BUSCADOR RECOLECTA

Current State-of-the-Art (SotA) chatbots are able to produce high-quality sentences, handling different conversation topics and larger interaction times. Unfortunately, the generated responses depend greatly on the data on which they have been trained, the specific dialogue history and current turn used for guiding the response, the internal decoding mechanisms, and ranking strategies, among others. Therefore, it may happen that for semantically similar questions asked by users, the chatbot may provide a different answer, which can be considered as a form of hallucination or producing confusion in long-term interactions. In this research paper, we propose a novel methodology consisting of two main phases: (a) hierarchical automatic detection of topics and subtopics in dialogue interactions using a zero-shot learning approach, and (b) detecting inconsistent answers using k-means and the Silhouette coefficient. To evaluate the efficacy of topic and subtopic detection, we use a subset of the DailyDialog dataset and real dialogue interactions gathered during the Alexa Socialbot Grand Challenge 5 (SGC5). The proposed approach enables the detection of up to 18 different topics and 102 subtopics. For the purpose of detecting inconsistencies, we manually generate multiple paraphrased questions and employ several pre-trained SotA chatbot models to generate responses. Our experimental results demonstrate a weighted F-1 value of 0.34 for topic detection, a weighted F-1 value of 0.78 for subtopic detection in DailyDialog, then 81% and 62% accuracy for topic and subtopic classification in SGC5, respectively. Finally, to predict the number of different responses, we obtained a mean squared error (MSE) of 3.4 when testing smaller generative models and 4.9 in recent large language models.

Personal assistants and social robotics have evolved significantly in recent years thanks to the development of artificial intelligence and affective computing. Today’s main challenge is achieving a more natural and human interaction with these systems. Integrating emotional models into social robotics is necessary to accomplish this goal. This paper presents an emotional model whose design has been supervised by psychologists, and its implementation on a social robot. Based on social psychology, this dimensional model has six dimensions with twelve emotions. Fuzzy logic has been selected for defining: (i) how the input stimuli affect the emotions and (ii) how the emotions affect the responses generated by the robot. The most significant contribution of this work is that the proposed methodology, which allows engineers to easily adapt the robot personality designed by a team of psychologists. It also allows expert psychologists to define the rules that relate the inputs and outputs to the emotions, even without technical knowledge. This methodology has been developed and validated on a personal assistant robot. It consists of three input stimuli, (i) the battery level, (ii) the brightness of the room, and (iii) the touch of caresses. In a simplified implementation of the general model, these inputs affect two emotions that generate an externalized emotional response through the robot’s heartbeat, facial expression, and tail movement. The three experiments performed verify the correct functioning of the emotional model developed, demonstrating that stimuli, independently or jointly, generate changes in emotions that, in turn, affect the robot’s responses.

The next generation of conversational AI has brought incredible capabilities such as high contextuality, naturalness, multimodality, and extended knowledge, but also important challenges such as high user expectations, high latencies, large computational requirements, as well as more subtle problems such as mismatch on existing databases for fine-tuning purposes, difficulties for pre-trained LLMs models to handle dialogue interactions, and the integration of multimodal capabilities.

This paper describes the architecture, methodology, and results of our THAURUS chatbot developed for the Alexa Prize Socialbot Grand Challenge (SGC5). Our proposal relies on several innovative ideas to take advantage of existing LLMs to create engaging user experiences that are capable of handling real users in a scalable way and without compromising the competition rules. Different SotA dialogue generators were fine-tuned and incorporated to give variability and handling the wide range of topic conversations; we also developed mechanisms to control the quality of the responses (e.g., detecting and handling toxic interactions, keeping topic coherence, and increasing engagement by providing up-to-date information in a conversational style).

In addition, our system extends the capabilities of the Cobot architecture by incorporating modules to automatically generate images, provide voice cloning capabilities with fictional characters, serve contextual sounds for detected entities in the dialogue, better capitalization and punctuation capabilities, and to provide natural expressions of interest.

Finally, we also included a trained generative selector and a reference-free model for automatic evaluation of turns that could reduce latencies and complement the ranker’s capabilities to select the best generative answer.

In this work, a new methodology for the dynamic analysis of non-linear systems is developed by applying the Mamdani fuzzy model. With this model, parameters such as settling time, peak time and overshoot will be obtained. The dynamic analysis of non-linear fuzzy systems with triangular membership functions is performed, and linguistic variables describing overly complex or ill-defined phenomena are used to fit the model. Scaling factors will simplify the modification of the variables, making them easier to find the system model. The specifications of second-order characteristics in the time domain, such as overshoot and peak time, will be represented graphically. As a case study, the proposed methods are implemented to analyse the dynamics of a tank and a simple pendulum for first-order and second-order systems, respectively, where it is observed that the proposed methodology offers highly positive results.

