Training and Test Traffic Scenarios

该数据集包含了用于训练和评估HR-PPO智能体在自动驾驶领域性能的交通场景。其中，训练场景总计200,200个，测试场景则有10,000个未见过的场景。这些场景在车辆数量上有所不同，从1辆到58辆不等，平均每个场景有12辆车。为了减少随机性，每个场景都进行了15次采样。整个数据集的规模包括200,200个训练场景和10,000个测试场景，其任务旨在进行自动驾驶智能体的训练与评估。This dataset comprises traffic scenarios for training and evaluating the performance of HR-PPO AI Agents in the autonomous driving domain. Specifically, it includes 200,200 training scenarios and 10,000 unseen test scenarios. The number of vehicles in these scenarios ranges from 1 to 58, with an average of 12 vehicles per scenario. To reduce randomness, each scenario is sampled 15 times. This dataset is dedicated to the training and performance evaluation of autonomous driving AI agents.