Anàlisi, disseny i construcció d'un robot educatiu basat en microcontrolador AVR-ATMEGA

Aquest projecte presenta el disseny, construcció i programació d’un robot autònom, com a base per una proposta educativa. Per aconseguir aquest objectiu s’ha dotat el robot d’una unitat de procés, un sistema de locomoció i un seguit de sensors que proporcionaran a la unitat informació respecte l’entorn. Per gestionar totes aquestes funcionalitats, s’ha fet servir un sistema operatiu en temps real capaç de gestionar amb efectivitat les tasques que puguin ser executades pel robot. Finalment, s’ha exposat una detallada descripció dels costos per una producció de volum mig i de caire merament educatiu.

Estudi: microcontrolador LPC2119 - ARM. Cas pràctic: anàlisi, disseny i construcció d'un robot amb pèndul invertit

Estudi de l'arquitectura i prestacions del microcontrolador LPC2119 tot implementant la proposta d’un cas pràctic. En la besant teòrica, es fa una anàlisi acurada del dispositiu LPC2119, enumerant les principals característiques i exposant les seves parts, aprofundint sobretot en l’arquitectura i core ARM que incorpora. En l'àmbit pràctic, s'introdueix el problema del pèndul invertit com a proposta per a ser integrada sobre un robot que exploti les funcionalitats del dispositiu integrat presentades a l'estudi teòric.

Utilització del robot mòbil com a eina d'aprenentatge en ensenyaments tècnics

En aquest Projecte de Millora de la Qualitat Docent es descriu el disseny, la construcció i la utilització d’un robot mòbil com a eina docent en titulacions d’Enginyeria. El robot mòbil té com a element de control un PC portàtil convencional per tal de facilitar el procés d’aprenentatge de l’alumnat estigui centrat en l’objectiu de les pràctiques i no en el funcionament i control del robot. A més a més, el robot disposa d’un elevat nombre de sensors i actuadors per tal d’oferir un elevat grau d’interdisciplinaritat.

Trouble dissociatif: une clinique à l'interface de la neurologie et de la psychiatrie [Dissociative disorders: neurologists and psychiatrists working together]

Les troubles dissociatifs se présentent souvent par une clinique neurologique atypique impliquant une démarche diagnostique complexe à l'interface de la neurologie et de la psychiatrie. La restitution du diagnostic aux patients et leur prise en charge nécessitent une étroite collaboration interdisciplinaire. Les connaissances actuelles sont encore limitées, mais ce domaine est enrichi par des études récentes en neurosciences cliniques. Cet article présente les principaux aspects des troubles dissociatifs et formule un concept de prise en charge. Dissociative disorders often have an atypical neurological presentation requiring a complex diagnostic process at the interface between neurology and psychiatry. A strong interdisciplinary collaboration is needed for diagnosis restitution and patient treatment. Current knowledge is still scarce but recent studies in clinical neuroscience enrich this field. This article presents the main aspects of dissociative disorders and suggests a treatment framework

La consultation systémique: une interface entre recherche et clinique

Le but de la consultation systémique est l'évaluation des interactions familiales à des fins cliniques aussi bien qu'à des fins de recherche, avec mise en lumière des ressources aussi bien que des difficultés de la famille. Elle est soit demandée spontanément par les parents soit par le(s) thérapeute(s) qui sui(ven)t la famille. Lors d'une première rencontre, nous proposons d'une part à la famille de faire des jeux familiaux standardisés que nous filmons et d'autre part de poser les questions qui motivent les parents ou thérapeute(s) à nous consulter. Lors d'une deuxième rencontre, un visionnement des films avec la famille (et les thérapeutes) permet une discussion ainsi que l'élaboration de réponses aux questions posées. Après une description de la pratique de la consultation systémique, avec ses objectifs et ses principes, les situations utilisées dans ce contexte sont présentées (comme le Lausanne Trilogue Play). Enfin, une vignette clinique en illustre la richesse et l'utilité, aussi bien pour la recherche que pour la clinique. The aim of the systems consultation is to assess the family interaction in order to enlighten, in a clinical perspective, the resources as well as the difficulties of the family. The family itself or a therapist may request it. During the first session, we propose to the family, on the one hand, to play standardized games which are recorded and, on the other hand, to ask the questions they (or the therapist/s) may have. During the second session, a video feedback takes place to discuss and elaborate on the questions. After a description of the practice of the systems consultation, including aims and principles, the observational situations used in this context will be presented (e.g. the Lausanne Trilogue Play). Finally a case illustration will show its richness and usefulness for research as well as for clinical purposes, in particular as a bridge between research and clinical domains.

Molecular design of the Calphabeta interface favors specific pairing of introduced TCRalphabeta in human T cells.

