Collaborative search by air-ground unmanned swarm, as a pivotal and efficient approach for intelligence gathering and disaster relief, highlights the critical role of search path planning in enhancing overall performance. Addressing the inefficiency resulting from insufficient collaboration between air and ground unmanned platforms in current research, this paper delves into the fundamental characteristics and challenges of collaborative search by air-ground unmanned swarm. This paper clarifies the objectives and constraints of path planning and introduces a method for collaborative search path planning based on the Learning Wolf Pack Algorithm (LWPA). This method initially constructs an optimization model that comprehensively considers area coverage, target detection probability, and search uncertainty. It integrates Distributed Model Predictive Control (DMPC) with the Distributed Constraint Optimization Problem (DCOP) framework, forming an architecture for real-time search path planning. To overcome the limitation of existing DCOP solution methods, which tend to get stuck in local optimal solutions, the LWPA employs a Q-learning mechanism for hierarchical learning and dynamically adjusts parameters to balance local refinement and global exploration. Experimental results demonstrate that this method offers significant advantages in improving search efficiency, coverage, and target detection rates, with an average area coverage of 99.28% and uncertainty as low as 0.86%. These results fully validate its effectiveness and superiority in search tasks in complex urban environments. Furthermore, tests on dynamic adaptability and scalability further verify the potential and value of this method in practical applications.

As defense technology develops, it is essential to study the pursuit–evasion (PE) game problem in hypersonic vehicles, especially in the situation where a head-on scenario is created. Under a head-on situation, the hypersonic vehicle’s speed advantage is offset. This paper, therefore, establishes the scenario and model for the two sides of attack and defense, using the twin delayed deep deterministic (TD3) gradient strategy, which has a faster convergence speed and reduces over-estimation. In view of the flight state–action value function, the decision framework for escape control based on the actor–critic method is constructed, and the solution method for a deep reinforcement learning model based on the TD3 gradient network is presented. Simulation results show that the proposed strategy enables the hypersonic vehicle to evade successfully, even under an adverse head-on scene. Moreover, the programmed maneuver strategy of the hypersonic vehicle is improved, transforming it into an intelligent maneuver strategy.

In order to overcome the drawbacks of the convergence time boundary dependent on tuning parameters in existing finite/fixed-time cooperative guidance law, this paper presents a three-dimensional prescribed-time pinning group cooperative guidance scheme that ensures multiple unpowered missiles to intercept multiple stationary targets. Firstly, combining a prescribed-time scaling function with pinning group consensus theory, the prescribed-time consensus-based cooperative guidance law is proposed. Secondly, the prescribed-time convergence of the proposed pinning group consensus-based cooperative guidance law proves that the convergence can be achieved at a specified time, regardless of initial conditions and parameters. Furthermore, the design steps including two stages of the proposed guidance law are given for engineering application. Extensive simulations are carried out in three cases to verify the properties. Simulation results show the effectiveness and superiority of the proposed prescribed-time consensus-based cooperative guidance scheme.

In view of the cooperative guidance problem with time delay, this paper proposes a two-stage time-delay prescribed-time cooperative guidance law in the three-dimensional (3D) space. In the first stage, by introducing a time scaling function and time-delay consensus, the proposed cooperative guidance law can overcome the negative influence of time delay to guaranteed the desired convergence performance. Derived from the Lyapunov convergence analysis, the time-delay stability of the first stage can be ensured and the convergence time can be described as the relationship between delayed time and mission-assigned convergence time. Then, taking the prescribed-time-related convergence time as the switching point, the second stage begins with suitable initial conditions and all interceptors are governed by proportional navigation guidance. Finally, comparative simulations are performed to demonstrate the effectiveness and superiority of the proposed time-delay guidance law.

Unmanned aircraft systems (UASs) with autonomous maneuvering decision capabilities are expected to play a key role in future unmanned systems applications. While reinforcement learning has proven successful in solving UAS path planning problems in simple urban environments, it remains under-researched for some complex mountain environments. In this paper, the path planning of UAS swarm for the low-altitude rapid traverse in diverse environments is studied when facing the threats of complex terrain, radars and swarm failure. First, a UAS swarm radar detection probability is built up for evaluating the radar detection threat by a networked radar system, where the detection probability of a UAS swarm is equated to a single UAS with appropriate position and radar cross section named as the swarm virtual leader. Second, a reinforcement learning based path planning method is proposed to seek the optimal path for the swarm virtual leader which balances instantaneous reward, including detection probability and path constraints with terminal reward, including normal rate. Third, a formation optimization strategy is designed to further reduce the threat of radar detection through dynamically adjusting the formation geometry. Final, simulations in the complex environment have been carried out to evaluate the performance of the proposed method, where the path quality, task success rate and normal rate are counted as the performance indicators.