Effect of bandwidth knowledge of results on the learning of a grip force control task

An experiment was conducted to investigate the persistence of the effect of ""bandwidth knowledge of results (KR)"" manipulated during the learning phase of performing a manual force-control task. The experiment consisted of two phases, an acquisition phase with the goal of maintaining 60% maximum force in 30 trials, and a second phase with the objective of maintaining 40% of maximum force in 20 further trials. There were four bandwidths of KR: when performance error exceeded 5, 10, or 15% of the target, and a control group (0% bandwidth). Analysis showed that 5, 10, and 15% bandwidth led to better performance than 0% bandwidth KR at the beginning of the second phase and persisted during the extended trials.

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

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

New strings for old : the role of similarity processing in an incidental learning task

This paper reports three experiments that examine the role of similarity processing in McGeorge and Burton's (1990) incidental learning task. In the experiments subjects performed a distractor task involving four-digit number strings, all of which conformed to a simple hidden rule. They were then given a forced-choice memory test in which they were presented with pairs of strings and were led to believe that one string of each pair had appeared in the prior learning phase. Although this was not the case, one string of each pair did conform to the hidden rule. Experiment 1 showed that, as in the McGeorge and Burton study, subjects were significantly more likely to select test strings that conformed to the hidden rule. However, additional analyses suggested that rather than having implicitly abstracted the rule, subjects may have been selecting strings that were in some way similar to those seen during the learning phase. Experiments 2 and 3 were designed to try to separate out effects due to similarity from those due to implicit rule abstraction. It was found that the results were more consistent with a similarity-based model than implicit rule abstraction per se.

Learning curve for robotic-assisted laparoscopic colorectal surgery.

BACKGROUND: Robotic-assisted laparoscopic surgery (RALS) is evolving as an important surgical approach in the field of colorectal surgery. We aimed to evaluate the learning curve for RALS procedures involving resections of the rectum and rectosigmoid. METHODS: A series of 50 consecutive RALS procedures were performed between August 2008 and September 2009. Data were entered into a retrospective database and later abstracted for analysis. The surgical procedures included abdominoperineal resection (APR), anterior rectosigmoidectomy (AR), low anterior resection (LAR), and rectopexy (RP). Demographic data and intraoperative parameters including docking time (DT), surgeon console time (SCT), and total operative time (OT) were analyzed. The learning curve was evaluated using the cumulative sum (CUSUM) method. RESULTS: The procedures performed for 50 patients (54% male) included 25 AR (50%), 15 LAR (30%), 6 APR (12%), and 4 RP (8%). The mean age of the patients was 54.4 years, the mean BMI was 27.8 kg/m(2), and the median American Society of Anesthesiologists (ASA) classification was 2. The series had a mean DT of 14 min, a mean SCT of 115.1 min, and a mean OT of 246.1 min. The DT and SCT accounted for 6.3% and 46.8% of the OT, respectively. The SCT learning curve was analyzed. The CUSUM(SCT) learning curve was best modeled as a parabola, with equation CUSUM(SCT) in minutes equal to 0.73 × case number(2) - 31.54 × case number - 107.72 (R = 0.93). The learning curve consisted of three unique phases: phase 1 (the initial 15 cases), phase 2 (the middle 10 cases), and phase 3 (the subsequent cases). Phase 1 represented the initial learning curve, which spanned 15 cases. The phase 2 plateau represented increased competence with the robotic technology. Phase 3 was achieved after 25 cases and represented the mastery phase in which more challenging cases were managed. CONCLUSIONS: The three phases identified with CUSUM analysis of surgeon console time represented characteristic stages of the learning curve for robotic colorectal procedures. The data suggest that the learning phase was achieved after 15 to 25 cases.

Self-learning of fault diagnosis identification

A good and early fault detection and isolation system along with efficient alarm management and fine sensor validation systems are very important in today¿s complex process plants, specially in terms of safety enhancement and costs reduction. This paper presents a methodology for fault characterization. This is a self-learning approach developed in two phases. An initial, learning phase, where the simulation of process units, without and with different faults, will let the system (in an automated way) to detect the key variables that characterize the faults. This will be used in a second (on line) phase, where these key variables will be monitored in order to diagnose possible faults. Using this scheme the faults will be diagnosed and isolated in an early stage where the fault still has not turned into a failure.

