Search

© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting /republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works This paper presents a health-aware Model Predictive Control (MPC) including fault-tolerant capabilities for drinking water transport networks. When a fault has occurred, the predictive controller must be redesigned to deal with the fault. This is done by considering the system reliability that is incorporated into the MPC model. The inclusion of components and system reliability in the MPC model is done through the Linear Parameter Varying (LPV) modelling approach with the aim of maximizing the availability of the system by considering system reliability. As a result, the MPC design is modified by considering the reliability model such that additionally to achieve the best achievable performance after the fault, the controller try to preserve the remaining useful life. The solution to the optimization problem related to the MPC problem is achieved by solving a series of Quadratic Programming (QP) problems thanks to the proposed LPV formulation. The proposed approach is applied to a part of a real drinking water transport network of Barcelona for demonstrating the performance of the method.

Not all video frames are equally informative for recognizing an action. It is computationally infeasible to train deep networks on all video frames when actions develop over hundreds of frames. A common heuristic is uniformly sampling a small number of video frames and using these to recognize the action. Instead, here we propose full video action recognition and consider all video frames. To make this computational tractable, we first cluster all frame activations along the temporal dimension based on their similarity with respect to the classification task, and then temporally aggregate the frames in the clusters into a smaller number of representations. Our method is end-to-end trainable and computationally efficient as it relies on temporally localized clustering in combination with fast Hamming distances in feature space. We evaluate on UCF101, HMDB51, Breakfast, and Something-Something V1 and V2, where we compare favorably to existing heuristic frame sampling methods., Accepted to CVPR 2021

Parkinson's disease (PD) is a complex neurode-generative disorder that results into poverty of movements. Freezing of gait (FOG) is a common and severely incapacitating symptom of PD. However, the underlying neurophysiological mechanism of FOG is still unclear. Electroencephalogram (EEG) as a non-invasive brain signal measurement method has been repeatedly used to reveal the pathological reasons behind PD. Since PD affects movement abilities, one particular type of EEG pattern called Movement-related cortical potential (MRCP), which was shown to be related to the generation of movement intention, can be a useful tool in PD research. However, most MRCP studies in PD area focus on the upper-limb movement like finger movement, to the best knowledge of the authors, no study explored MRCP characteristics during lower extremity movement in PD patients, particularly PD patients with FOG. This paper intends to investigate the relationship between the characteristics of MRCP and FOG by comparing MRCP signals among three groups: healthy controls group, PD patients without FOG group, and PD patients with FOG group. These three groups of participants were recruited based on age-matching and gender proportion matching criteria. During the experiment, the Electromyography (EMG) from the tibialis anterior (TA) muscle and EEG were recorded when performing ankle dorsiflexion (AD) by the dominant foot. MRCP signals were extracted from EEG data based on the muscle activities in the TA muscle of the dominant foot. The results of this study show that, during lower limb activities, MRCP shares considerable similarities between PD patients and healthy participants. However, the amplitude of MRCP in PD patients (with and without FOG) is lower compared to the healthy participants. In addition, direct observations from the results show that MRCP from PD patients with FOG has higher amplitude but more variability in terms of amplitude and latency than those from PD patients without FOG.

Trabajo presentado en la European Control Conference (ECC), celebrada en San Petersburgo (Rusia), del 12 al 15 de mayo de 2020, This paper proposed an economic reliabilityaware model predictive control (MPC) for the management of drinking water transport networks (DWNs) that includes a new goal to increase the system and components reliability based on a finite horizon stochastic optimization problem with joint probabilistic (chance) constraints. The proposed approach is based on a single-layer economic optimization problem with dynamic constraints. The inclusion of components and system reliability in the MPC model using an LPV modelling approach aims at maximizing the availability of the system by estimating system reliability. The solution of the optimization problem related to the MPC problem is obtained by solving a series of Quadratic Programming (QP) problem. The use of chanceconstraint programming allows computing an optimal policy based on a desirable risk acceptability level and managing dynamically volume tank stocks to cope with non-stationary flow demands. Finally, the proposed approach is applied to a part of a real drinking water transport network of Barcelona for demonstrating the performance of the method., This work has been partially funded by the Spanish State Research Agency (AEI) and the European Regional Development Fund (ERFD) through the projects DEOCS (ref. MINECO DPI2016-76493) and SCAV (ref. MINECO DPI2017-88403-R)

