Fish monitoring has become increasingly popular due to its growing real-world applications and recent advancements in intelligent technologies such as AI, Computer Vision, and Robotics. The primary objective of this article is to review benchmark datasets used in fish monitoring while introducing a novel framework that categorizes fish monitoring applications into four main domains: Fish Detection and Recognition (FDR), Fish Biomass Estimation (FBE), Fish Behavior Classification (FBC), and Fish Health Analysis (FHA). Additionally, this study proposes dedicated workflows for each domain, marking the first comprehensive effort to establish such a structured approach in this field. The detection and recognition of fish involve identifying fish and fish species. Estimating fish biomass focuses on counting fish and measuring their size and weight. Fish Behavior Classification tracks and analyzes movement and extracts behavioral patterns. Finally, health analysis assesses the general health of the fish. The methodologies and techniques are analyzed separately within each domain, providing a detailed examination of their specific applications and contributions to fish monitoring. These innovations enable fish species classification, fish freshness evaluation, fish counting, and body length measurement for biomass estimation. The study concludes by reviewing the development of key datasets and techniques over time, identifying existing gaps and limitations in current frameworks, and proposing future research directions in fish monitoring applications.

Distributed source seeking in a 3-dimensional environment without explicit estimation of the gradient of the field is challenging. Nevertheless, for a swarm of an arbitrary number of agents, we develop a strategy to perform a source seeking behavior by requiring agents to synchronize their direction of motion using only local interactions while modulating their speed based only on the field value measured by each agent. The agents collectively move toward the source without estimating gradient and without sharing measurements of the field. We formulate a cascaded input-to-state stability problem from which we obtain Lyapunov-based convergence and robustness results. We validate the convergence of the swarm to the minimum of the field through simulated source seeking behavior in a 3-dimensional scalar field.

Algorithms for multiagent systems to locate a source or to follow a desired level curve of spatially distributed scalar fields generally require sharing field measurements among the agents for gradient estimation. Yet, in this article, we propose a distributed active perception strategy that enables swarms of various sizes and graph structures to perform source seeking and level curve tracking without the need to explicitly estimate the field gradient or explicitly share measurements. The proposed method utilizes a consensus-like principal component analysis perception algorithm that does not require explicit communication in order to compute a local body frame. This body frame is used to design a distributed control law where each agent modulates its motion based only on its instantaneous field measurement. Several stability results are obtained within a singular perturbation framework that justifies the convergence and robustness of the strategy. Additionally, efficiency is validated through robots experiments.

In this letter, we propose a bio-inspired derivative-free optimization algorithm capable of minimizing objective functions with vanishing or exploding gradients. The proposed method searches for improvements by leveraging a PCA-based strategy similar to fish foraging. The strategy does not require explicit gradient computation or estimation and is shown in simulation to require few function evaluations. Additionally, our analysis proves that the proposed algorithm's search direction converges to the gradient direction everywhere outside of small neighborhoods around local minima. Applications to a data-driven LQR problem and noisy Rosenbrock optimization problem are demonstrated. Empirical results show the proposed method exhibits fast convergence and is able to find the LQR gains for any controllable system, including unstable systems, and is robust to noisy function evaluations.

Physics-Informed Neural Networks (PINNs) have emerged as a powerful tool for modeling and controlling complex energy systems by embedding physical laws into deep learning architectures. This review paper highlights the application of PINNs in grid-connected inverter systems (GCISs), categorizing them by key tasks: parameter estimation, state estimation, control strategies, fault diagnosis and detection, and system identification. Particular focus is given to the use of PINNs in enabling accurate parameter estimation for aging and degradation monitoring. Studies show that PINN-based approaches can outperform purely data-driven models and traditional methods in both computational efficiency and accuracy. However, challenges remain, mainly related to high training costs and limited uncertainty quantification. To address these, emerging strategies such as advanced PINN frameworks are explored. The paper also explores emerging solutions and outlines future research directions to support the integration of PINNs into practical inverter design and operation.