Incorporating curvature information into on-line learning

We analyse the dynamics of a number of second order on-line learning algorithms training multi-layer neural networks, using the methods of statistical mechanics. We first consider on-line Newton's method, which is known to provide optimal asymptotic performance. We determine the asymptotic generalization error decay for a soft committee machine, which is shown to compare favourably with the result for standard gradient descent. Matrix momentum provides a practical approximation to this method by allowing an efficient inversion of the Hessian. We consider an idealized matrix momentum algorithm which requires access to the Hessian and find close correspondence with the dynamics of on-line Newton's method. In practice, the Hessian will not be known on-line and we therefore consider matrix momentum using a single example approximation to the Hessian. In this case good asymptotic performance may still be achieved, but the algorithm is now sensitive to parameter choice because of noise in the Hessian estimate. On-line Newton's method is not appropriate during the transient learning phase, since a suboptimal unstable fixed point of the gradient descent dynamics becomes stable for this algorithm. A principled alternative is to use Amari's natural gradient learning algorithm and we show how this method provides a significant reduction in learning time when compared to gradient descent, while retaining the asymptotic performance of on-line Newton's method.

An incremental multilinear system for human face learning and recognition

This dissertation establishes a novel system for human face learning and recognition based on incremental multilinear Principal Component Analysis (PCA). Most of the existing face recognition systems need training data during the learning process. The system as proposed in this dissertation utilizes an unsupervised or weakly supervised learning approach, in which the learning phase requires a minimal amount of training data. It also overcomes the inability of traditional systems to adapt to the testing phase as the decision process for the newly acquired images continues to rely on that same old training data set. Consequently when a new training set is to be used, the traditional approach will require that the entire eigensystem will have to be generated again. However, as a means to speed up this computational process, the proposed method uses the eigensystem generated from the old training set together with the new images to generate more effectively the new eigensystem in a so-called incremental learning process. In the empirical evaluation phase, there are two key factors that are essential in evaluating the performance of the proposed method: (1) recognition accuracy and (2) computational complexity. In order to establish the most suitable algorithm for this research, a comparative analysis of the best performing methods has been carried out first. The results of the comparative analysis advocated for the initial utilization of the multilinear PCA in our research. As for the consideration of the issue of computational complexity for the subspace update procedure, a novel incremental algorithm, which combines the traditional sequential Karhunen-Loeve (SKL) algorithm with the newly developed incremental modified fast PCA algorithm, was established. In order to utilize the multilinear PCA in the incremental process, a new unfolding method was developed to affix the newly added data at the end of the previous data. The results of the incremental process based on these two methods were obtained to bear out these new theoretical improvements. Some object tracking results using video images are also provided as another challenging task to prove the soundness of this incremental multilinear learning method.

Robotic Manipulation of DLOs for Wiring Harness Assembly: a Machine Learning Approach

In recent times, a significant research effort has been focused on how deformable linear objects (DLOs) can be manipulated for real world applications such as assembly of wiring harnesses for the automotive and aerospace sector. This represents an open topic because of the difficulties in modelling accurately the behaviour of these objects and simulate a task involving their manipulation, considering a variety of different scenarios. These problems have led to the development of data-driven techniques in which machine learning techniques are exploited to obtain reliable solutions. However, this approach makes the solution difficult to be extended, since the learning must be replicated almost from scratch as the scenario changes. It follows that some model-based methodology must be introduced to generalize the results and reduce the training effort accordingly. The objective of this thesis is to develop a solution for the DLOs manipulation to assemble a wiring harness for the automotive sector based on adaptation of a base trajectory set by means of reinforcement learning methods. The idea is to create a trajectory planning software capable of solving the proposed task, reducing where possible the learning time, which is done in real time, but at the same time presenting suitable performance and reliability. The solution has been implemented on a collaborative 7-DOFs Panda robot at the Laboratory of Automation and Robotics of the University of Bologna. Experimental results are reported showing how the robot is capable of optimizing the manipulation of the DLOs gaining experience along the task repetition, but showing at the same time a high success rate from the very beginning of the learning phase.

