TL;DR
This paper introduces a novel cooperative control strategy for Unmanned Surface Vessels (USVs) and Unmanned Aerial Vehicles (UAVs) to jointly follow a path. It leverages a new "Logic Virtual Ship-Logic Virtual Aircraft" (LVS-LVA) guidance principle and a robust adaptive neural control system that reduces communication load through an event-triggered mechanism, demonstrating effective path following even with disturbances.
What Problem Does This Paper Address?
Coordinating the actions of different types of unmanned vehicles, like surface vessels and aerial drones, for tasks such as maritime surveillance or search and rescue, presents a significant challenge. While USVs offer endurance and payload capacity, UAVs provide speed and superior observation. Combining their strengths through cooperative control can greatly enhance mission effectiveness. However, achieving this requires a robust system that can handle the distinct dynamics of each platform, uncertainties in their models, and external disturbances, all while minimizing communication overhead. Existing guidance principles often focus on single agents or homogeneous multi-agent systems, leaving a gap for heterogeneous USV-UAV cooperation.
Methods
The core of this work lies in two main innovations:
- LVS-LVA Guidance Principle: This novel guidance framework establishes a crucial link between the USV and the UAV. It uses a "Logic Virtual Ship" to define the reference path and then maps this to a "Logic Virtual Aircraft" to generate coordinated reference signals for the UAV. This allows the USV and UAV to navigate together along a desired route.
- Robust Adaptive Neural Cooperative Control: To manage the complex dynamics and uncertainties, the authors employ a sophisticated control architecture. This includes:
- Dynamic Surface Control (DSC): Used to manage the complexity of control design and avoid "explosion of terms" when dealing with cascaded systems.
- Radial Basis Function Neural Networks (RBF-NNs): These are employed to approximate unknown nonlinear functions and model uncertainties within the USV-UAV system.
- Event-Triggered Technique: This is a key contribution for reducing communication load. Instead of sending control commands continuously, the system only updates the control signal when a predefined error threshold is crossed. This significantly reduces the transmission burden between the controller and actuators.
The control law is designed to directly manipulate the steerable variables: the main engine speed and rudder angle for the USV, and the rotor angular velocities for the UAV.
Datasets
Not discussed in the paper. The evaluation relies on numerical simulations using a mathematical model of the USV-UAV system.
Key Insights and Ideas
- Bridging Heterogeneous Systems: The LVS-LVA principle is a clever way to create a unified guidance framework for two fundamentally different platforms (USV and UAV), enabling them to act in concert.
- Communication Efficiency: The event-triggered approach is crucial for practical deployments. By reducing the frequency of control updates, it significantly lowers the communication bandwidth requirements, which is vital in remote or congested operational environments.
- Robustness through Adaptation: The combination of RBF-NNs for uncertainty approximation and adaptive control ensures that the system can maintain performance even when faced with unknown disturbances and model variations.
- Simplified Control Inputs: The paper focuses on directly controlling the physical actuators (engine speed, rudder, rotor speeds), making the control design more directly applicable to real-world systems compared to abstract thrust or torque commands.
Key Results
The paper demonstrates the effectiveness of the proposed robust adaptive neural cooperative control algorithm through numerical simulations. Key findings include:
- Successful Path Following: The USV and UAV were shown to navigate along the desired cooperative path with satisfactory control performance.
- Reduced Control Burden: The event-triggered control mechanism resulted in "stair-step" control inputs, significantly reducing the continuous transmission load compared to a non-event-triggered approach.
- Error Convergence: Tracking errors for the USV-UAV cooperative control system were shown to converge to a small compact set, indicating good tracking accuracy.
- Guaranteed Stability: The authors theoretically guarantee that all signals in the closed-loop system are semi-globally uniformly ultimately bounded (SGUUB) by selecting appropriate control and adaptive parameters.
- Disturbance Rejection: The simulations were conducted in the presence of time-varying disturbances (wind and wave), and the proposed algorithm demonstrated robustness.
Limitations and Future Work
The authors acknowledge several limitations and suggest future research directions:
- Parameter Tuning: Adjusting the control parameters and, particularly, the triggered threshold parameters can be a challenging task to optimize performance and transmission load.
- High-Speed Operation: The paper notes that while the algorithm is designed for path following, operating at very high speeds might lead to instability, suggesting that reference speed design needs careful consideration.
- Multi-UAVs-USVs Formation Control: The current work focuses on a single USV and UAV. Extending this to control formations of multiple UAVs and USVs presents a significant future challenge.
- Real-World Experiments: While simulations are presented, the ultimate validation would involve implementing and testing the algorithm on actual USV and UAV platforms.
Summary updated