Abstract
The article will present the development of the tool FEUPAutom, used at the Faculty of Engineering of the University of Porto (FEUP) in the automation engineering Technological & Scientific area. In FEUP, the pressure to deliver well trained engineers is steadily high in the last two decades, thus producing a well-known situation of massification in Higher Education Institutions, namely in engineering degrees. In the school year of 2013/14, the course where the tool was used had about 270 students, despite quite low retention rates. The article also includes a brief characterization of the engineering program, course and expected outcomes in full alignment with the ideas promoted by the EUR-ACE referential for the accreditation of engineering programs and also in strong consonance with the ideas defended by the Bologna process. The course includes lab work and a part of those uses Problem Based Learning (PBL) methodology. In the last decade, the professors of the mentioned course have tried to limit the usage of real world industrial equipments because of budget concerns, always without hindering the learning process. Adequate simulation tools were sought on the market but not found, mainly because the needs of a full blown engineer are frequently not the same as those of an early engineering student. At that point, the decision was made to develop an in-house tool, adequate for students. Industrial-grade equipment was not totally set aside, only reserved for latter stages and the actual usage strategy allowed the number of equipments to be halved. The article will go on briefly describing the FEUPAutom tool and new strategies available for lab classes and PBL. As control groups would be unethical, students' quiz data from the two last editions of the course are used to evaluate learning (self-assessed). Grading strategy and coordination with the university's LMS is also addressed. Final grades of the course and satisfaction are also discussed. The students' assessment is that the FEUPAutom tool is very useful for the learning process and easier to use than the available industrial counterpart. Continuous improvement efforts have tried to push students to adequate PBL work only possible with the tool, with some results hinting deep learning in the technical area at stake. Some final thoughts, lessons learned and future work are also present in the article.

Abstract
This paper presents a novel localization method for small mobile robots. The proposed technique is especially designed for the Robot@Factory, a new robotic competition which is started in Lisbon in 2011. The real-time localization technique resorts to low-cost infra-red sensors, a map-matching method and an Extended Kalman Filter (EKF) to create a pose tracking system that performs well. The sensor information is continuously updated in time and space according to the expected motion of the robot. Then, the information is incorporated into the map-matching optimization in order to increase the amount of sensor information that is available at each moment. In addition, the Particle Swarm Optimization (PSO) relocates the robot when the map-matching error is high, meaning that the map-matching is unreliable and the robot gets lost. The experiments presented in this paper prove the ability and accuracy of the presented technique to locate small mobile robots for this competition. Extensive results show that the proposed method presents an interesting localization capability for robots equipped with a limited amount of sensors, but also less reliable sensors.

Abstract
In the past few years, cable-driven robots have received some attention by the scientific community and the industry. They have special characteristics that made them very reliable to operate with the level of safeness that is required by different environments, such as, handling of hazardous materials in construction sites. This paper presents a cable-driven robot called SPIDERobot, that was developed for automated construction of architectural projects. This robot has a rotating claw and it is controlled by a set of 4 cables that allow 4 degrees of freedom. In addition to the robot, this paper introduces a Dynamic Control System (DCS) that controls the positioning of the robot and assures that the length of cables is always within a safe value. Results show that traditional force-feasible approaches are more influenced by the pulling forces or the geometric arrangement of all cables and their positioning is significantly less accurate than the DCS. Therefore, the architecture of the SPIDERobot is designed to enable an easily scaling up of the solution to higher dimensions for operating in realistic environments.

Abstract
There are several approaches to create the Humanoid robot gait planning. This problem presents a large number of unknown parameters that should be found to make the humanoid robot to walk. Optimization in simulation models can be used to find the gait based on several criteria such as energy minimization, acceleration, step length among the others. The energy consumption can also be reduced with elastic elements coupled to each joint. The presented paper addresses an optimization method, the Stretched Simulated Annealing, that runs in an accurate and stable simulation model to find the optimal gait combined with elastic elements. Final results demonstrate that optimization is a valid gait planning technique.

Abstract
This paper describes a procedure applied to model DC motors. An example of the procedure apply is shown for a 12V brushed DC motor, equipped with a 29:1 metal gearbox and an integrated quadrature encoder. It is described the developed setup applied to obtain the experimental data and the developed algorithm applied to estimate the actuator parameters. It was obtained an electro-mechanical dynamical model that describes the motor, its gear box and the encoder. The motivation to develop a simple and easy to assemble procedure that allows to model DC motors is due to the fact that these actuators are intensively used in mobile robotics, being realistic simulation, based in accurate sensor and actuator models, the key to speed up Robot Software developing time.

Abstract
This research studies motion segmentation based on dense optical flow fields for mobile robotic applications. The optical flow is usually represented in the Euclidean space however, finding the most suitable motion space is a relevant problem because techniques for motion analysis have distinct performances. Factors like the processing-time and the quality of the segmentation provide a quantitative evaluation of the clustering process. Therefore, this paper defines a methodology that evaluates and compares the advantage of clustering dense flow fields using different feature spaces, for instance, Euclidean and Polar space. The methodology resorts to conventional clustering techniques, Expectation-Maximization and K-means, as baseline methods. The experiments conducted during this paper proved that the K-means clustering is suitable for analyzing dense flow fields.