Policy gradient based Reinforcement Learning for real autonomous underwater cable tracking

This paper proposes a field application of a high-level reinforcement learning (RL) control system for solving the action selection problem of an autonomous robot in cable tracking task. The learning system is characterized by using a direct policy search method for learning the internal state/action mapping. Policy only algorithms may suffer from long convergence times when dealing with real robotics. In order to speed up the process, the learning phase has been carried out in a simulated environment and, in a second step, the policy has been transferred and tested successfully on a real robot. Future steps plan to continue the learning process on-line while on the real robot while performing the mentioned task. We demonstrate its feasibility with real experiments on the underwater robot ICTINEU AUV

Análisis de la arquitectura de webs mediante tests de estrés de navegación, de usabilidad y eye tracking

Evaluar una arquitectura de la información en un sitio web ya desplegado no resulta una tarea sencilla. La mayoría de las técnicas se centran en examinar la usabilidad del sistema que, aunque afecta a la arquitectura de la información, no es el único factor que influye en ella. La principal técnica que se utiliza es el test de estrés de navegación. Se muestra un aporte metodológico para hacer dicha técnica más informativa, llevándola más allá de la simple anotación en papel por parte del usuario de respuestas a las preguntas de navegación planteadas. Se propone la combinación de ésta con otras técnicas de evaluación de la usabilidad: la técnica de pensar en voz alta o thinking aloud y un cuestionario de usabilidad. Se ha utilizado un sistema de seguimiento de la mirada o eye tracking para complementar la información obtenida mediante las técnicas aplicadas. El enfoque metodológico planteado se ha puesto a prueba analizando una serie de sitios web de bibliotecas de universidades públicas españolas. Se muestra en los resultados la validez del enfoque empleado, así como el valor que dicho enfoque y el uso del eye tracking aportan al análisis de la arquitectura de la información respecto al test de estrés de navegación tradicional.

The learning process as a cooperative task among teachers and students

This research project gathers several ideas and guidelines on professional improvement as a teacher. This study includes two empirical studies. The first one focuses mainly on the teacher's figure. It is meant to be a study of the several resources that the teacher uses in order to construct the student's knowledge in an English classroom context. The second empirical study focuses on the students. It is a study on how students learn cooperatively by analyzing their oral productions when working in small groups

Teachers' reactions on students' responses: essential for an adequate task monitoring

This research project analyzes the reactions the teacher has on students' responses. Different techniques as discourse markers, types of questions and repair sequences are taken into account, but the author puts a special emphasis on non-verbal communication. To be aware of all these ways of reacting in a class interaction is essential for an adequate task monitoring

Eye tracking for on-screen reading

A medida que avanza la tecnolog a, cada vez son m as comunes los libros digitales. Por eso, existen varias formas de mejorar la experiencia de lectura del usuario, como mostrar la de nici on de una palabra que resulte dif cil, o resaltar lo importante del texto cuando se pasa la vista por encima. En este proyecto, se ha investigado la base de esto con la ayuda de un Eye Tracker. Se ha implementado una clasi caci on en palabras f aciles y dif ciles dependiendo de c omo una persona lee, y una forma de saber si se est a leyendo el texto o pasando la vista por encima.

Dam wall detection and tracking using a Mechanically Scanned Imaging Sonar

In dam inspection tasks, an underwater robot has to grab images while surveying the wall meanwhile maintaining a certain distance and relative orientation. This paper proposes the use of an MSIS (mechanically scanned imaging sonar) for relative positioning of a robot with respect to the wall. An imaging sonar gathers polar image scans from which depth images (range & bearing) are generated. Depth scans are first processed to extract a line corresponding to the wall (with the Hough transform), which is then tracked by means of an EKF (Extended Kalman Filter) using a static motion model and an implicit measurement equation associating the sensed points to the candidate line. The line estimate is referenced to the robot fixed frame and represented in polar coordinates (rho&thetas) which directly corresponds to the actual distance and relative orientation of the robot with respect to the wall. The proposed system has been tested in simulation as well as in water tank conditions

A Monte Carlo-Based Fiber Tracking Algorithm using Diffusion Tensor MRI

Diffusion tensor magnetic resonance imaging, which measures directional information of water diffusion in the brain, has emerged as a powerful tool for human brain studies. In this paper, we introduce a new Monte Carlo-based fiber tracking approach to estimate brain connectivity. One of the main characteristics of this approach is that all parameters of the algorithm are automatically determined at each point using the entropy of the eigenvalues of the diffusion tensor. Experimental results show the good performance of the proposed approach

Controladores PI con acción de reset

El control automàtic exerceix un paper important en molts processos de la industria. Cada un dels sistemes de control requereix d’un controlador, la majoria dels quals són del tipus PI. L’objectiu d’aquest projecte es investigar tècniques que permetin superar les limitacions que tenen els controladors PI lineals. En la resposta d’un sistema de control es poden distingir dues tasques diferents: El seguiment a un canvi d’entrada o consigna correspon a la tasca de servo, mentre que el rebuig a pertorbacions correspon a la tasca de regulatori. Al típic esquema de control realimentat, aquestes dues tasques estan enfrontades, és a dir, una millora a la tasca de servo implica un empitjorament a la tasca de regulatori i a l’inversa. Això suposa un problema al rendiment del sistema, així com la necessitat d’establir un cert compromís entre les dues tasques. El que es pretén en aquest projecte es implementar senzilles regles de control no lineal amb la finalitat de millorar el rendiment del sistema i evitar la necessitat d’establir un compromís entre les dues tasques. Així, es pretén superar les limitacions que aquest té, obtenint controladors PI alternatius fàcilment sintetitzables.

