data for "UAV Path Planning Based on Proximal Policy Optimization Algorithm with Long Short-Term Memory Networks and Generalized Integral Compensator"

This study proposes a more competitive and sample-efficient algorithm: Memory-GIC-PPO, specifically to address POMDPs in UAV path planning. The effectiveness of the proposed algorithm is thoroughly evaluated through simulations conducted on the Airsim platform. The results convincingly demonstrate that Memory-GIC-PPO enables the UAV to achieve optimal path planning in complex environments and outperforms the benchmark algorithms in terms of sampling efficiency and success rates. This dataset contains the rewards obtained by the Memory-GIC-PPO, LSTM-TD3, and PPO algorithms in the designed scenarios, as well as their respective success rates. In addition, the rewards and success rates associated with the results of the ablation experiments were also demonstrated in the dataset. 

本研究提出了一种兼具更强竞争力与更优样本效率的算法：Memory-GIC-PPO，专门用于解决无人机（Unmanned Aerial Vehicle，UAV）路径规划场景下的部分可观察马尔可夫决策过程（Partially Observable Markov Decision Processes，POMDPs）。研究团队通过在Airsim仿真平台上开展的多组仿真实验，对所提算法的有效性进行了全面评估。实验结果充分证明，Memory-GIC-PPO可使无人机在复杂环境中实现最优路径规划，且在采样效率与任务成功率两大指标上均优于基准算法。本数据集收录了Memory-GIC-PPO、LSTM-TD3以及近端策略优化（Proximal Policy Optimization，PPO）算法在预设实验场景中获得的奖励值，以及各算法对应的任务成功率。此外，数据集还包含了消融实验结果所对应的奖励值与任务成功率数据。