Underwater target detection techniques have been extensively applied to underwater vehicles for marine surveillance, aquaculture, and rescue applications. However, due to complex underwater environments and insufficient training samples, the existing underwater target recognition algorithm accuracy is still unsatisfactory. A long-term effort is essential to improving underwater target detection accuracy. To achieve this goal, in this work, we propose a modified YOLOv5s network, called YOLOv5s-CA network, by embedding a Coordinate Attention (CA) module and a Squeeze-and-Excitation (SE) module, aiming to concentrate more computing power on the target to improve detection accuracy. Based on the existing YOLOv5s network, the number of bottlenecks in the first C3 module was increased from one to three to improve the performance of shallow feature extraction. The CA module was embedded into the C3 modules to improve the attention power focused on the target. The SE layer was added to the output of the C3 modules to strengthen model attention. Experiments on the data of the 2019 China Underwater Robot Competition were conducted, and the results demonstrate that the mean Average Precision (mAP) of the modified YOLOv5s network was increased by 2.4%.

Nonlinear positive control system can be found in many real-world applications but the positivity requirements lead to challenges in system analysis and control design. In this article, we approach the problem by fuzzy-model-based control techniques and overcome some challenges including transforming the nonconvexity conditions when both positive and stability conditions exist into convexity conditions that can be solved by the convex programming techniques. This article focuses on the static output-feedback tracking control issue of positive polynomial fuzzy-model-based systems. The purpose of the tracking control is to design an appropriate static output feedback polynomial fuzzy controller which can drive the system states of the nonlinear plant to follow those of a stable reference model subject to an $H_\infty$ performance. The concept of imperfectly matched premises is employed to enhance the design and implementation flexibility. To circumvent the problem of nonconvex stability conditions, an approach is employed to transform the nonconvex stability conditions into convex ones by introducing a novel scalar implantation transformation technique. Besides, the partition approximation of membership functions with local information of membership functions is used to promote stability analysis and synthesis of controllers. The positive and relaxed stability conditions for static output-feedback tracking control with $H_\infty$ performance being taken into account are obtained in terms of sum-of-squares. Finally, a simulation example is presented to verify the effectiveness of the proposed tracking control approach.

Nonlinear positive systems are ubiquitous in practical applications. The constraint of positive conditions makes the design of this system still quite challenging even for the existing advanced control theory. In addition, when uncertainty and external disturbance also act on the positive nonlinear system at the same time, it will aggravate the complexity of the problem. In response to this problem, for the first time, this article presents a feasible robust tracking control strategy for the positive nonlinear systems with external disturbance and subject to uncertainty. In the proposed control strategy, the positive Takagi–Sugeno (T–S) fuzzy model is used to effectively represent the nonlinearity of the positive system, and the interval type-2 (IT2) fuzzy sets are used to deal with the uncertainty. The focus of tracking control is to design a reasonable IT2 fuzzy controller, which can make the states of the plant follow those of the given stable positive reference system as accurately as possible under the $H_\infty$ performance. Different from the tracking control research of general systems in the existing literature, in the tracking control analysis of positive systems, the generation of nonconvex conditions is inevitable. This article smoothly removes the obstacle of this problem. In addition, this article successfully attempts to relax the conservativeness of the basic analysis conditions by improving IT2 membership-function-dependent method, thereby improving the tracking effect under $H_\infty$ performance. Finally, a detailed simulation example is given to verify the effectiveness of the proposed strategy.

This paper deals with the stability and positivity analysis of polynomial-fuzzy-model-based (PFMB) control systems with time delay, which is formed by a polynomial fuzzy model and a polynomial fuzzy controller connected in a closed loop, under imperfect premise matching. To improve the design and realization flexibility, the polynomial fuzzy model and the polynomial fuzzy controller are allowed to have their own set of premise membership functions. A sum-of-squares-based stability analysis approach using the Lyapunov stability theory is employed to investigate the positivity and stability of the PFMB control systems and synthesize the polynomial fuzzy controller. In order to relax the stability results, we propose two methods: first, membership functions are considered as symbolic variables in the stability analysis; and second, the property of the membership functions and the boundary information of the membership functions are considered in the stability analysis. A simulation example is given to illustrate the effectiveness of the proposed approach.

This work focuses on the filtering problem and stability analysis for positive Takagi-Sugeno (T-S) fuzzy systems with time delay under L1-induced performance. Due to the importance of estimation of system states but the few filter design results on positive nonlinear systems, it is an attractive and meaningful topic well worth studying. In order to fully exploit and take advantage of the positivity of positive T-S fuzzy systems, many commonly used methods, for instance free-weighting matrix approach and similarity transformation are probably not suitable for positive systems. To address the hard-nut-to-crack problem, an auxiliary variable is introduced so that the augmentation approach can be employed to carry out the positivity and stability analysis of filtering error systems. In addition, another obstacle that cannot be ignored is the existence of non-convex terms in the stability and positivity conditions. For getting around this barrier, some iterative linear matrix inequality (ILMI) algorithms have been proposed in the literature. However, considering the weakness that these methods cannot guarantee the convergence to a numerical solution and the iterative process is exhaustive, we present an effective matrix decoupling method to convert the non- convex conditions into convex ones in this paper. Furthermore, a linear co-positive Lyapunov function which incorporates the positivity of system states and time delay at the same time is chosen so that the positivity characteristic of filtering error systems can be captured further. However, because of plenty of valuable information of membership functions (MFs) being ignored, hence, the obtained results are conservative. For the sake of relaxing the conservativeness, the advanced piecewise- linear membership functions (PLMFs) approximate method is utilized to facilitate the stability and positivity analysis. Therefore, the relaxed stability and positivity conditions which are cast as sum of squares (SOS) are obtained and can be solved numerically. Finally, the effectiveness of the designed fuzzy filtering strategy with satisfying L1-induced performance are demonstrated by a simulation example.