NVlabs/alpasim

AlpaSim is an open-source autonomous vehicle simulation platform designed for development and testing of end-to-end AV policies

What it solves

AlpaSim provides a modular and lightweight simulation environment for autonomous vehicle (AV) research. It allows developers to test end-to-end driving policies in a closed-loop setting, enabling the validation of algorithms, safety analysis of edge cases, and performance benchmarking without needing a physical vehicle.

How it works

The platform uses a microservices architecture with a gRPC interface, allowing researchers to easily swap components or scale individual parts of the system across multiple nodes. It simulates realistic sensor data and vehicle dynamics, supporting pluggable renderers like NuRec and video-model rendering via OmniDreams and FlashDreams to provide high-fidelity camera feeds and environmental conditions.

Who it’s for

It is designed for researchers and developers working on autonomous driving algorithms, specifically those needing a hackable, scalable testbed to prototype and debug complex AV behaviors.

Highlights

  • High-Fidelity Rendering: Supports stateful video-model rendering for better visual fidelity of dynamic and non-rigid objects.
  • Modular Design: Python-based implementation with a pluggable architecture for rapid prototyping.
  • Horizontal Scalability: Distributed computing capabilities via microservices, supporting multi-node deployments.
  • Policy Support: Integrated support for various driving policies, including Alpamayo-R1, Alpamayo 1.5, VaVAM, and Transfuser.

Related

  • Project
    microsoft/AirSim

    An open-source simulator for drones and cars built on Unreal Engine, designed as a platform for AI research in deep learning, computer vision, and reinforcement learning.

  • Project
    Zihao-Felix-Zhou/UavNetSim

    A Python-based simulation platform for UAV communication networks that models physical, MAC, and network layers to test routing protocols and mobility models.

  • Project
    carla-simulator/carla

    An open-source simulator for autonomous driving research used to develop, train, and validate autonomous driving systems in simulated urban environments.

  • Project
    Thinklab-SJTU/Bench2Drive

    A closed-loop dataset and benchmark for autonomous driving systems using the CARLA simulator to evaluate driving scores, success rates, and specific driving abilities.

  • Project
    OpenDriveLab/SimScale

    A scalable simulation pipeline and co-training strategy for end-to-end autonomous driving that synthesizes high-fidelity reactive scenarios to improve model robustness.