This paper presents an innovative and cost-effective solution for the absolute localization of mobile robots using ultrasound beacons. The proposed system addresses the challenge of precise positioning within a controlled environment by employing Differential Time-of-Flight (ToF) measurements to determine the relative distances between the robot and optimally placed beacons. Unlike other ToF methods that require synchronization pulses, the proposed approach eliminates this requirement, significantly simplifying the setup and reducing system complexity. Furthermore, the system achieves a higher sampling rate than conventional synchronization-based systems, enhancing real-time performance. Detailed analysis and simulation demonstrate the system's ability to provide accurate and reliable localization. The results highlight the potential for broad application in various robotic environments, offering a robust solution for absolute positioning without complex synchronization strategies. This work underscores the advantages of using ToF measurements with ultrasound beacons and contributes to the ongoing development of efficient and cost-effective robotic localization systems.

This work presents a multi-strategy fuzzy controller (MSFC) based on the Takagi–Sugeno (T–S) model, where the membership functions are based on polar coordinates. Adopting a polar coordinate system helps to reduce the number of rules significantly, especially when fuzzy variables exhibit extreme values. This approach also allows more rules to be assigned in certain directions (angles) than others, while the radius represents the distance to the reference. The MSFC uses three strategies: constant input (CI), where a predefined gain is set in the exterior rules; standard discrete state–space model linear–quadratic regulator (LQR), in the intermediate rules; and an incremental state–space model LQR (INC-LQR) in the central rule. A technique employing fuzzy fusion is suggested to facilitate smooth transitions among the various methods. The results show that the proposed MSFC in polar coordinates has a faster transient response and zero steady-state error compared to other fuzzy controllers, and it requires fewer rules compared to an MSFC based on Cartesian coordinates.

In this paper, a new multidimensional Takagi-Sugeno (T-S) identification technique is proposed for multivariable nonlinear systems. In this technique, multidimensional membership functions are designed using concepts from solid mechanics. The design of membership functions is carried out in multidimensional space, defining the principal axes from the eigenvectors of the inertia matrix, and it has the characteristic of dividing the space into regions with the same inertia. These regions are analyzed to define the center of gravity for each rule. Illustrative examples of multivariable nonlinear systems, such as a thermal mixing process and a binary distillation column, are selected to evaluate the effectiveness of the proposed method. The proposed method is compared with traditional T-S identification that uses one-dimensional membership functions and shows a reduction in the relative identification error and the algorithm execution time. Additionally, the proposed method prevents rules from being positioned outside the system's range, thereby avoiding the generation of unnecessary rules.

The increasing interest in developing robots capable of navigating autonomously has led to the necessity of developing robust methods that enable these robots to operate in challenging and dynamic environments. Visual odometry (VO) has emerged in this context as a key technique, offering the possibility of estimating the position of a robot using sequences of onboard cameras. In this paper, a VO algorithm is proposed that achieves sub-pixel precision by combining optical flow and direct methods. This approach uses only a downward-facing, monocular camera, eliminating the need for additional sensors. The experimental results demonstrate the robustness of the developed method across various surfaces, achieving minimal drift errors in calculation.

In this work, a fuzzy predictive optimal control for multivariable nonlinear systems with pure time delays is presented. Therefore, dynamic local linear state models are used at each point of the state space obtained by fuzzy Takagi-Sugeno (T-S) modelling. The modelling error is considered as white noise, and the state is observed using the Kalman Filter (KF). This method does not take into account possible input constraints. Therefore, it applies to systems where there are no saturation problems. In this work, a new approach in the Model Predictive Control (MPC) method is proposed by calculating the control signal increment as a function of the error between a reference state vector and the prediction at N-steps of the state vector instead of using the traditional MPC approach which is based on calculating the error between a reference and the predicted output. The proposed method has a satisfactory tracking performance. The main feature of the proposed method is that it attains computational savings compared to other methods that have used incremental state models, which making it more appropriate for real-time applications.

In this work, a new multi-strategy fuzzy control (MSFC) based on fuzzy fusion techniques for the control of multivariable nonlinear systems is proposed. This new nonlinear control method is a mixture of different control techniques by fuzzy interpolation using the fuzzy Takagi-Sugeno model. It combines various control strategies to achieve smooth transient responses and zero steady state error. The following techniques are merged within the proposed MSFC structure: an optimal state feedback control that yields an optimal transient response, optimal control through an incremental state model that returns a zero steady state error, and a constant input control to maintain the system behavior within predefined boundaries whenever the MSFC method is applied.

