments, handles traffic signals, and improves safety and navigation success rates over existing approaches.

Contribution

It presents a novel direct perception approach that maps video to intermediate representations for urban driving, handling traffic lights and signs with image-level labels, outperforming prior methods.

Findings

Achieves up to 68% improvement in goal-directed navigation in CARLA

First to handle traffic lights and speed signs with image-level labels

Reduces traffic accidents in simulation

Abstract

Most existing approaches to autonomous driving fall into one of two categories: modular pipelines, that build an extensive model of the environment, and imitation learning approaches, that map images directly to control outputs. A recently proposed third paradigm, direct perception, aims to combine the advantages of both by using a neural network to learn appropriate low-dimensional intermediate representations. However, existing direct perception approaches are restricted to simple highway situations, lacking the ability to navigate intersections, stop at traffic lights or respect speed limits. In this work, we propose a direct perception approach which maps video input to intermediate representations suitable for autonomous navigation in complex urban environments given high-level directional inputs. Compared to state-of-the-art reinforcement and conditional imitation learning approaches, we achieve an improvement of up to 68 % in goal-directed navigation on the challenging CARLA simulation benchmark. In addition, our approach is the first to handle traffic lights and speed signs by using image-level labels only, as well as smooth car-following, resulting in a significant reduction of traffic accidents in simulation.