MetaDriverse for Autonomy Research

metadrive

MetaDrive: Open-source driving simulator

scenarionet

ScenarioNet: Scalable Traffic Scenario Management System for Autonomous Driving

trafficgen

[ICRA 2023] The official code for paper "TrafficGen: Learning to Generate Diverse and Realistic Traffic Scenarios"

ACO

[ECCV 2022] Learning to Drive by Watching YouTube Videos: Action-Conditioned Contrastive Policy Pretraining

cat

[CoRL'23] Adversarial Training for Safe End-to-End Driving

policydissect

[NeurIPS 2022] Official implementation of the paper: "Human-AI Shared Control via Policy Dissection"

metadrive-scenario

This repo contains scenarios from different source for training and testing autonomous vehicles.

pvp

Official release for the code used in paper: Learning from Active Human Involvement through Proxy Value Propagation (NeurIPS 2023 Spotlight)

metadrive-benchmark

Benchmark result of different RL algorithms on MetaDrive environments, including Multi-agent RL (IPPO, centralized critics, CoPO).

TS2C

[ICLR 2023] The official code for paper "Guarded Policy Optimization with Imperfect Online Demonstrations"

HACO

[ICLR 22] Official implementation of paper: "Efficient Learning of Safe Driving Policy via Human-AI Copilot Optimization"

CoPO

[NeurIPS 2021] Official implementation of paper "Learning to Simulate Self-driven Particles System with Coordinated Policy Optimization".

metadriverse.github.io

bevgen

openpilot

openpilot is an open source driver assistance system. openpilot performs the functions of Automated Lane Centering and Adaptive Cruise Control for over 200 supported car makes and models.