该数据集建立于2025年3月，利用Python仿真平台构建生成，主要面向动态未知环境下的人机协同无人集群空战控制问题，旨在为多智能体强化学习、人类意图嵌入、层级元命令决策与OOD泛化能力评估提供数据支撑。数据集基于Harfang三维空战仿真平台生成，主要包含6v6训练场景，以及6v6、8v8、12v12、20v20等对称场景和6v8、8v12、12v20等非对称OOD测试场景。数据集包含一个txt文本文件，即replay_bufer.txt。replay_bufer.txt为无人集群执行任务过程中的状态、动作、奖励等强化学习要素数据，包含4列数据，分别为无人集群状态(state)、无人集群个体动作(action),团队奖励函数值(reward)、无人集群下一状态(next state)以及当前任务是否完成标志(done)。数据内容包括无人机状态、局部观测、连续控制动作、元命令选择、人类偏好值、奖励反馈、任务终止信息，以及胜率、毁伤比和平均回合时长等评估指标，可用于支撑人机协同集群控制策略训练与跨场景泛化性能测试。
", "ds_source": "本数据集为完全仿真生成数据，不源自特定文献、实测或第三方下载。
", "ds_process_way": "本数据集基于Harfang三维仿真环境和人机协同多智能体强化学习交互过程生成。首先初始化红蓝双方无人机的位置、速度、姿态和健康状态，并在仿真过程中依据飞行动力学和近距空战规则更新状态、执行攻击判定并记录交互轨迹。随后，将高层战术意图加工为“攻击/规避—目标组”的元命令，并结合人类对不同元命令的偏好值，形成带有人类意图标注的决策样本。最后，通过6v6训练场景和不同敌方策略、不同规模配置下的OOD测试场景生成多样化数据，并统计胜率、毁伤比和平均回合时长等指标，形成用于层级人机协同集群控制研究的结构化数据。
", "ds_quality": "该数据呈现了无人集群子任务执行的详细数据。
", "ds_acq_start_time": "2025-01-01 00:00:00", "ds_acq_end_time": null, "ds_acq_place": "合肥", "ds_acq_lon_east": 117.43, "ds_acq_lat_south": 31.57, "ds_acq_lon_west": 117.07, "ds_acq_lat_north": 31.97, "ds_acq_alt_low": 4.0, "ds_acq_alt_high": 80.0, "ds_share_type": "open-access", "ds_total_size": 150751480, "ds_files_count": 0, "ds_format": ".txt", "ds_space_res": "", "ds_time_res": "", "ds_coordinate": "无", "ds_projection": "", "ds_thumbnail": "0428e11a-ce1e-467a-a721-6ed121c5f0ca.png", "ds_thumb_from": 0, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "", "ds_from_station": "", "organization_id": "9ecaaa78-39e9-411e-9f24-274e12aa643f", "ds_serv_man": "虞文武", "ds_serv_phone": "15051861330", "ds_serv_mail": "wwyu@seu.edu.cn", "doi_value": "", "subject_codes": [ "410" ], "quality_level": 0, "publish_time": "2026-07-09 10:58:45", "last_updated": "2026-07-09 10:58:45", "protected": false, "protected_to": "2028-06-30 00:00:00", "lang": "zh", "cstr": "", "i18n": { "en": { "title": "Human-Machine Collaborative Cluster Air Combat Generalization Dataset", "ds_format": ".txt", "ds_source": "This dataset is completely synthetically generated through simulation and does not originate from specific literature, field measurements, or third-party downloads.", "ds_quality": "This dataset presents detailed information on the execution of sub-tasks by unmanned swarms.", "ds_ref_way": "", "ds_abstract": "This dataset was established in March 2025 and generated using a Python simulation platform. It is primarily designed for human-machine collaborative unmanned swarm air combat control in dynamic unknown environments, aiming to provide data support for multi-agent reinforcement learning, human intention embedding, hierarchical meta-command decision-making, and OOD generalization capability evaluation. The dataset is generated based on the Harfang 3D air combat simulation platform and mainly includes 6v6 training scenarios, as well as symmetric scenarios such as 6v6, 8v8, 12v12, and 20v20, and asymmetric OOD test scenarios such as 6v8, 8v12, and 12v20. The dataset contains a single TXT text file, namely replay_buffer.txt. This file records the reinforcement learning elements—states, actions, rewards, etc.—during the unmanned swarm's mission execution. It comprises four columns of data: the swarm state, individual swarm actions, the team reward function value, the next swarm state, and a flag indicating whether the current task is completed (done). The data content includes UAV states, local observations, continuous control actions, meta-command selections, human preference values, reward feedback, task termination information, as well as evaluation metrics such as win rate, damage ratio, and average episode length. The dataset can support training of human-machine collaborative swarm control policies and cross-scenario generalization performance testing.", "ds_time_res": "", "ds_acq_place": "Hefei", "ds_space_res": "", "ds_projection": "", "ds_process_way": "This dataset was generated using the Harfang 3D simulation environment and human-machine collaborative multi-agent reinforcement learning interaction processes. Initially, the positions, velocities, attitudes, and health statuses of the Red and Blue team UAVs are initialized; during the simulation, these states are updated—and attack outcomes determined—based on flight dynamics and close-range aerial combat rules, while interaction trajectories are recorded. Subsequently, high-level tactical intentions are processed into \"attack/evade—target group\" meta-commands; these are combined with human preference values for various meta-commands to create decision-making samples annotated with human intent. Finally, diverse data are generated through 6v6 training scenarios and out-of-distribution (OOD) testing scenarios featuring varying enemy strategies and configuration scales. Metrics such as win rates, kill-loss ratios, and average episode durations are calculated to form a structured dataset suitable for research on hierarchical human-machine collaborative swarm control.", "ds_ref_instruction": "" } }, "submit_center_id": "ncdc", "data_level": 0, "recommendation_value": 0, "license_type": "https://creativecommons.org/licenses/by/4.0/", "doi_reg_from": "reg_local", "cstr_reg_from": "reg_local", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "belong_to_nieer": false, "ds_topic_tags": [ "协同拦截", "分层强化学习" ], "ds_subject_tags": [ "工程与技术科学基础学科" ], "ds_class_tags": [], "ds_locus_tags": [], "ds_time_tags": [], "ds_contributors": [ { "true_name": "虞文武", "email": "wwyu@seu.edu.cn", "work_for": "东南大学数学学院", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "虞文武", "email": "wwyu@seu.edu.cn", "work_for": "东南大学数学学院", "country": "中国" } ], "ds_managers": [ { "true_name": "虞文武", "email": "wwyu@seu.edu.cn", "work_for": "东南大学数学学院", "country": "中国" } ], "category": "其他" }