Agent-based middleware architecture for reconfigurable manufacturing systems

Modern manufacturing systems are expected to be flexible and efficient in order to cope with challenging market demands. Thus, they must be flexible enough as to meet changing requirements such as changes in production, energy efficiency, performance optimization, fault tolerance to process or controller faults, among others. Demanding requirements can be defined as a set of quality of service (QoS) requirements to be met. This paper proposes a generic and customizable multi-agent architecture that, making use of distributed agents, monitors QoS, triggering, if needed, a reconfiguration of the control system to recover QoS. As a proof of concept, the architecture has been implemented to provide availability of the control system understood as service continuity. The prototype has been tested in a case study consisting of an assembly cell where assessment of the approach has been conducted.     

Providing universally accessible interactive services through TV sets: implementation and validation with elderly users

One of the challenges that Ambient Intelligence (AmI) faces is the provision of a usable interaction concept to its users, especially for those with a weak technical background. In this paper, we describe a new approach to integrate interactive services provided by an AmI environment with the television set, which is one of the most widely used interaction client in the home environment. The approach supports the integration of different TV set configurations, guaranteeing the possibility to develop universally accessible solutions. An implementation of this approach has been carried out as a multimodal/multi-purpose natural human computer interface for elderly people, by creating adapted graphical user interfaces and navigation menus together with multimodal interaction (simplified TV remote control and voice interaction). In addition, this user interface can also be suited to other user groups. We have tested a prototype that adapts the videoconference and the information service with a group of 83 users. The results from the user tests show that the group found the prototype to be both satisfactory and efficient to use.     

Novel control algorithm for MPPT with Boost converters in photovoltaic systems

The great advances in efficiency and performance of photovoltaic modules would not be very useful if they do not work close to their maximum power point (MPP). In this paper a novel Sliding Mode Control (SMC) based algorithm is proposed to be implemented in a DC/DC converter in order to make an autonomous photovoltaic system to work at the MPP. Once that the design of the novel algorithm has been detailed (especially the novel part relative to the current reference signal) and its stability has been demonstrated, its performance has been compared with two of the most commonly used algorithms in this scope, i.e., Perturbation & Observation (P&O) and Incremental Conductance (IC) algorithms, in addition to a PI controller because it is one of the preferred controllers in industrial applications. This comparison has been carried out taking into account both simulated and experimental tests. The first focused on their behavior when sudden changes in irradiance and temperature, while the lasts analyzed them when the load resistance was varying arbitrarily in actual facilities (composed of a photovoltaic module Mitsubishi PV-TD185MF5, a Boost converter, a variable load and a real-time data acquisition card dSPACE DSP1104 used as the interface between the control algorithm implemented in Simulink/Matlab and the real photovoltaic module). After completing tests under different conditions, we found that the proposed SMC based algorithm outperforms the PI controller and the P&O and IC algorithms, especially in experiments carried out using actual facilities.     

Dual model oriented modeling of monocrystalline PV modules based on artificial neuronal networks

The deep insight into the different elements that compose photovoltaic (PV) systems is capital to boost the optimization of each one of them and consequently, increment of the overall performance of the whole PV systems. In this paper we address the open problem of obtaining empirical accurate models of monocrystalline PV modules in a systematic and unattended fashion. In order to tackle this issue, we used a dual model oriented modeling approach based on artificial neural networks (ANN) due to their advantages, being the generalization capability the most outstanding one. We tried two different model approaches with different input/outputs specifications to learn the electrical behavior of a monocrystalline PV module Atersa A-55 placed on the roof of the Faculty of Engineering of Vitoria-Gasteiz (Basque Country University, Spain). Following these approaches we found two season oriented models of I_PV with a RMSE accuracy of 0.20 mA and 0.26 mA respectively, which is better than the precision of the measurement devices. After comparing these results with the state-of-art ones, we conclude that we have outperformed the previously existing results.     

Performance enhancement of the artificial neural network–based reinforcement learning for wind turbine yaw control

The yaw angle control of a wind turbine allows maximization of the power absorbed from the wind and, thus, the increment of the system efficiency. Conventionally, classical control algorithms have been used for the yaw angle control of wind turbines. Nevertheless, in recent years, advanced control strategies have been designed and implemented for this purpose. These advanced control strategies are considered to offer improved features in comparison to classical algorithms. In this paper, an advanced yaw control strategy based on reinforcement learning (RL) is designed and verified in simulation environment. The proposed RL algorithm considers multivariable states and actions, as well as the mechanical loads due to the yaw rotation of the wind turbine nacelle and rotor. Furthermore, a particle swarm optimization (PSO) and Pareto optimal front (PoF)‐based algorithm have been developed in order to find the optimal actions that satisfy the compromise between the power gain and the mechanical loads due to the yaw rotation. Maximizing the power generation and minimizing the mechanical loads in the yaw bearings in an automatic way are the objectives of the proposed RL algorithm. The data of the matrices Q (s,a) of the RL algorithm are stored as continuous functions in an artificial neural network (ANN) avoiding any quantification problem. The NREL 5‐MW reference wind turbine has been considered for the analysis, and real wind data from Salt Lake, Utah, have been used for the validation of the designed yaw control strategy via simulations with the aeroelastic code FAST.     

Modeling and evaluating beat gestures for social robots

Multimedia Tools and Applications - Natural gestures are a desirable feature for a humanoid robot, as they are presumed to elicit a more comfortable interaction in people. With this aim in mind, we...