Levantamento, reparação e beneficiação do simulador TERCO PST 2200 do Laboratório de Sistemas de Energia do ISEP

Este trabalho é baseado no simulador de redes PST2200 do Laboratório de Sistemas de Energia (LSE) pois está avariado com vários problemas conhecidos, designadamente: - Defeito de isolamento (disparo de diferencial), - Desregulação da velocidade da máquina primária (motor DC), - Circuito de excitação da máquina síncrona inoperacional, - Inexistência de esquemas elétricos dos circuitos do simulador, - Medidas desreguladas e com canais de medida com circuito impresso queimado. O trabalho executado foi: - O levantamento e desenho de raiz (não existe qualquer manual) dos esquemas dos 10 módulos do simulador, designadamente naqueles com avaria ou com desempenho problemático a fim de que se possa ter uma visão mais pormenorizada dos circuitos e seus problemas, por forma a intervir para os minimizar e resolver, - Foi realizado o diagnóstico de avaria do simulador e foram propostas soluções para os mesmos, - Realizaram-se as intervenções propostas e aprovadas. Nas intervenções realizadas, os princípios orientadores foram: - Aumentar a robustez do equipamento por forma a garantir a sua integridade a utilizações menos apropriados e manobras 'exóticas' próprias de alunos, que pela sua condição, estão em fase de aprendizagem, - Atualizar o equipamento, colocando-o em sintonia com o 'estado da arte', - Como fator de valorização suplementar, foi concebida e aplicada a supervisão remota do funcionamento do simulador através da rede informática. Foram detetados inúmeros erros: - Má ligação do motor de corrente continua ao variador, resultando a falta de controlo da frequência da rede do sistema, - Ligações entre painéis trocadas resultando em avarias diversas das fontes de alimentação, - Cartas eletrónicas de medidas avariadas e que além de se reparar, foram também calibradas. Devido ao mecenato da empresa Schnitt + Sohn participando monetariamente, fez-se o projeto de alteração e respetiva execução de grande parte do simulador aumentando a fiabilidade do mesmo, diminuindo assim a frequência das avarias naturais mais as que acontecem involuntariamente devido a este ser um instrumento didático. Além do trabalho elétrico, foi feito muito trabalho de chaparia para alteração de estrutura e suporte do material com diferenças de posicionamento. Neste trabalho dá-se também alguns exemplos de cálculo e simulação das redes de transporte que se pode efetuar no simulador como estudo e simulação de avarias num sistema produtivo real. Realizou-se a monitorização de dois aparelhos indicadores de parâmetros de energia (Janitza UMG96S) através duma rede com dois protocolos ethernet e profibus utilizando o plc (Omron CJ2M) como valorização do trabalho.

Data gathering approach in dense sensor networks

Sensor/actuator networks promised to extend automated monitoring and control into industrial processes. Avionic system is one of the prominent technologies that can highly gain from dense sensor/actuator deployments. An aircraft with smart sensing skin would fulfill the vision of affordability and environmental friendliness properties by reducing the fuel consumption. Achieving these properties is possible by providing an approximate representation of the air flow across the body of the aircraft and suppressing the detected aerodynamic drags. To the best of our knowledge, getting an accurate representation of the physical entity is one of the most significant challenges that still exists with dense sensor/actuator network. This paper offers an efficient way to acquire sensor readings from very large sensor/actuator network that are located in a small area (dense network). It presents LIA algorithm, a Linear Interpolation Algorithm that provides two important contributions. First, it demonstrates the effectiveness of employing a transformation matrix to mimic the environmental behavior. Second, it renders a smart solution for updating the previously defined matrix through a procedure called learning phase. Simulation results reveal that the average relative error in LIA algorithm can be reduced by as much as 60% by exploiting transformation matrix.

