In this study, we propose an aerial rendezvous method to facilitate the recovery of unmanned aerial vehicles (UAVs) using carrier aircrafts, which is an important capability for the future use of UAVs. The main contribution of this study is the development of a promising method for online generation of feasible rendezvous trajectories for UAVs. First, the wake vortex of a carrier aircraft is analyzed using the finite element method, and a method for establishing a safety constraint model is proposed. Subsequently, a model-reference reinforcementearning algorithm is proposed based on the potential function method, which can ensure the convergence and stability of training. A combined reward function is designed to solve the UAV trajectory generation problem under non-convex constraints. The simulation results show that, compared with the traditional artificial potential field method under different working conditions, the success rate of this method under non-convex constraints is close to 100%, with high accuracy, convergence, and stability, and has greater application potential in the aerial recovery scenario, providing a solution to the trajectory generation problem of UAVs under non-convex constraints.

Unmanned aerial vehicles play a significant role in the automated inspection of future smart cities, which can ensure the safety of urban residents’ lives and property and the normal operation of the city. However, there may be situations where small targets in drone images are difficult to detect and the detection is unclear when the targets are similar to the environment. In response to the above problems, this paper proposes a real-time target detection method for unmanned aerial vehicle images based on Transformer. Aiming at the problem of small targets lacking visual features, a feature fusion module was designed, which realizes the interaction and fusion of features at different levels and improves the feature expression ability of small targets. Aiming at the problem of discontinuous features when the target is similar to the environment, a multi-head attention algorithm based on Transformer is designed. By extracting the context information of the target, the recognition ability of targets similar to the environment is improved. On the target image dataset collected by unmanned aerial vehicles in smart cities, the detection accuracy of the method described in this paper has reached 85.9%.

The effectiveness of autonomous decision-making of unmanned aerial vehicles (UAVs) exceeds that of a single one. However, existing methods face critical gaps in within-visual-range (WVR) combat: single-step strategies lack decision depth and precision; non-cooperative multi-UAV frameworks lose tactical advantages-both conflicting with WVR's stringent demands for rapid response and high precision. To bridge these issues, an UAV autonomous decision-making model is proposed. This model incorporates several key components, including a dynamics model with overload control, a relative motion model, an overload maneuver library, and a situation assessment (SA) model. To systematically address the decision-making challenge in WVR engagements, a two-step strategy is adopted to design a 1v1 decision-making method. Specifically, in the first step, preliminary maneuver selection is made using an improved matrix decision-making method, which incorporates a Kalman filter for improved accuracy. Subsequently, the second step involves reoptimizing the decision using an improved genetic algorithm (IGA). Building upon this 1v1 decision-making framework, a 2v1 cooperative decision-making method is proposed-leveraging distributed cooperative principles and information exchange mechanisms. The simulation results show that, compared to the single-step strategy method, the proposed two-step strategy matrix decision-making method significantly reduces the time required to lock onto the enemy and increases tactical advantage. In addition, the proposed cooperative decision-making method outperforms non-cooperative approaches, both in terms of locking time and advantage. The proposed cooperative decision-making framework for UAVs enables significantly enhance efficiency and precision, offering advancements in autonomous operations and real-time decision-making systems.

An effective solution method is proposed for the scanning path planning problem of heterogeneous multi-agent systems. Target mission is divided into multiple task clusters by density clustering to balance the task difficulty. Afterwards, the method innovatively first performs path planning for each task cluster, then systematically allocates tasks by utilizing mixed-integer linear programming (MILP) based on the planning outcomes and the diverse capabilities within the heterogeneous system, ensuring that agents possess unique adaptability to tasks that align with their specific capabilities. In the optimization solving phase, a neural combinatorial optimization (NCO) algorithm based on graph neural network (GNN) is proposed. Through supervised learning and cross-instance parallel training, the automation of heuristic design and the enhancement of heuristic rules are achieved, avoiding the inefficiency and suboptimality of expert design and manual parameter tuning. Simulation results show that the algorithm has better performance in terms of task completion time, execution time and deviation rate, showing its application potential in the rapid task planning problem of heterogeneous multi-agent systems.