Brain-Computer Interface (BCI) systems aim at providing a channel through electroencephalogram (EEG) to control external devices. Movement-Related Cortical Potential (MRCP) is a low-frequency and low-amplitude EEG component that has been recently introduced to control real-time BCI systems. The performance of BCI systems based on MRCP highly depends on the accuracy of single trial MRCP detection, which is a challenging task due to high amount of EEG background noise and various types of artifacts. In this paper, a semi-blind source extraction algorithm based on second order statistics and a reference signal, designed according to the BCI paradigm, is presented to extract MRCP. The performance of the method is investigated during a complex motor task, gait initiation. The reference signal utilizes the temporal information from the training set to extract MRCP. The proposed method is compared with two commonly used spatial filtering techniques as well as two source extraction methods based on second order statistics. Two different training sets including only ankle dorsiflexion data (AD) and ankle dorsiflexion plus gait initiation (GI) data was used to train the proposed algorithm and the methods that require training. The performance of the proposed algorithm and other extraction methods on the extracted MRCP was evaluated by a performance index (PI) quantifying the signal-to-noise ratio of the extracted MRCP. The PI obtained with the proposed algorithm trained with AD only training set $(2.43 \pm 1.23)$ was greater than all other investigated methods. When trained with both AD and GI trials, the performance of the method was the second best among all methods ($2.52 \pm 0.83)$. The results from this study show that the proposed method is robust against differences between training and testing sets and provides high quality MRCP extraction.

Trabajo presentado en el 18th European Control Conference (ECC), celebrado en Nápoles (Italia), del 25 al 28 de junio de 2019, This paper presents a Model Predictive Control (MPC) strategy based on Linear Parameter Varying (LPV) models with varying delays affecting states and inputs. The proposed control approach allows the controller to accommodate the scheduling parameters and delay change. By computing the prediction of the state variables and delay along a prediction time horizon, the system model can be modified according to the evaluation of the estimated state and delay at each time instant. Moreover, the solution of the optimization problem associated with the MPC design is achieved by solving a series of Quadratic Programming (QP) problem at each time instant. This iterative approach reduces the computational burden compared to the solution of a non-linear optimization problem. A pasteurization plant system is used as a case study to demonstrate the effectiveness of the proposed approach.

This paper proposes an economic health-aware model predictive control (MPC) for Drinking Water Transport Networks (DWN) that includes an additional goal to extend the components and system reliability evaluated using a Bayesian Network. The components and system reliability are incorporated in the MPC model using the LPV approach. The MPC controller uses additionally an economic objective function that determines the optimal filling/emptying sequence of the tanks considering that electricity price varies between day and night. The proposed LPV-MPC control approach allows the controller to accommodate the parameter changes. By computing an estimation of the state variables during prediction, the MPC model can be modified considering the estimated state evolution at each time instant. Moreover, the solution of the optimization problem associated with the MPC problem is achieved by solving a series of Quadratic Programs (QP) at each sampling time. This iterative approach reduces the computational load compared to the solution of a non-linear optimization problem. A part of a real drinking water transport network of Barcelona is used as a case study for illustrating the performance of the proposed approach.

Movement related cortical potential (MRCP), a slow cortical potential from scalp electroencephalogram (EEG), has been the used in real-time brain-computer-interface (BCI) systems designed for neurorehabilitation applications. Detecting MPCPs in real time with high accuracy is essential for a reliable BCI system in these applications. In this study, we investigated the effect of four spatial filters on MRCP detection during executed and imagined dorsiflexion of healthy participants. Our analysis indicated that, in both executed and imagined movement datasets, none of the methods investigated significantly improved trial-to-trial consistency of MRCP. However, all four spatial filters significantly enhanced signal to noise ratio (SNR) of the MRCPs, which is critical to achieve a high true-positive rate in MRCP detection.