Relatório de estágio curricular realizado na Câmara Municipal de Lisboa

The experience of an internship is always considered an experience that should be enriching, a first contact with the employment world. The intention is to build a bridge between the theory and practice - applying what has been learnt throughout the course. Therefore, it was started a new learning phase that seeks to take advantage and obtain experiences useful for a professional life based in efficiency and efficacy. The curricular internship was carried out within the Masters in Public Law and held by Faculdade de Direito da Universidade Nova de Lisboa (FDUNL)(Law School Of University New Of Lisbon) in the Câmara Municipal de Lisboa (CML) (Lisbon's Town Hall), in the Departamento de Política de Solos e Valorização Patrimonial (DPSVP) (Soil Politics and Patrimonial Valuation Department) of the Direção Municipal de Planeamento, Reabilitação e Gestão Urbanística (Municipal Direction of Planning, Rehabilitation and Urban Management) between September and December 2013. Throughout this internship, several activities within the DPSVP were developed. In an early stage, there was a presentation of the organic structure of Lisbon's Town Hall, and the Department where I was going to do the internship, and its competencies. Therefore I acquired key concepts and researched jurisprudence and legislation needed to the analysis and understanding of the activities done in the internship. In a second stage, it was done the analysis and understanding of the division into lots administrative procedures, as well as the solving of the problems occurred throughout that analysis. Besides that, there was the need to help preparing some procedural acts to be applied regarding the Department competencies, namely within the alienation, procurement, encumbrance and rental of immovable assets owned by the Municipality of Lisbon.

Relatório de estágio curricular realizado na Câmara Municipal de Lisboa

The experience of an internship is always considered an experience that should be enriching, a first contact with the employment world. The intention is to build a bridge between the theory and practice - applying what has been learnt throughout the course. Therefore, it was started a new learning phase that seeks to take advantage and obtain experiences useful for a professional life based in efficiency and efficacy. The curricular internship was carried out within the Masters in Public Law and held by Faculdade de Direito da Universidade Nova de Lisboa (FDUNL)(Law School Of University New Of Lisbon) in the Câmara Municipal de Lisboa (CML) (Lisbon's Town Hall), in the Departamento de Política de Solos e Valorização Patrimonial (DPSVP) (Soil Politics and Patrimonial Valuation Department) of the Direção Municipal de Planeamento, Reabilitação e Gestão Urbanística (Municipal Direction of Planning, Rehabilitation and Urban Management) between September and December 2013. Throughout this internship, several activities within the DPSVP were developed. In an early stage, there was a presentation of the organic structure of Lisbon's Town Hall, and the Department where I was going to do the internship, and its competencies. Therefore I acquired key concepts and researched jurisprudence and legislation needed to the analysis and understanding of the activities done in the internship. In a second stage, it was done the analysis and understanding of the division into lots administrative procedures, as well as the solving of the problems occurred throughout that analysis. Besides that, there was the need to help preparing some procedural acts to be applied regarding the Department competencies, namely within the alienation, procurement, encumbrance and rental of immovable assets owned by the Municipality of Lisbon.

Equilibrium selection through incomplete information in coordination games: An experimental study

We perform an experiment on a pure coordination game with uncertaintyabout the payoffs. Our game is closely related to models that have beenused in many macroeconomic and financial applications to solve problemsof equilibrium indeterminacy. In our experiment each subject receives anoisy signal about the true payoffs. This game has a unique strategyprofile that survives the iterative deletion of strictly dominatedstrategies (thus a unique Nash equilibrium). The equilibrium outcomecoincides, on average, with the risk-dominant equilibrium outcome ofthe underlying coordination game. The behavior of the subjects convergesto the theoretical prediction after enough experience has been gained. The data (and the comments) suggest that subjects do not apply through"a priori" reasoning the iterated deletion of dominated strategies.Instead, they adapt to the responses of other players. Thus, the lengthof the learning phase clearly varies for the different signals. We alsotest behavior in a game without uncertainty as a benchmark case. The gamewith uncertainty is inspired by the "global" games of Carlsson and VanDamme (1993).

Anàlisi i disseny d'una aplicació gestora de curses via GPS per a dispositius mòbils amb SO Android

Aquest treball consisteix en analitzar la viabilitat d'una aplicació gestora de curses via GPS per a dispositius Android i, en cas afirmatiu, desenvolupar-ne un prototip. Per tal d'aconseguir-ho, es passa per una sèrie de fases d'estudi de l'aplicació i d'aprenentatge en la implementació.