Autonomous Racing Simulation Dataset
收藏资源简介:
Autonomous Racing Simulation Dataset是由加州大学圣地亚哥分校等机构创建的高保真赛车模拟数据集,旨在为自动驾驶赛车算法提供测试和验证。数据集包含6400万步的训练数据和超过900圈的人类驾驶员数据,涵盖多种车辆和赛道。数据集的创建过程利用了Assetto Corsa模拟器的插件接口,记录了车辆的实时状态和控制数据。该数据集主要应用于自动驾驶赛车的研究和开发,特别是强化学习(RL)和模型预测控制(MPC)算法的评估和优化。
The Autonomous Racing Simulation Dataset is a high-fidelity racing simulation dataset developed by the University of California, San Diego and other research institutions, aiming to provide testing and validation support for autonomous racing algorithms. This dataset contains 64 million steps of training data and over 900 laps of human driver data, covering a wide range of vehicle types and racing tracks. The dataset was constructed using the plugin interface of the Assetto Corsa simulator, which records real-time vehicle status and control data. It is primarily utilized for the research and development of autonomous racing technologies, particularly for the evaluation and optimization of reinforcement learning (RL) and model predictive control (MPC) algorithms.

- 1A Simulation Benchmark for Autonomous Racing with Large-Scale Human Data加州大学圣地亚哥分校, 格拉茨理工大学, 意大利摩德纳大学, Know-Center GmbH · 2024年



