Abstract
This paper presents results of the INESC TEC participation in the maritime environment (both at surface and underwater) integrated in the ICARUS team in the euRathlon 2015 robotics search and rescue competition. These relate to the marine robots from INESC TEC, surface (ROAZ USV) and underwater (MARES AUV) autonomous vehicles participation in multiple tasks such as situation assessment, underwater mapping, leak detection or victim localization. This participation was integrated in the ICARUS Team resulting of the EU funded project aimed to develop robotic tools for large scale disasters. The coordinated search and rescue missions were performed with an initial surface survey providing data for AUV mission planning and execution. A situation assessment bathymetry map, sidescan sonar imaging and location of structures, underwater leaks and victims were achieved, with the global ICARUS team (involving sea, air and land coordinated robots) participating in the final grand Challenge and achieving the second place.

Abstract

Abstract
In this paper we introduce three main features extracted from Moodle logs in order to be uses a possible means to predict future student grades. We discuss the statistical analysis on these features and show how they cannot be applied isolatedly to model our data. We then apply them as a whole and use principal component analysis to derive a decision tree based on the features. With derived tree we are able to predict grades in three intervals, namely to predict failures. Our proposed analysis methodology can be incorporated in an LMS and be used during a course. As the course unfolds, the system can to trigger alarms regarding possible failure situations.

Abstract
The efficiency of Public Transportation (PT) Networks is a major goal of any urban area authority. Advances on both location and communication devices drastically increased the availability of the data generated by their operations. Adequate Machine Learning methods can thus be applied to identify patterns useful to improve the Schedule Plan. In this paper, the authors propose a fully automated learning framework to determine the best Schedule Coverage to be assigned to a given PT network based on Automatic Vehicle location (AVL) and Automatic Passenger Counting (APC) data. We formulate this problem as a clustering one, where the best number of clusters is selected through an ad-hoc metric. This metric takes into account multiple domain constraints, computed using Sequence Mining and Probabilistic Reasoning. A case study from a large operator in Sweden was selected to validate our methodology. Experimental results suggest necessary changes on the Schedule coverage. Moreover, an impact study was conducted through a large-scale simulation over the affected time period. Its results uncovered potential improvements of the schedule reliability on a large scale. © Springer International Publishing Switzerland 2016.

Abstract
Assistive technology enables people to achieve independence when performing daily tasks and it enhances their overall quality of life. Visual information is the basis for most navigational tasks, so visually impaired individuals are at disadvantage due to the lack of sufficient information about their surrounding environment. With recent advances in inclusive technology it is possible to extend the support given to people with visual disabilities in terms of their mobility. In this context we present and describe a wearable system (Blavigator project), whose global objective is to assist visually impaired people in their navigation on indoor and outdoor environments. This paper is focused mainly on the Computer Vision module of the Blavigator prototype. We propose an object collision detection algorithm based on stereo vision. The proposed algorithm uses Peano-Hilbert Ensemble Empirical Mode Decomposition (PHEEMD) for disparity image processing and a two layer disparity image segmentation to detect nearby objects. Using the adaptive ensemble empirical mode decomposition (EEMD) image analysis real time is not achieved, with PH-EEMD results on a fast implementation suitable for real time applications.

Abstract
We introduce in this paper visibility maps for robots of any shape, representing the reachability limit of the robot's motion and sensing in a 2D gridmap with obstacles. The brute-force approach to determine the optimal visibility map is computationally expensive, and prohibitive with dynamic obstacles. We contribute the Robot-Dependent Visibility Map (RDVM) as a close approximation to the optimal, and an effective algorithm to compute it. The RDVM is a function of the robot's shape, initial position, and sensor model. We first overview the computation of RDVM for the circular robot case in terms of the partial morphological closing operation and the optimal choice for the critical points position. We then present how the RDVM for any-shape robots is computed. In order to handle any robot shape, we introduce in the first step multiple layers that discretize the robot orientation. In the second step, our algorithm determines the frontiers of actuation, similarly to the case of the the circular robot case. We then derive the concept of critical points to the any-shape robot, as the points that maximize expected visibility inside unreachable regions. We compare our method with the ground-truth in a simulated map compiled to capture a variety of challenges of obstacle distribution and type, and discuss the accuracy of our approximation to the optimal visibility map.