Recent advancements in reinforcement learning (RL) and humanoid robotics are rapidly addressing the challenge of adapting to complex, dynamic environments in real time. This letter introduces a novel approach that integrates two key concepts: experience variability (a criterion for detecting changes in loco-manipulation) and experience accumulation (an efficient method for storing acquired experiences based on a selection criterion). These elements are incorporated into the development of RL agents and humanoid robots, with an emphasis on stability. This focus enhances adaptability and efficiency in unpredictable environments. Our approach enables more sophisticated modeling of such environments, significantly improving the system's ability to adapt to real-world complexities. By combining this method with advanced RL techniques, such as Proximal Policy Optimization (PPO) and Model-Agnostic Meta-Learning (MAML), and incorporating self-learning driven by stability, we improve the system's generalization capabilities. This facilitates rapid learning from novel and previously unseen scenarios. We validate our algorithm through both simulations and real-world experiments on the HRP-4 humanoid robot, utilizing an intrinsically stable model predictive controller.

The single-stage (SS) ac–dc converter is a promising solution for bidirectional electric vehicle chargers (BEVCs) due to its smaller volume, higher efficiency, and durability compared with the two-stage ac–dc converter. Nevertheless, typically SS ac–dc structures implement highly complex modulation schemes because most of them use bipolar voltage switches. In this article, a novel modulation and control strategy is introduced for the SS quad-active-bridge resonant-type ac–dc converter, which only uses unipolar voltage switches. In the proposed modulation, grid voltages are modulated by duty ratio (DR), whereas the dc source voltage is modulated by phase-shift (PS) and DR. The PS angle takes a constant value according to grid currents amplitude. Moreover, the DR angles, used to modulate grid voltages, control the sinusoidal shape of grid currents according to proportional-integral-resonant controllers cascaded to a second-order low-pass filters and an active damping. Whereas, the DR angle, used to modulate the dc source voltage, is applied to reduce the high-frequency current, in case of reactive power transfer to the grid. Experimental results from a 1.5 kW prototype show high efficiency (96.4%), bidirectional active-reactive power transfer capability, and low harmonic distortion in the grid currents (1.1%), making it an ideal alternative for BEVCs.

The control of nonlinear systems poses significant challenges due to their inherent complexities, limiting the effectiveness of traditional control strategies. This paper presents an improved fuzzy identification and control method for nonlinear industrial systems, using Takagi-Sugeno fuzzy inference to model nonlinear dynamics as an interpolation of multiple linear subsystems. A key improvement of this approach lies in the accurate identification of the nonlinear model, which leads to fewer control system failures. The research contribution is the development of a control strategy that enhances system reliability while simplifying implementation. The method involves minimizing a cost function that optimizes the system’s output error, refining the fuzzy identification process for dynamic adaptation to varying operating conditions. The strategy also enables the design of linear controllers for each subsystem and applies Parallel Distributed Compensation (PDC) to regulate the overall nonlinear system. This approach is validated through experimental testing on an aero-pendulum system. The results show that the PDCbased control scheme not only ensures high performance across a wide operational range but also significantly reduces identification errors compared to traditional methods. Given its improved accuracy, reduced complexity, and adaptability, this approach holds significant potential for practical application in industrial environments, where robust and efficient control of nonlinear systems is crucial for operational success.

State convergence is a control strategy that was proposed in the early 2000s to ensure stability and transparency in a teleoperation system under specific control gains values. This control strategy has been implemented for a linear system with or without time delay. This paper represents the first attempt at demonstrating, theoretically and experimentally, that this control strategy can also be applied to a nonlinear teleoperation system with \(n\) degrees of freedom and delay in the communication channel. It is assumed that the human operator applies a constant force on the local manipulator during the teleoperation. In addition, the interaction between the remote manipulator and the environment is considered passive. Communication between the local and remote sites is made by means of a communication channel with variable time delay. In this article the theory of Lyapunov–Krasovskii was used to demonstrate that the local–remote teleoperation system is asymptotically stable.

En esta propuesta se plantean algunas sugerencias para la fabricación de sensores inteligentes basados en nanoestructuras amorfas, para lo cual la elaboración de los transductores de los sensores debe tener en cuenta la correlación entre su geometría interna (en escala nanométrica) y el material que lo compone, además de los requerimientos de la empresa que los necesite y el espacio geográfico de la comunidad donde estos sean usados. Tal es así, que la fabricación de estos sensores requiere procesos complejos cual pulverización catódica, deposición atómica, e incluso deposición atómica mediante reacciones electro-químicas. Por lo tanto, se sugiere estudiar qué minerales y componentes químicos se pueden encontrar en los alrededores de las montañas Andinas, para así poderlos usar en la fabricación de los sensores, teniendo un estricto compromiso del cuidado ambiental con los residuos acabado los procesos de fabricación. Además, se plantea tomar en cuenta la comprensión de las características geográficas y climatológicas del lugar de fabricación, cual también el lugar donde se someterá a prueba y uso de los sensores diseñados, que generalmente para las actividades de la población Andina, puede darse en la minería, pesca, agricultura y transporte público. En este contexto, los sensores diseñados en base a nanoestructuras tienen la ventaja de un corto tiempo de respuesta y robustez frente a perturbaciones, lo cual es muy útil para tareas de sistemas mecánicos o mecatrónicos donde el carácter activo de los sensores mejore el performance de las tareas, también esta ventaja es un soporte desde el punto de vista de la programación del microcontrolador que le dé el carácter de inteligencia artificial al sensor elaborado, pues permite ejecutar algoritmos complejos y adaptativos para trabajar en mejor respuesta de la transducción.

'Correction to: Nonlinear Fuzzy Control of Position and Velocity Using Crisp Outputs Implemented in a General-Purpose Microprocessor' published in 'Intelligent Technologies: Design and Applications for Society'