Path planning is crucial in the automatic navigation of USVs (unmanned underwater vehicles), which directly affects the operational efficiency and safety of USVs. In this paper, we propose a path-planning algorithm based on DDPG (Deep Deterministic Policy Gradient) and make a detailed comparison with the traditional A-Star algorithm and the recent Actor–Critical algorithm. Through a series of simulation experiments, it can be observed that the optimal path for USVs found by the DDPG-based path planning algorithm is faster and more accurate than that found by the other two methods. The experimental results show that the DDPG algorithm has a significant advantage in processing time and better performance in terms of path quality and safety. These results provide a strong reference for future research on automatic navigation for USVs and demonstrate the potential of DDPG-based path planning for USVs.

This paper explores the significance and feasibility of addressing a notion that the system error of a nonlinear feedback control can be decorated by an inverse tangent function in order to attain a sound energy-efficient performance. The related mathematical model and relevant evaluation of this concept are further illustrated by demonstrating a case study about the control performance of water tank level. The rationale of robust control and theoretical algorithm of Lyapunov stability theorem are outlined to evaluate the effectiveness of nonlinear feedback with inverse tangent function in terms of improving robustness of PID (Proportional–Integral–Derivative) controller and energy-saving capability. By demonstrating five simulations of different scenarios, it ultimately proves that the modified robust PID controller by inverse tangent function meets the requirement of energy-saving capacity. Comparing with the routine PID control, the mean control input of controlling water tank level can be reduced up to 39.2% by using modified nonlinear feedback controller. This nonlinear feedback PID controller is energy efficient and concise for its convenient use, which is feasible to expand its utility to other applications.

The rudder is used for damping the roll motion of a ship. Study of this motion is the essence of this paper. The responses of yaw and roll motions are different as the rudder angle varies. The low frequency motion of the rudder mainly affects the ship’s yaw motion, whereas the high frequency motion of the rudder mainly affects the roll motion. As long as the controller is well designed, the characteristic of the rudder can be used to reduce the roll motion. In order to save energy and reduce steer frequency, a nonlinear feedback controller is proposed based on the arc tangent function processing feedback error to save energy consumption. In addition, a ZOH component is applied in the system to reduce the steer frequency. In heavy sea state, simulation results illustrate that controllers based on pole assignment with and without nonlinear feedback can reduce roll motion by 32.1% and 30.3%, respectively, when the rudder turn rate is limited within 7°/s. Furthermore, the former reduces the amplitude of rudder angle by 23.3% compared with the latter, which means the nonlinear feedback control consumes less energy. Consequently, the ZOH can lower steer frequency to once every 1 s, which protects steering gear from abrasion.

Automatic route planning and tracking control are important for unmanned surface vehicles (USVs). This study presents a final-optimized A* (FO A*) algorithm for automatic route planning of USVs and designs a nonlinear tracking controller to ensure that the ship navigates on the route. Ship domain theory is employed in the FO A* algorithm to ensure that the planning route is far from shorelines and reefs. Meanwhile, water depth is used to determine the safety area in the search process. The nonlinear feedback algorithm is also applied to reduce the rudder angle when the ship tracks the planning route in wind and waves. Finally, the proposed algorithms are well validated on the ship-based Electronic Chart Display and Information System (ECDIS). The experiment shows that the FO A* algorithm and the nonlinear tracking controller are effective and deterministic for route planning offline. It can be put into navigation practice, which will make some contribution in accelerating the USV development.

The goal of 6G is to achieve ubiquitous connectivity in heterogeneous and large-scale networks, and Hierarchical Federated Learning (HiFL) is an effective approach to assist AI technologies at the network edge. However, the “waist -shaped client-channel-cloud” architecture of maritime communication network greatly limits the implementation of hierarchical federated edge learning. Considering the aforementioned limitations, a hierarchical federated data distillation method based on UAV-assisted MEC systems is proposed, called Data Augmentation Distilled Federated Learning (DADFL). Specifically, DADFL reduce the communication process between clients and edge servers to just one round. Each client extracts their private dataset, sends synthesized data to the server, and collectively trains a global model. Building upon this, we propose a Global Optimal Iterative Search Algorithm (GOISA) considering client heterogeneity and resource constraints. GOISA minimizes overall latency by optimizing client heterogeneity, bandwidth allocation, and UAV transmission rates. Simulation experiments demonstrate that our proposed system achieves performance comparable to or even better than traditional federated learning, while reducing communication overhead by three orders of magnitude, greatly improving communication efficiency.