Abstract
This paper proposes the concepts of sliding mode schemes for fault tolerant control. The theoretical ideas developed in the paper were applied to a multi-motor electric vehicle (EV). To address this problem, we started by revisiting the modeling of vehicle dynamics focusing on nonlinear single-track model. The vehicle response to external perturbation forces, which are generated by unbalanced right and left traction forces due to malfunctions in motor drive, is described using the transfer functions analysis. The designing of a sliding mode controller for handling faults is described. The proposed scheme shows that certain motor drive failures can be handled directly without reconfiguring the controller. Simulation results obtained with CarSim vehicle model show the effectiveness of the fault tolerant control in various driving scenarios.

Abstract
Fault detection has been an open problem in power converters for several years. In this paper, we aim at detecting faults by means of left invertibility techniques. Using a mathematical model of the power converter it is possible to exploit the principle of injectivity of I/O map, which allows the recovery of unknown inputs applied to the system from the measure of the outputs. To achieve this end, we utilize the existing current and voltage sensors of the converter without the need for any additional sensors. Finally, a case-study of a multiport converter is presented and simulated to illustrate the methodology. Experimental results obtained under realistic conditions illustrate the effectiveness of the scheme and prove that fault detection based on the inverse method is possible.

Abstract
This paper presents a comparative study between two non-causal algorithms for the energy management problem of electric vehicles, endowed with batteries and supercapacitors(SCs). Toward that goal, an optimization-based energy problem is formulated, which targets the minimization of the source's energy losses throughout a given driving cycle. This problem is solved, firstly, with the help of a fast (but locally optimal) non-linear programming solver; and, secondly, with a slow, but globally optimal, dynamic programming (DP) approach. Simulation results will demonstrate that, despite the different theoretical properties associated with these two solver approaches, both generate similar solutions. In the second part of the work, we will develop a filter-based energy management algorithm, i.e., employ batteries to provide the low-frequency content of the power demand, while SCs cover the high-frequency demand. Our approach builds on the idea of adapting the filter's time constant throughout the vehicle's journey, using, for that purpose, a fuzzy logic algorithm and the information of the state of the vehicle. In comparison with the traditional fixed time-constant approach, the simulation results show that under some conditions the adaptive time-constant algorithm has the potential to reduce the energy losses of the sources by up to 62%. © 2013 IEEE.

Abstract

Abstract
In recent years, computer vision has been widely used on industrial environments, allowing robots to perform important tasks like quality control, inspection and recognition. Vision systems are typically used to determine the position and orientation of objects in the workstation, enabling them to be transported and assembled by a robotic cell (e.g. industrial manipulator). These systems commonly resort to CCD (Charge-Coupled Device) Cameras fixed and located in a particular work area or attached directly to the robotic arm (eye-in-hand vision system). Although it is a valid approach, the performance of these vision systems is directly influenced by the industrial environment lighting. Taking all these into consideration, a new approach is proposed for eye-on-hand systems, where the use of cameras will be replaced by the 2D Laser Range Finder (LRF). The LRF will be attached to a robotic manipulator, which executes a pre-defined path to produce grayscale images of the workstation. With this technique the environment lighting interference is minimized resulting in a more reliable and robust computer vision system. After the grayscale image is created, this work focuses on the recognition and classification of different objects using inherent features (based on the invariant moments of Hu) with the most well-known machine learning models: k-Nearest Neighbor (kNN), Neural Networks (NNs) and Support Vector Machines (SVMs). In order to achieve a good performance for each classification model, a wrapper method is used to select one good subset of features, as well as an assessment model technique called K-fold cross-validation to adjust the parameters of the classifiers. The performance of the models is also compared, achieving performances of 83.5% for kNN, 95.5% for the NN and 98.9% for the SVM (generalized accuracy). These high performances are related with the feature selection algorithm based on the simulated annealing heuristic, and the model assessment (k-fold cross-validation). It makes possible to identify the most important features in the recognition process, as well as the adjustment of the best parameters for the machine learning models, increasing the classification ratio of the work objects present in the robot's environment.

Abstract