Optimization of multi-UAV systems based on differential game theory and sliding mode control

Purpose This study aims to optimize control strategies for multi-unmanned aerial vehicle (UAV) systems by integrating differential game theory with sliding mode control and neural networks. This approach addresses challenges in dynamic and uncertain environments, enhancing UAV system coordination, operational stability and precision under varying flight conditions. Design/methodology/approach The methodology combines sliding mode control, differential game theory and neural network algorithms to devise a robust control framework for multi-UAV systems. Using a nonsingular fast terminal sliding mode observer and Nash equilibrium concepts, the approach counters external disturbances and optimizes UAV interactions for complex task execution. Findings Simulations demonstrate the effectiveness of the proposed control strategy, showcasing enhanced stability and robustness in managing multi-UAV operations. The integration of neural networks successfully solves high-dimensional Hamilton–Jacobi–Bellman equations, validating the precision and adaptability of the control strategy under simulated external disturbances. Originality/value This research introduces a novel control framework for multi-UAV systems that uniquely combines differential game theory, sliding mode control and neural networks. The approach significantly enhances UAV coordination and operational stability in dynamic environments, providing a robust solution to high-dimensional control challenges. The use of neural networks to solve complex Hamilton–Jacobi–Bellman equations for real-time multi-UAV management represents a groundbreaking advancement in autonomous aerial vehicle research.