Abstract
We describe a data mining workflow for predictive maintenance of the Air Pressure System in heavy trucks. Our approach is composed by four steps: (i) a filter that excludes a subset of features and examples based on the number of missing values (ii) a metafeatures engineering procedure used to create a meta-level features set with the goal of increasing the information on the original data; (iii) a biased sampling method to deal with the class imbalance problem; and (iv) boosted trees to learn the target concept. Results show that the metafeatures engineering and the biased sampling method are critical for improving the performance of the classifier.

Abstract
In this paper, a multi energy system (MES) model incorporating the traffic behavior of plug-in electric vehicles (PEVs) is proposed. It is assumed that in a micro MES two charging options are available for the PEVs: the home charging (HC) stations and the PEV parking lot (PL). The operation of these elements within the micro MES concept is studied. The matrix model of the micro MES is adapted to enable the integration of PL and HC. Moreover, the traffic flow of the PEVs is added to the model as an input to the micro MES. The model is tested for various case studies and possible traffic behavior between the PL and HC. The results show that the presence of these two elements leads to effective integration of reduced system operation costs.

Abstract
The aim of this work is to develop an application for Android able to classifying urban sounds in a real life context. It also enables the collection and classification of new sounds. To train our classifier we use the UrbanSound8K data set available online. We have used a hybrid approach to obtain features, by combining SAX-based multiresolution motif discovery with Mel-Frequency Cepstral Coefficients (MFCC). We also describe different configurations of motif discovery for defining attributes and compare the use of Random Forest and SVM algorithms on this kind of data.

Abstract
Cyber-Physical Systems (CPS) is an emergent approach that focuses on the integration of computational applications with physical devices, being designed as a network of interacting cyber and physical elements. CPS control and monitor real-world physical infrastructures and thus is starting having a high impact in industrial automation. As such design, implementation and operation of CPS and management of the resulting automation infrastructure is of key importance for the industry. In this work, an overview of key aspects of industrial CPS, their technologies and emerging directions, as well as challenges for their implementation is presented. Based on the hands-on experiences gathered from four European, innovation projects over the last decade (i.e. SOCRADES, IMC-AESOP, GRACE and ARUM), a key challenges have been identified and a prioritization and timeline are pointed out with the aim to increase Technology Readiness Levels and lead to their usage in industrial automation environments.

Abstract
Our work proposes a hardware architecture for a Long Short-Term Memory (LSTM) Neural Network, aiming to outperform software implementations, by exploiting its inherent parallelism. The main design decisions are presented, along with the proposed network architecture. A description of the main building blocks of the network is also presented. The network is synthesized for various sizes and platforms, and the performance results are presented and analyzed. Our synthesized network achieves a 251 times speed-up over a custom-built software network, running on an i7-3770k Desktop computer, proving the benefits of parallel computation for this kind of network.

Abstract
This study presents a scheme to detect and isolate faults in over-actuated electric vehicles. Although this research work is still emerging, it already provides a view of the main challenges on the problem and discusses some possible approaches that can be useful to overcome the key difficulties. This paper intends to present a fault detection algorithm based on Unknown Input Observer (UIO). The residuals are built through the difference of signals between the measured outputs and the output estimations from the observer. The main idea is to detect fault in the electric motors and steering wheel actuator. The algorithm is presented and tested with some fault scenarios using the co-simulation tool between Simulink/MATLAB and the high-fidelity model from Carsim software.