Nocturne: a scalable driving benchmark for bringing multi-agent learning one step closer to the real world

@article{Vinitsky2022NocturneAS,
  title={Nocturne: a scalable driving benchmark for bringing multi-agent learning one step closer to the real world},
  author={Eugene Vinitsky and Nathan Lichtl'e and Xiaomeng Yang and Brandon Amos and Jakob N. Foerster},
  journal={ArXiv},
  year={2022},
  volume={abs/2206.09889}
}
We introduce Nocturne , a new 2D driving simulator for investigating multi-agent coordination under partial observability. The focus of Nocturne is to enable research into inference and theory of mind in real-world multi-agent settings without the computational overhead of computer vision and feature extraction from images. Agents in this simulator only observe an obstructed view of the scene, mimicking human visual sensing constraints. Unlike existing benchmarks that are bottlenecked by… 

Figures and Tables from this paper

References

SHOWING 1-10 OF 34 REFERENCES

Large Scale Interactive Motion Forecasting for Autonomous Driving : The Waymo Open Motion Dataset

TLDR
This work introduces the most diverse interactive motion dataset to their knowledge, and provides specific labels for interacting objects suitable for developing joint prediction models, in a bid to provide strong baseline models for individual-agent prediction and joint-prediction.

Symphony: Learning Realistic and Diverse Agents for Autonomous Driving Simulation

TLDR
S Symphony is proposed, which greatly improves realism by combining conventional policies with a parallel beam search, and factoring agent behaviour into goal generation and goal conditioning, and ensures that agent diversity neither disappears during adversarial training nor is pruned away by the beam search.

DriverGym: Democratising Reinforcement Learning for Autonomous Driving

TLDR
DriverGym is proposed, an open-source OpenAI Gym-compatible environment tailored for developing RL algorithms for autonomous driving, and behavior cloning baselines using supervised learning and RL, trained in DriverGym are provided.

MetaDrive: Composing Diverse Driving Scenarios for Generalizable Reinforcement Learning

TLDR
The generalization experiments conducted on both procedurally generated scenarios and real-world scenarios show that increasing the diversity and the size of the training set leads to the improvement of the RL agent's generalizability.

Scalable Evaluation of Multi-Agent Reinforcement Learning with Melting Pot

TLDR
This work creates over 80 unique test scenarios covering a broad range of research topics such as social dilemmas, reciprocity, resource sharing, and task partitioning, and demonstrates how Melting Pot reveals weaknesses not apparent from training performance alone.

The Surprising Effectiveness of MAPPO in Cooperative, Multi-Agent Games

TLDR
This work shows that PPO-based multi-agent algorithms achieve surprisingly strong performance in four popularMulti-agent testbeds: the particle-world environments, the StarCraft multi- agent challenge, the Hanabi challenge, and Google Research Football, with minimal hyperparameter tuning and without any domain-specific algorithmic modiflcations or architectures.

TrafficSim: Learning to Simulate Realistic Multi-Agent Behaviors

TLDR
This work proposes TrafficSim, a multi-agent behavior model for realistic traffic simulation that generates significantly more realistic traffic scenarios as compared to a diverse set of baselines and can exploit trajectories generated by TrafficSim as effective data augmentation for training better motion planner.

Emergent Road Rules In Multi-Agent Driving Environments

TLDR
Empirical evidence is provided that suggests that -- instead of hard-coding road rules into self-driving algorithms -- a scalable alternative may be to design multi-agent environments in which road rules emerge as optimal solutions to the problem of maximizing traffic flow.

SMARTS: Scalable Multi-Agent Reinforcement Learning Training School for Autonomous Driving

TLDR
The design goals of SMARTS (Scalable Multi-Agent RL Training School) are described, its basic architecture and its key features are explained, and its use is illustrated through concrete multi-agent experiments on interactive scenarios.

MADRaS : Multi Agent Driving Simulator

TLDR
MADRaS provides a platform for constructing a wide variety of highway and track driving scenarios where multiple driving agents can train for motion planning tasks using reinforcement learning and other machine learning algorithms.