Intelligent task-level grasp mapping for robot control

In the future, robots will enter our everyday lives to help us with various tasks.For a complete integration and cooperation with humans, these robots needto be able to acquire new skills. Sensor capabilities for navigation in real humanenvironments and intelligent interaction with humans are some of the keychallenges.Learning by demonstration systems focus on the problem of human robotinteraction, and let the human teach the robot by demonstrating the task usinghis own hands. In this thesis, we present a solution to a subproblem within thelearning by demonstration field, namely human-robot grasp mapping. Robotgrasping of objects in a home or office environment is challenging problem.Programming by demonstration systems, can give important skills for aidingthe robot in the grasping task.The thesis presents two techniques for human-robot grasp mapping, directrobot imitation from human demonstrator and intelligent grasp imitation. Inintelligent grasp mapping, the robot takes the size and shape of the object intoconsideration, while for direct mapping, only the pose of the human hand isavailable.These are evaluated in a simulated environment on several robot platforms.The results show that knowing the object shape and size for a grasping taskimproves the robot precision and performance

Chroma binary similarity and local alignment applied to cover song identification

We present a new technique for audio signal comparison based on tonal subsequence alignment and its application to detect cover versions (i.e., different performances of the same underlying musical piece). Cover song identification is a task whose popularity has increased in the Music Information Retrieval (MIR) community along in the past, as it provides a direct and objective way to evaluate music similarity algorithms.This article first presents a series of experiments carried outwith two state-of-the-art methods for cover song identification.We have studied several components of these (such as chroma resolution and similarity, transposition, beat tracking or Dynamic Time Warping constraints), in order to discover which characteristics would be desirable for a competitive cover song identifier. After analyzing many cross-validated results, the importance of these characteristics is discussed, and the best-performing ones are finally applied to the newly proposed method. Multipleevaluations of this one confirm a large increase in identificationaccuracy when comparing it with alternative state-of-the-artapproaches.

Multiple target tracking using random sets

This paper presents several algorithms for joint estimation of the target number and state in a time-varying scenario. Building on the results presented in [1], which considers estimation of the target number only, we assume that not only the target number, but also their state evolution must be estimated. In this context, we extend to this new scenario the Rao-Blackwellization procedure of [1] to compute Bayes recursions, thus defining reduced-complexity solutions for the multi-target set estimator. A performance assessmentis finally given both in terms of Circular Position Error Probability - aimed at evaluating the accuracy of the estimated track - and in terms of Cardinality Error Probability, aimed at evaluating the reliability of the target number estimates.

Using qualia information to identify lexical semantic classes in an unsupervised clustering task

Acquiring lexical information is a complex problem, typically approached by relying on a number of contexts to contribute information for classification. One of the first issues to address in this domain is the determination of such contexts. The work presented here proposes the use of automatically obtained FORMAL role descriptors as features used to draw nouns from the same lexical semantic class together in an unsupervised clustering task. We have dealt with three lexical semantic classes (HUMAN, LOCATION and EVENT) in English. The results obtained show that it is possible to discriminate between elements from different lexical semantic classes using only FORMAL role information, hence validating our initial hypothesis. Also, iterating our method accurately accounts for fine-grained distinctions within lexical classes, namely distinctions involving ambiguous expressions. Moreover, a filtering and bootstrapping strategy employed in extracting FORMAL role descriptors proved to minimize effects of sparse data and noise in our task.

Cognitive stimulation through task-oriented telerobotics

The project presented, iCognos, consists of a flexible platform to assist end-users in performing a series of mental tasks with a sensitized mobile telerobotic platform aimed at mitigating the problems associated to cognitive disorders with an ecological cognition approach.

Tracking the invisible hand: Convergence of double auctions to competitive equilibrium

Economics is the science of want and scarcity. We show that want andscarcity, operating within a simple exchange institution (double auction),are sufficient for an economy consisting of multiple inter--related marketsto attain competitive equilibrium (CE). We generalize Gode and Sunder's(1993a, 1993b) single--market finding to multi--market economies, andexplore the role of the scarcity constraint in convergence of economies to CE.When the scarcity constraint is relaxed by allowing arbitrageurs in multiple markets to enter speculative trades, prices still converge to CE,but allocative efficiency of the economy drops. \\Optimization by individual agents, often used to derive competitive equilibria,are unnecessary for an actual economy to approximately attain such equilibria.From the failure of humans to optimize in complex tasks, one need not concludethat the equilibria derived from the competitive model are descriptivelyirrelevant. We show that even in complex economic systems, such equilibriacan be attained under a range of surprisingly weak assumptions about agentbehavior.

Cultural differences on seeking information: an eye tracking study

The main goal of this research is to investigate how people with different cultural background differ in their interaction style and visual behavior on search engine results pages (SERP), more specifically between groups from the Middle Eastern region vs. Western Europe. We conducted a controlled eye-tracking experiment to explore and evaluate the visual behavior of Arabs and Spaniardusers when scanning through the first page of the search results in Google. Big differences can be observed in the 4 aspects studied: U.A.E. participants stayed on the SERPs for longer, they read more results and they read each snippet in a more complete way than Spaniards. In Spain, people tended to scan the SERP, reading less text on each snippet, and choose a result among the first top rankedones without hardly seeing those in bottom positions.