Reacher and Tracker Environments

该数据集专为训练智能体使用深度强化学习而设计，包含了一系列针对UR10e机器人手臂进行到达和跟踪任务的环境。此外，数据集还包含了PPO和DQN算法的性能指标，如每集步骤的平均奖励及其标准误，以及用于评估的剧集长度图表。该数据集的规模涵盖了多次试验（每个智能体独立进行3次试验），其任务重点是机器人控制的强化学习。

This dataset is specifically designed for training AI Agents with deep reinforcement learning, and includes a series of environments for reaching and tracking tasks using the UR10e robotic arm. Additionally, the dataset contains performance metrics of PPO and DQN algorithms, such as the average reward per episode and its standard error, as well as charts of episode lengths for evaluation purposes. The dataset covers multiple trials (3 independent trials per AI Agent), with its tasks focusing on reinforcement learning for robotic control.