Drone Path Planning Simulation Dataset

该数据集包含了在受限的室内Gazebo环境中模拟的无人机导航场景，涵盖了多种配置和障碍物设置，旨在评估使用预训练基础模型进行路径规划的效果。此外，数据集还包括了成功率（SR）、平均完成时间（ACT）、路径长度（PL）以及效率-成功率得分（ESS）等指标，用于评估不同语言模型在路径规划方面的性能。该数据集所涉及的空间规模为具有静态立方障碍物（2米 x 2米 x 2米）的三维空间，任务重点是自主无人机路径规划。This dataset comprises drone navigation scenarios simulated in a constrained indoor Gazebo environment, with diverse configurations and obstacle arrangements, intended to evaluate the efficacy of path planning using pre-trained foundation models. Furthermore, the dataset includes metrics including Success Rate (SR), Average Completion Time (ACT), Path Length (PL), and Efficiency-Success Score (ESS) for benchmarking the performance of various language models in path planning tasks. The spatial scale involved in this dataset is a 3D environment with static cubic obstacles measuring 2m × 2m × 2m, and the core task is autonomous drone path planning.

搜集汇总

背景与挑战

背景概述

该数据集用于无人机路径规划模拟，包含标准化场景以支持基准测试研究，旨在评估基础模型在自主导航中的性能。数据集与FM-Planner项目关联，侧重于实际应用验证。