In this work, a new nonlinear control method is proposed, which integrates the Takagi-Sugeno (T-S) fuzzy model with the Principal Component Analysis (PCA) technique. The approach uses PCA to reduce the system's dimensionality, minimizing the number of fuzzy rules required in the T-S fuzzy model. This reduction not only simplifies the system variables but also decreases the computational complexity, resulting in a more efficient control with smooth transient responses and zero steady-state error. To validate the performance of this PCA-based approach for both system identification and control, an interconnected double-tank system was employed. The results demonstrate the method's capacity to maintain control accuracy while reducing computational load, making it a promising solution for applications in industrial and engineering systems that require robust, efficient control mechanisms.

In this work, an experimental methodology is shown that allows selecting the best controller alternative to govern a two-axis solar tracking system in trajectory tracking tasks. This, based on the comparison of the performance of different controllers during the tracking of a solar trajectory (obtained offline with the help of a numerical method) using a prototype solar tracking system in a simulation environment. The selection of the controllers is based on the performance reported in the literature and commercially for other alternatives based on parameters such as: tracking error and energy consumption, computational cost, ease of implementation, ease of tuning, among others. However, for the final validation and selection of the controller, a brief analysis of the numerical results obtained during the follow-up tests is provided.

The increasing complexity of nonlinear multivariable systems poses significant challenges for effective modeling and control. Fuzzy modeling and control typically use fuzzy inference with one-dimensional membership functions. However, the use of multidimensional membership functions can provide significant benefits in optimizing and reducing the computational cost of a fuzzy controller. In this work, we propose the use of fuzzy clustering techniques to adjust and design multidimensional membership functions. These techniques represent a well-developed and comprehensive framework, though they are often disconnected from traditional fuzzy modeling and control methodologies. Thus, this work also seeks to combine fuzzy techniques of different applications with a single ultimate goal, namely, to optimize the modeling and control of nonlinear systems. Our main objective is system identification, modeling, and control using the Takagi–Sugeno method based on one-dimensional and multidimensional membership functions. Moreover, a comparison of various fuzzy clustering techniques for the design of multidimensional membership functions is carried out to demonstrate the effectiveness of the proposed methods in optimizing control performance and reducing computational cost.

Social networks have become a powerful communication tool, with millions of people exchanging information, opinions, and experiences daily. Companies, organizations, and even people have turned this tool into a marketing platform to position themselves and gain popularity. However, not only do companies present products or services to society, but society also provides feedback. This feedback also has a significant impact. It is impossible to process all this vast information manually in time, but it is crucial. This information is precious even to governmental or public entities such as universities. Potential future students will use social media to learn about the general feel of the institution. Therefore, this study presents a new dataset called CEIMaT2021, which compiles all tweets in Spanish related to the Technical School of Industrial Engineering of the Universidad Politecnica de Madrid (ETSII-UPM). This dataset is designed for two main tasks of Online Reputation Management: 1) automatic detection of topics and 2) polarity. Furthermore, this study shows that the BETO model obtains better performance for topic detection for these tasks. Meanwhile, the MarIA model obtains better results for polarity detection.

Este dataset contiene nubes de puntos parciales derivadas de los modelos Bunny, Dragon, Armadillo y Buda, originalmente del reconocido Stanford 3D Scanning Repository. Se incluyen 15 vistas parciales para cada modelo, generadas mediante simulaciones de un sensor tipo ToF (Time-of-Flight) utilizando el software Blender y la extensión Blensor. Estas simulaciones permiten evaluar el rendimiento de algoritmos de registro al enfrentar nubes parciales y su correspondiente nube completa de referencia.

Directorio raíz: dataset
- Subdirectorio bunny: contiene 15 nubes de puntos parciales (scanX.pcd) del modelo Bunny, junto con una imagen asociada de cada escaneo parcial y la nube completa de referencia (source.pcd).
- Subdirectorio dragon: contiene 15 nubes de puntos parciales (scanX.pcd) del modelo Dragon, junto con una imagen asociada de cada escaneo parcial y la nube completa de referencia (source.pcd).
- Subdirectorio armadillo: contiene 15 nubes de puntos parciales (scanX.pcd) del modelo Armadillo, junto con una imagen asociada de cada escaneo parcial y la nube completa de referencia (source.pcd).
- Subdirectorio buda: contiene 15 nubes de puntos parciales (scanX.pcd) del modelo Buda, junto con una imagen asociada de cada escaneo parcial y la nube completa de referencia (source.pcd).

