<b>Urban Multi-UAV Path Planning Simulation Dataset (2-D Dynamic Urban MEC Scenarios)</b>

This dataset accompanies the manuscript “PAIR: A Hybrid A* with PPO Path Planner for Multi-UAV Navigation in 2-D Dynamic Urban MEC Environments” by Bahaa Hussien Taher et al., submitted to the MDPI journal <i>Drones</i>. It provides a reproducible benchmark for evaluating multi-UAV path planning algorithms in cluttered, dynamic urban airspace.The dataset consists of nine procedurally generated 2-D urban scenarios on a 100×100 grid. Each scenario includes (i) static obstacles with three density levels (approximately 15%, 25%, and 40%), (ii) two spatial regimes (“clustered” and “dispersed”), (iii) static or Markovian dynamic obstacles (Move / Pause / Detour states), and (iv) start–goal positions for nine UAVs. Dynamic scenarios contain trajectories for up to 30 moving obstacles over 1,000 time steps.Data are provided as CSV tables (scenarios.csv, static_obstacles.csv, uav_positions.csv, dynamic_obstacles.csv) and optionally as a single JSON file aggregating all scenarios. These files allow users to benchmark classical planners (A*, D* Lite, CBS–D*), metaheuristics (PSO), and learning-based planners (e.g., PPO-based PAIR) on metrics such as mission success rate, travel time, energy surrogate, and unified path-quality score.A plain-text README file in the repository documents the file structure, field definitions, recommended Python environment, and an example workflow for loading the data and running simulations.本数据集配套于Bahaa Hussien Taher等人提交至MDPI期刊《Drones》的论文《PAIR：面向二维动态城市多接入边缘计算（Multi-Access Edge Computing，MEC）环境的混合A*与近端策略优化（Proximal Policy Optimization，PPO）路径规划器的多无人机（Multi-UAV）导航算法》。本数据集为杂乱动态城市空域中的多无人机路径规划算法评估提供了可复现的基准测试集。数据集包含9个基于100×100网格程序生成的二维城市场景。每个场景涵盖以下内容：(i) 三种密度等级（约15%、25%与40%）的静态障碍物；(ii) 两种空间分布类型：“集群式”与“离散式”；(iii) 静态或马尔可夫动态障碍物（包含移动、暂停与绕行三种状态）；(iv) 9架无人机的起始-目标位置。动态场景中包含最多30个移动障碍物在1000个时间步长内的运动轨迹。数据以逗号分隔值（Comma-Separated Values，CSV）表格文件（scenarios.csv、static_obstacles.csv、uav_positions.csv、dynamic_obstacles.csv）形式提供，同时可选择包含所有场景的单一JSON聚合文件。上述文件可支持研究人员针对经典路径规划器（如A*、D* Lite、CBS–D*）、元启发式算法（粒子群优化（Particle Swarm Optimization，PSO））以及基于学习的路径规划器（例如基于PPO的PAIR）开展基准测试，评估指标涵盖任务成功率、航行时间、能耗代理指标与统一路径质量评分。该数据集仓库附带纯文本README文档，其中说明了文件结构、各字段定义、推荐的Python运行环境，以及加载数据与运行仿真的示例工作流程。