The reconnaissance of high-value targets is prerequisite for effective operations. The recent appreciation of deep reinforcement learning (DRL) arises from its success in navigation problems, but due to the competitiveness and complexity of the military field, the applications of DRL in the military field are still unsatisfactory. In this paper, an end-to-end DRL-based intelligent reconnaissance mission planning is proposed for dual unmanned aerial vehicle (dual UAV) cooperative reconnaissance missions under high-threat and dense situations. Comprehensive consideration is given to specific mission properties and parameter requirements through the whole modelling. Firstly, the reconnaissance mission is described as a Markov decision process (MDP), and the mission planning model based on DRL is established. Secondly, the environment and UAV motion parameters are standardized to input the neural network, aiming to deduce the difficulty of algorithm convergence. According to the concrete requirements of non-reconnaissance by radars, dual-UAV cooperation and wandering reconnaissance in the mission, four reward functions with weights are designed to enhance agent understanding to the mission. To avoid sparse reward, the clip function is used to control the reward value range. Finally, considering the continuous action space of reconnaissance mission planning, the widely applicable proximal policy optimization (PPO) algorithm is used in this paper. The simulation is carried out by combining offline training and online planning. By changing the location and number of ground detection areas, from 1 to 4, the model with PPO can maintain 20% of reconnaissance proportion and a 90% mission complete rate and help the reconnaissance UAV to complete efficient path planning. It can adapt to unknown continuous high-dimensional environmental changes, is generalizable, and reflects strong intelligent planning performance.

In recent years, significant progress has been made in the multi-target tracking (MTT) of unmanned aerial vehicle (UAV) swarms. Most existing MTT approaches rely on the ideal assumption of a pre-set target trajectory. However, in practice, the trajectory of a moving target cannot be known by the UAV in advance, which poses a great challenge for realizing real-time tracking. Meanwhile, state-of-the-art multi-agent value-based methods have achieved significant progress for cooperative tasks. In contrast, multi-agent actor-critic (MAAC) methods face high variance and credit assignment issues. To address the aforementioned issues, this paper proposes a learning-based factored multi-agent soft actor-critic (FMASAC) scheme under the maximum entropy framework, where the UAV swarm is able to learn cooperative MTT in an unknown environment. This method introduces the idea of value decomposition into the MAAC setting to reduce the variance in policy updates and learn efficient credit assignment. Moreover, to further increase the detection tracking coverage of a UAV swarm, a spatial entropy reward (SER), inspired by the spatial entropy concept, is proposed in this scheme. Experiments demonstrated that the FMASAC can significantly improve the cooperative MTT performance of a UAV swarm, and it outperforms existing baselines in terms of the mean reward and tracking success rates. Additionally, the proposed scheme scales more successfully as the number of UAVs and targets increases.

Rapid and precise air operation mission planning is a key technology in unmanned aerial vehicles (UAVs) autonomous combat in battles. In this paper, an end‐to‐end UAV intelligent mission planning method based on deep reinforcement learning (DRL) is proposed to solve the shortcomings of the traditional intelligent optimization algorithm, such as relying on simple, static, low‐dimensional scenarios, and poor scalability. Specifically, the suppression of enemy air defense (SEAD) mission planning is described as a sequential decision‐making problem and formalized as a Markov decision process (MDP). Then, the SEAD intelligent planning model based on the proximal policy optimization (PPO) algorithm is established and a general intelligent planning architecture is proposed. Furthermore, three policy training tricks, i.e., domain randomization, maximizing policy entropy, and underlying network parameter sharing, are introduced to improve the learning performance and generalizability of PPO. Experiments results show that the model in this work is efficient and stable, and can be adapted to the unknown continuous high‐dimensional environment. It can be concluded that the UAV intelligent mission planning model based on DRL has powerful intelligent planning performance, and provides a new idea for researching UAV autonomy.

Aiming at the problems of difficult handling of three-dimensional flight conflicts and unfair distribution of resolution costs, we propose a multi-aircraft conflict resolution method based on the network cooperative game. Firstly, we establish a flight conflict network model with aircraft as nodes and the conflict relationship between node pairs as edges. After that, we propose a comprehensive network index that can evaluate the effect of resolution strategy. Based on the concept of “nucleolus solution”, we establish a conflict network alliance with all nodes as participants, and balance the interests of all participants through the resolution cost function. In order to improve the timeliness of the method, we propose two optimization methods: adjusting high-priority nodes and customizing the initial resolution scheme. Finally, we combine the NSGA-II algorithm to solve the optimal conflict resolution scheme. The simulation results show that our method can adjust 10 aircraft in 15.17 s and resolve 12 flight conflicts in a complex conflict scenario containing 40 aircraft; our method reduces the resolution cost by more than 22.1% on average compared with the method without considering the resolution cost. The method ensures both the conflict resolution capability and the reduction in resolution cost.

Cooperative path planning for unmanned aerial vehicle (UAV) swarms has attracted considerable research attention, yet it remains challenging in complex, uncertain environments. To tackle this problem, we model the cooperative path planning task as a heterogeneous decentralized Markov Decision Process (MDP), emphasizing the symmetry-inspired role assignment between leader and wingmen UAVs, which ensures balanced and coordinated behaviors in dynamic settings. We address the problem using a Multi-Agent Soft Actor-Critic (MASAC) framework enhanced with a symmetry-aware reward mechanism designed to optimize multiple cooperative objectives: simultaneous arrival, formation topology preservation, dynamic obstacle avoidance, trajectory smoothness, and inter-agent collision avoidance. This design promotes behavioral symmetry among agents, enhancing both coordination efficiency and system robustness. Simulation results demonstrate that our method achieves efficient swarm coordination and reliable obstacle avoidance. Quantitative evaluations show that our MASAC-CA algorithm achieves a Mission Success Rate (MSR) of 99.0% with 2–5 wingmen, representing approximately 13% improvement over baseline MASAC, while maintaining Formation Keeping Rates (FKR) of 59.68–26.29% across different swarm sizes. The method also reduces collisions to near zero in cluttered environments while keeping flight duration, path length, and energy consumption at levels comparable to baseline algorithms. Finally, the proposed model’s robustness and effectiveness are validated in complex uncertain environments, underscoring the value of symmetry principles in multi-agent system design.

The self-organization capabilities of massive aerial swarms pose challenges to conventional methods of grouping aerial targets. These traditional approaches struggle with issues such as stability, reliability, and the recognition of deep intentions. To overcome these challenges, we propose an architecture called Unsupervised Contrastive Learning for Aerial Targets Grouping (UCL-ATG). Our approach utilizes random time periods within the scenario as training batches for contrastive learning. The UCL-ATG model consists of four key modules: feature extraction, time series generation, fine-tuning, and clustering. The feature extraction and time series generation modules form the network architecture of Contrastive Predictive Coding (CPC). Positive and negative samples are obtained using different sampling methods. To effectively collaborate with the clustering module, we design a contrastive predictive loss function specifically tailored for clustering. This enables the extraction of low-dimensional representations that capture the high-dimensional temporal characteristics of aircraft. Furthermore, we introduce a real-time fine-tuner to enhance the model's transferability to specific tasks. The inclusion of the real-time fine-tuner greatly alleviates problems associated with evolving confrontation styles and unevenly distributed datasets, as confirmed by the performance on the validation set. Extensive comparative experimental results demonstrate our model's superior training outcomes and its ability to extract improved clustering features. In wargame applications, our model no longer relies solely on static position information of the aircraft. It exhibits outstanding capabilities in historical information memory, information synthesis, and even demonstrates aptitude in identifying friend or foe and making tactical inferences.