Cada carpeta incluye las nubes parciales y una imagen por cada escaneo parcial, facilitando la visualización de las capturas desde diferentes ángulos. Las nubes de puntos completas (source.pcd) sirven como referencia para el registro en cada subdirectorio.


5. Notas

Given a graph, the maximum clique problem (MCP) asks for determining a complete subgraph with the largest possible number of vertices. We propose a new exact algorithm, called CliSAT, to solve the MCP to proven optimality. This problem is of fundamental importance in graph theory and combinatorial optimization due to its practical relevance for a wide range of applications. The newly developed exact approach is a combinatorial branch-and-bound algorithm that exploits the state-of-the-art branching scheme enhanced by two new bounding techniques with the goal of reducing the branching tree. The first one is based on graph colouring procedures and partial maximum satisfiability problems arising in the branching scheme. The second one is a filtering phase based on constraint programming and domain propagation techniques. CliSAT is designed for structured MCP instances which are computationally difficult to solve since they are dense and contain many interconnected large cliques. Extensive experiments on hard benchmark instances, as well as new hard instances arising from different applications, show that CliSAT outperforms the state-of-the-art MCP algorithms, in some cases by several orders of magnitude., This publication is part of the R&D project “Cognitive Personal Assistance for Social Environments (ACOGES)”, reference PID2020-113096RB-I00, funded by MCIN/AEI/10.13039/501100011033., Peer reviewed

In this paper, a new multidimensional Takagi–Sugeno (T-S) identification technique is proposed for multivariable nonlinear systems. In this technique, multidimensional membership functions are designed using concepts from solid mechanics. The design of membership functions is carried out in multidimensional space, defining the principal axes from the eigenvectors of the inertia matrix, and it has the characteristic of dividing the space into regions with the same inertia. These regions are analyzed to define the center of gravity for each rule. Illustrative examples of multivariable nonlinear systems, such as a thermal mixing process and a binary distillation column, are selected to evaluate the effectiveness of the proposed method. The proposed method is compared with traditional T-S identification that uses one-dimensional membership functions and shows a reduction in the relative identification error and the algorithm execution time. Additionally, the proposed method prevents rules from being positioned outside the system’s range, thereby avoiding the generation of unnecessary rules., This publication is part of the R&D project “Cognitive Personal Assistance for Social Environments (ACOGES)”, reference PID2020-113096RB-I00, funded by MCIN/AEI/10.13039/501100011033., Peer reviewed

This article belongs to the Special Issue Advances in Robotics and Autonomous Systems., This paper presents an innovative and cost-effective solution for the absolute localization of mobile robots using ultrasound beacons. The proposed system addresses the challenge of precise positioning within a controlled environment by employing Differential Time-of-Flight (ToF) measurements to determine the relative distances between the robot and optimally placed beacons. Unlike other ToF methods that require synchronization pulses, the proposed approach eliminates this requirement, significantly simplifying the setup and reducing system complexity. Furthermore, the system achieves a higher sampling rate than conventional synchronization-based systems, enhancing real-time performance. Detailed analysis and simulation demonstrate the system’s ability to provide accurate and reliable localization. The results highlight the potential for broad application in various robotic environments, offering a robust solution for absolute positioning without complex synchronization strategies. This work underscores the advantages of using ToF measurements with ultrasound beacons and contributes to the ongoing development of efficient and cost-effective robotic localization systems., This paper is part of the R&D project “Cognitive Personal Assistance for Social Environments (ACOGES)”, reference PID2020-113096RB-I00, funded by MCIN/AEI/10.13039/501100011033., Peer reviewed

The increasing interest in developing robots capable of navigating autonomously has led to the necessity of developing robust methods that enable these robots to operate in challenging and dynamic environments. Visual odometry (VO) has emerged in this context as a key technique, offering the possibility of estimating the position of a robot using sequences of onboard cameras. In this paper, a VO algorithm is proposed that achieves sub-pixel precision by combining optical flow and direct methods. This approach uses only a downward-facing, monocular camera, eliminating the need for additional sensors. The experimental results demonstrate the robustness of the developed method across various surfaces, achieving minimal drift errors in calculation., This paper is part of the R&D project “Cognitive Personal Assistance for Social Environments (ACOGES)”, reference PID2020-113096RB-I00, funded by MCIN/AEI/10.13039/501100011033., Peer reviewed