Evolutionary Algorithms (EA's) are often well-suited for optimization problems involving several, often conflicting objectives. In this paper an improved hybrid method based on Strength Pareto Evolutionary Algorithm2 (SPEA2) and Tabu Search (TS) is proposed to handle the multi-objective optimization problems (MOPs). This method uses the exploration capacity of SPEA2 beside the power of TS in neighborhood research to find Pareto optimal solutions in different multi objective problems. To have a good distribution of solutions, proposed approach also uses an Improved Diversificator Tabu Search (IDTS) to find unexplored zones of Pareto front and achieve a comprehensive coverage. To test and compare this model with previous works, functions (ZDT1, ZDT2, and ZDT3) are selected. Experimental results show not only clear improvement of the proposed method, in contrast to the previous approaches, but good coverage and distribution of the Pareto front points have been achieved.

Determining the shape, area, volume, and direction of flaws using ultrasonic imaging of metallic pieces, is a method estimating the severity of their defects. Different methods are used to process ultrasound images. Among these methods are spectral analyses, statistical, mathematical and intelligent methods. Within each of these, there are some advantages as well as limitations. Prony algorithm, which has been used as a parametric method for extracting exponential components of a signal, has several applications in signal modeling, system identification and classification. In this paper, after simulating pieces of oil pipeline, digital Wavelet transform has been used to reduce the noise of simulated images. Then the two dimensional Prony algorithm is applied to the images. After estimating the poles and residue of the signal using the Prony method, some features are extracted from these components in order to detect and classify defects. The method proposed in this paper, unlike conventional methods, will not only locate a flaw position in the image, but it can also be used to simultaneously estimate the depth and area of the defect.

By scaling down the feature size and emersion of multi-cores that are usually multi-thread processors, the performance requirements almost guaranteed. Despite the ubiquity of multi-cores, it is as important as ever to deliver high single-thread performance. Multithreaded processors, by simultaneously using both the thread-level parallelism and the instruction-level parallelism of applications, achieve larger instruction per cycle rate than single-thread processors. In the recent multi-core multi-thread systems, the performance and power consumption is severely related to the average memory access time and its power consumption. This makes the cache as a major and important part in designing multi-thread multi-core embedded processor architectures. In this paper we perform a comprehensive design space exploration to find cache sizes that create the best tradeoffs between performance, power, and area of the processor. Finally we run multiple threads on the proposed optimum architecture to find out the maximum thread level parallelism based on performance per power and area efficient uni-thread architecture.

In this paper we have used quantum mechanical transport approach to analyze electrical characteristics of silicon nanowire transistor and have compared the results with those obtained using semi classical Boltzmann transport model. The analyses employs a three dimensional simulation of Silicon nanowire transistor based on self consistent solution of Poisson, Schrodinger equations. Quantum mechanical transport model uses the non equilibrium Green's function (NEGF) [1] while the semi classic model doesn't account for tunneling current. We investigate the effect of tunneling current on I-V characteristics of Nanowire transistor with the different channel length. We have used of NANO TCAD ViDES software to analyze a DG (double gate) silicon nano wire transistor. We get that when the channel length increases to20nm and upper, tunneling is significant only for inversion condition, while for low gate voltages the error between these two models is negligible.

In this paper we investigate the electrical characteristics of silicon nanowire transistors using a fully ballistic quantum mechanical transport approach to analyze rectangular silicon nanowire transistor. We investigate the impact of structural parameters of nano scale Gate all around Silicon nano wire transistor (GAA-SNWT)on its electrical characteristics in subthreshold regime. In particular we show that increase in Source/Drain length (L S , L D ) negligibly affects the current while increasing the L S /L D will affect the gate capacitance. We also investigate the effect of increasing the gate underlap on short channel effects and on the switching speed of device. We show that if the L un is increased the gate capacitance and DIBL will reduces while the I ON /I OFF ratio is increased. This parameter affect the power consumption and delay and is useful in nanowire design for low power application.

As integrated circuits (ICs) become more densely packed with transistors, the conventional MOSFETs approach their physical limits of operation. Thus many novel device structures are being extensively explored. As a substitute Ultra thin Nano scale MOSFETs are one of the potential candidates. It is very attractive one-dimensional material and is useful for future nanoelectronic applications. We have used quantum transport to investigate the effect of scattering due to coupling of the gate contact on the gate tunnelling current. We have developed a proper self energy term which is included within the Hamiltonian formalism for the Poisson Schrodinger solver. It will manifest the coupling of scattering parameter on the gate tunnelling current.