Underwater target detection is the foundation and guarantee for the autonomous operation of underwater vehicles and is one of the key technologies in marine exploration. Due to the complex and special underwater environment, the detection effect is poor, and the detection precision is not high. In this paper, YOLOv5 (You Only Look Once v5) is used as the overall structural framework of the target detection algorithm, and improvement is made on the basis of its detection precision in the underwater environment. Specifically, an attention mechanism (Channel and Spatial Fusion Attention, CSFA) that fuses the channel attention and spatial attention is proposed and added to the YOLOv5 network framework, enabling the network to focus on both the prominent features of the detected object and the spatial information of the detected object. The proposed method was tested on the underwater target detection dataset provided by the China Underwater Robot Professional Competition. The experimental detection precision (P) reached 85%, the recall (R) reached 82.2%, and the mean average precision (mAP) reached 87.5%. The effectiveness of the proposed method was verified, and its underwater target detection performance was better than that of ordinary models.

In order to improve the trajectory tracking accuracy of an Underwater Vehicle Manipulator System (UVMS) under uncertain disturbance conditions of ocean current, a Model-free Adaptive Control (MFAC) method was used. Combined with Radial Basis Function Neural Networks (RBFNN), the RBFNN-MFAC method is proposed to improve the performance of the controller. A hydrodynamic model of UVMS was defined in the commercial software, Fluent, to calculate hydrodynamics disturbance, and the mechanism of the dynamic model of UVMS was defined in the commercial software, Adams, to simulate the motion of UVMS. The trajectory tracking performance with various control schemes, including PID (Proportional Integral Derivative), MFAC and RBFNN-MFAC, were analyzed with the Adams and Simulink joint simulation model. The results show that the position tracking accuracy and the speed tracking accuracy with the MFAC control scheme were 68.1% and 81.0% better, respectively, than those with PID control scheme. The position tracking accuracy and the speed tracking accuracy with the RBFNN-MFAC control scheme were 66.3% and 43.1% better, respectively, than those with the MFAC control scheme. The MFAC control scheme and the RBFNN-MFAC control scheme proposed in this paper exhibit good trajectory tracking performance without the precise dynamic model of UVMS, which is of great importance to applications in engineering.

To improve the efficiency and stability of an eccentric rotating wave energy converter (ERWEC), the adaptive hydraulic power-take-off (PTO) system with an accumulator is designed and developed. Experiments are performed to analyze the effects of trigger pressure, delay time, and open state duration on average output power and power fluctuation index. The results show that the effects of those three control parameters of accumulator on output power are strongly coupled. The experimental examples are designed based on the optimal Latin hypercube sampling (OLHS) method, and the nonparameterized agent models of control parameters to output power indices are established based on the Gaussian process regression (GPR) method. With the help of sensitivity analysis, it is found that the coupled effect of delay time and open state duration on the power fluctuation index is greater than that on the average output power. Furthermore, the optimal combination of control parameters is obtained by non-dominated sorting genetic algorithm-III (NSGA-III), which improves the amount and stability of output power from a hydraulic PTO system. This paper is of important significance for parameter setting of hydraulic PTO systems with an accumulator for wave energy converter, and provides the basis for the real-time adjustment of control parameters under complex sea conditions.

Deep-sea exploring and sampling technologies have become frontier topics. Generally, the movable exploring mode near the seabed with low disturbance is an important way to improve the measurement accuracy and expand the measurement range. Inspired by fish, the fishlike propulsion method has the characteristics of low disturbance and high flexibility, which is very suitable for near-seabed detection under complex terrain conditions. However, the swimming mechanism and surrounding flow field evolution law of the robotic fish under the constraints of complex terrain are still unclear. In this paper, the confined terrain space is constructed with an undulating seabed and a narrow channel, and the hydrodynamic changing law and flow field evolution law of the autonomous swimming process of the fishlike swimmer in the confined space are analyzed. Moreover, the influence mechanism of the terrain on the motion performance of the robotic fish is revealed, and the optimal motion mode of the robotic fish under a complex terrain constraint is discussed. The results show that the propulsion force, Froude efficiency, and swimming stability of the robotic fish vary with the distance from the bottom under the undulating seabed condition lightly. When the distance from the bottom exceeds a certain value, it can be considered that the undulating seabed no longer affects the swimmer. Furthermore, when the robotic fish swims through a narrow channel with certain width, the swimming performance obviously varies with the distance from the boundary surface. During swimming in the confined terrain space, the propulsion force and swimming stability of robotic fish will decrease. In order to maintain the forward speed, the robotic fish should improve the tail-beat frequency in real time. However, considering the swimming stability, the tail-beat frequency is not the larger the better. The relevant conclusions of this paper could provide theoretical support for the development of low-disturbance bionic exploring and sampling platforms for deep-sea resources and environments.