ROCVLGDec 26, 2020

Improving the Generalization of End-to-End Driving through Procedural Generation

arXiv:2012.13681v220 citations
AI Analysis

This work addresses the overfitting issue and lack of generalization in learning-based self-driving systems for researchers and developers by providing a highly configurable simulator.

This paper introduces PGDrive, a procedurally generated driving simulator designed to improve the generalization of end-to-end driving systems. By training with an increasing number of procedurally generated scenes, the generalization of the agent across different traffic densities and road networks is significantly improved.

Over the past few years there is a growing interest in the learning-based self driving system. To ensure safety, such systems are first developed and validated in simulators before being deployed in the real world. However, most of the existing driving simulators only contain a fixed set of scenes and a limited number of configurable settings. That might easily cause the overfitting issue for the learning-based driving systems as well as the lack of their generalization ability to unseen scenarios. To better evaluate and improve the generalization of end-to-end driving, we introduce an open-ended and highly configurable driving simulator called PGDrive, following a key feature of procedural generation. Diverse road networks are first generated by the proposed generation algorithm via sampling from elementary road blocks. Then they are turned into interactive training environments where traffic flows of nearby vehicles with realistic kinematics are rendered. We validate that training with the increasing number of procedurally generated scenes significantly improves the generalization of the agent across scenarios of different traffic densities and road networks. Many applications such as multi-agent traffic simulation and safe driving benchmark can be further built upon the simulator. To facilitate the joint research effort of end-to-end driving, we release the simulator and pretrained models at https://decisionforce.github.io/pgdrive

Code Implementations2 repos
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes