PDE-Based Neuro Adaptive Control for Multi-Agent Deployment With Non-Collocated Observer

Abstract

A neuro adaptive control for deploying a partial differential equation-based multi-agent system in 3D space with a non-collocated observer is proposed in this study. Since the full states of the system are unavailable in practice, an observer-based control is developed to ensure stability of the underlying closed-loop system. In addition, the system uncertainty is addressed by introducing a neural network control. By choosing appropriate system parameters, the desired control objectives can be achieved. The proposed strategy is simple to implement and its implementation condition is easily satisfied. Finally, the effectiveness of the designed method is verified by the simulation results. Note to Practitioners— In this paper, we introduce a neuro-adaptive control strategy for a multi-agent system based on partial differential equations in 3D space, using a non-collocated observer. This approach is particularly relevant for practitioners dealing with dynamic and uncertain environments in control systems. Neural networks are employed to manage system uncertainties, adapting to changing conditions, which is crucial in environments with variable system parameters. The non-collocated observer allows for state estimation, beneficial in situations where direct measurement is impractical. Ensuring the observer’s accuracy is key for effective control. Our strategy focuses on simplicity and ease of implementation, making it accessible for integration into existing systems. The observer-based control ensures the stability of the closed-loop system, a critical factor for consistent performance.