Abstract
Extreme and rare events, such as abnormal spikes in air pollution or weather conditions can have serious repercussions. Many of these sorts of events develop from spatio-temporal processes, and accurate predictions are a most valuable tool in addressing their impact, in a timely manner. In this paper, we propose a new set of resampling strategies for imbalanced spatiotemporal forecasting tasks, by introducing bias into formerly random processes. This spatio-temporal bias includes a hyperparameter that regulates the relative importance of the temporal and spatial dimensions in the selection of observations during under- or over-sampling. We test and compare our proposals against standard versions of the strategies on 10 different georeferenced numeric time series, using 3 distinct off-the-shelf learning algorithms. Experimental results show that our proposal provides an advantage over random resampling strategies in imbalanced spatio-temporal forecasting tasks. Additionally, we also find that valuing an observation's recency is more useful when over-sampling; while valuing its spatial distance to other cases with extreme values is more beneficial when under-sampling.

Abstract
We live in a digital society that needs new better prepared professionals for the new challenges and opportunities provided by the ICT. Students must learn how to deal with all the issues that emerge in this new context. They should acquire computational thinking skills by integrating STEAM, however this needs for changes in current learning curricula and also new learning approaches. RoboSTEAM project deals with this issue by the application of a Challenge Based Learning approach that uses Robotics and Physical Devices. One of the problems found during the project is the complexity of the application of a Challenge Based Learning approach due to the special needs of each educational institution. Given this situation the present work presents provides a flexible definition of challenge and describes also samples regarding how to use them. © 2019 ACM.

Abstract
In recent years, the intense urbanization, and consequently the traffic congestion, has been a major concern of large cities. In this context, the development of smart parkings is a suitable solution to deal with this problem. However, the complexity and requirements imposed by such large-scale systems are an obstacle to its easy implementation. In this sense, it is fundamental to adopt emergent ICT and Artificial Intelligence technologies that are capable to address the imposed requirements. Multi-Agent Systems (MAS) is a suitable approach to face this challenge by providing modularity, flexibility, reconfigurability and fast response to condition change based on its decentralized nature. The use of such agent-based solutions to control physical assets, create novel systems entitled Cyber-Physical Systems (CPS) where the interconnection between the cyber and the physical parts is a crucial issue. This paper focuses the interface between the software agents of a smart parking system with the physical control devices of the parking spots. For this purpose, different interface practices were implemented and tested, considering different interaction schemes and technologies. These alternative interface practices were analyzed taking into consideration the response time, scalability and re-usability parameters.

Abstract
Salinity measurement in water is typically performed with conductivity sensors. However, for long-term marine deployments, loss of precision is observed, mainly due to electrode drift (oxidation and degradation occurs in the presence of water, salts and bio-fouling), which results in inaccuracy of measurements. A cost-effective, low-power, four-probe salinity sensor is presented, to accurately measure long-term deployments in oceans, rivers and lakes. The four-probe methodology overcomes many of the drift problems, and the use of low-cost stainless-steel electrodes (avoiding platinum or titanium materials) can still achieve good long-term stability, in the practical salinity scale range from 2 to 42 PSU. Low-power electronics (200 µA in sleep-mode and 1 mA in active-mode) based on a ratiometric ADC conversion, and a low-power microcontroller with non-volatile memory, complements the proposed sensor, to achieve an autonomous salinity sensor for long-term marine deployments, with autonomy above 1 year with a 1 min-1 sample rate, using a common 2400 mA x 3.7 V lithium battery.

Abstract
Active contour models are one of the most emblematic algorithms of computer vision. Their strong theoretical foundations and high user interoperahility turned them into a reference approach for object segmentation and tracking tasks. A high number of modifications have already been proposed in order to overcome the known problems of traditional snakes, such as initialization dependence and poor convergence to concavities. In this paper, we address the scenario where the user wants to segment an object that has multiple dynamic regions but some of them do not correspond to the true object boundary. We propose a novel parametric active contour model, the Sparse Multi-Bending snake, which is capable of dividing the contour into a set of contiguous regions with different bending properties. We derive a new energy function that induces such behavior and presents a group optimization strategy that can be used to find the optimal bending resistance parameter for each point of the contour. We show the flexibility of our model in a set of synthetic images. In addition, we consider two real applications, lung segmentation in Computerized Tomography data and hand segmentation in depth images. We show how the proposed method is able to improve the segmentations obtained in both applications, when compared with other active contour models.

Abstract
The use of fossil fuel vehicles is one of the factors responsible for the degradation of air quality in urban areas. In order to reduce levels of air pollution in metropolitan areas, several countries have encouraged the use of electric vehicles in the cities. However, due to the high investment costs in this class of vehicles, it is expected that the spatial distribution of electric vehicles' adopters will be heterogeneous. The additional charging power required by electric vehicles' batteries can change operation and expansion planning of power distribution utilities. In addition, urban planning agencies should analyze the most suitable locations for the construction of electric vehicle recharging stations. Thus, in order to provide information for the planning of electric mobility services in the city, this paper presents a spatiotemporal model for estimating the rate of electric vehicles' adopters per subareas. Results are spatial databases that can be viewed in geographic information systems to observe regions with greater expectancy of residential electric vehicle adopters. These outcomes can help utilities to develop new services that ground on the rising availability of electric mobility in urban zones.