A promising approach to adoptive transfer therapy of tumors is to reprogram autologous T lymphocytes by TCR gene transfer of defined Ag specificity. An obstacle, however, is the undesired pairing of introduced TCRalpha- and TCRbeta-chains with the endogenous TCR chains. These events vary depending on the individual endogenous TCR and they not only may reduce the levels of cell surface-introduced TCR but also may generate hybrid TCR with unknown Ag specificities. We show that such hybrid heterodimers can be generated even by the pairing of human and mouse TCRalpha- and TCRbeta-chains. To overcome this hurdle, we have identified a pair of amino acid residues in the crystal structure of a TCR that lie at the interface of associated TCR Calpha and Cbeta domains and are related to each other by both a complementary steric interaction analogous to a "knob-into-hole" configuration and the electrostatic environment. We mutated the two residues so as to invert the sense of this interaction analogous to a charged "hole-into-knob" configuration. We show that this inversion in the CalphaCbeta interface promotes selective assembly of the introduced TCR while preserving its specificity and avidity for Ag ligand. Noteworthily, this TCR modification was equally efficient on both a Mu and a Hu TCR. Our data suggest that this approach is generally applicable to TCR independently of their Ag specificity and affinity, subset distribution, and species of origin. Thus, this strategy may optimize TCR gene transfer to efficiently and safely reprogram random T cells into tumor-reactive T cells.

Differential epipolar constraint in mobile robot egomotion estimation

The estimation of camera egomotion is a well established problem in computer vision. Many approaches have been proposed based on both the discrete and the differential epipolar constraint. The discrete case is mainly used in self-calibrated stereoscopic systems, whereas the differential case deals with a unique moving camera. The article surveys several methods for mobile robot egomotion estimation covering more than 0.5 million samples using synthetic data. Results from real data are also given

Efficient learning of reactive robot behaviors with a Neural-Q_learning approach

The purpose of this paper is to propose a Neural-Q_learning approach designed for online learning of simple and reactive robot behaviors. In this approach, the Q_function is generalized by a multi-layer neural network allowing the use of continuous states and actions. The algorithm uses a database of the most recent learning samples to accelerate and guarantee the convergence. Each Neural-Q_learning function represents an independent, reactive and adaptive behavior which maps sensorial states to robot control actions. A group of these behaviors constitutes a reactive control scheme designed to fulfill simple missions. The paper centers on the description of the Neural-Q_learning based behaviors showing their performance with an underwater robot in a target following task. Real experiments demonstrate the convergence and stability of the learning system, pointing out its suitability for online robot learning. Advantages and limitations are discussed

Semi-online neural-Q_leaming for real-time robot learning

Reinforcement learning (RL) is a very suitable technique for robot learning, as it can learn in unknown environments and in real-time computation. The main difficulties in adapting classic RL algorithms to robotic systems are the generalization problem and the correct observation of the Markovian state. This paper attempts to solve the generalization problem by proposing the semi-online neural-Q_learning algorithm (SONQL). The algorithm uses the classic Q_learning technique with two modifications. First, a neural network (NN) approximates the Q_function allowing the use of continuous states and actions. Second, a database of the most representative learning samples accelerates and stabilizes the convergence. The term semi-online is referred to the fact that the algorithm uses the current but also past learning samples. However, the algorithm is able to learn in real-time while the robot is interacting with the environment. The paper shows simulated results with the "mountain-car" benchmark and, also, real results with an underwater robot in a target following behavior

Vision-based localization of an underwater robot in a structured environment

This paper presents a vision-based localization approach for an underwater robot in a structured environment. The system is based on a coded pattern placed on the bottom of a water tank and an onboard down looking camera. Main features are, absolute and map-based localization, landmark detection and tracking, and real-time computation (12.5 Hz). The proposed system provides three-dimensional position and orientation of the vehicle along with its velocity. Accuracy of the drift-free estimates is very high, allowing them to be used as feedback measures of a velocity-based low-level controller. The paper details the localization algorithm, by showing some graphical results, and the accuracy of the system

Towards Direct Policy Search Reinforcement Learning for Robot Control

This paper proposes a high-level reinforcement learning (RL) control system for solving the action selection problem of an autonomous robot. Although the dominant approach, when using RL, has been to apply value function based algorithms, the system here detailed is characterized by the use of direct policy search methods. Rather than approximating a value function, these methodologies approximate a policy using an independent function approximator with its own parameters, trying to maximize the future expected reward. The policy based algorithm presented in this paper is used for learning the internal state/action mapping of a behavior. In this preliminary work, we demonstrate its feasibility with simulated experiments using the underwater robot GARBI in a target reaching task

Estimating the motion of an underwater robot from a monocular image sequence

When underwater vehicles perform navigation close to the ocean floor, computer vision techniques can be applied to obtain quite accurate motion estimates. The most crucial step in the vision-based estimation of the vehicle motion consists on detecting matchings between image pairs. Here we propose the extensive use of texture analysis as a tool to ameliorate the correspondence problem in underwater images. Once a robust set of correspondences has been found, the three-dimensional motion of the vehicle can be computed with respect to the bed of the sea. Finally, motion estimates allow the construction of a map that could aid to the navigation of the robot