As a specific category of artificial intelligence (AI), generative artificial intelligence (GenAI) generates new content that resembles what is created by humans. The rapid development of GenAI systems has created a huge amount of new data on the Internet, posing new challenges to current computing and communication frameworks. Currently, GenAI services rely on the traditional cloud computing framework due to the need for large computation resources. However, such services will encounter high latency because of data transmission and a high volume of requests. On the other hand, edge-cloud computing can provide adequate computation power and low latency at the same time through the collaboration between edges and the cloud. Thus, it is attractive to build GenAI systems at scale by leveraging the edge-cloud computing paradigm.  In this overview paper, we review recent developments in GenAI and edge-cloud computing, respectively.  Then, we use two exemplary GenAI applications to discuss technical challenges in scaling up their solutions using edge-cloud collaborative systems. Finally, we list design considerations for training and deploying GenAI systems at scale and point out future research directions.Â

The main idea of this research originates from the patients such that each patient with neural disorder should refer to a medical center to check his brain's health condition to sample his EEG signals and present the results to a specialist for further investigation. If this process can be done remotely by tele-medicine techniques, it will save cost and time. In tele-medicine method, the patient can record the EEG signal alone at home and send the results to his physician. To this end, this research employs Bluetooth to connect the interface system to the computer, and the patient can send the results to his physician after saving the data. Thus, the main purpose of designing this BCI system is to record EEG signals using a microcontroller and transmit them via Bluetooth to a computer and mobile phone such that the signal can be represented instantaneously in a GUI.

Focusing on GaN-based light-emitting diode (LED) design optimization, this paper evaluates simulation-based machine learning approaches from a device physics point of view. Strategies are suggested for achieving more realistic results.

We present a novel light field microscope (LFM) hardware design which benefits from scaled productions of cameras and other optical components. Our design has dramatically lower cost (< GBP 2000) yet only slightly compromised performance. This novel LFM design enables transient volumetric imaging at a cellular resolution at a much lower cost.

Intelligent omni-surfaces (IOSs) have been considered for assisting outdoor-to-indoor millimeter-wave (mmWave) communications. Nevertheless, the existing works have not adequately investigated how the number or the deployment locations of IOSs should be optimized for serving multiple indoor users. In this paper, we study IOS-assisted outdoor-to-indoor mmWave communications where IOSs are installed in an exterior wall of a building to refract mmWave signals from an outdoor base station (BS) to indoor users that locate among indoor blockages. Given a fixed total number of refracting elements, we formulate an optimization problem to maximize the downlink energy efficiency of the outdoor BS while satisfying the dowlink data rate requirements of the indoor users by jointly optimizing the number, locations and phase shifts of IOSs and the beamforming vectors of the BS. To address the varying dimensionality and the non-convexity of the optimization problem, we decompose it into two subproblems that optimize the IOSs' phase shifts together with the BS beamforming vectors and the number and locations of IOSs, respectively, and devise successive convex approximation and Continuous Population-Based Incremental Learning-based algorithms to solve them alternately. Simulation results demonstrate that the proposed algorithms can obtain the optimal number and locations of IOSs, resulting in significantly enhanced energy efficiency of the outdoor BS compared to benchmark schemes.

Differences in contact impedance of the ECG measurement electrodes lead to asymmetries of the signal paths and thus result in reduced common-mode rejection and artifacts. Here, the imbalance of contact impedance is investigated for different types of electrodes with capacitive coupling in terms of static imbalance as well as dynamic variation during body movement. Flexible and incompressible materials like conductive foam and fabric showed the best overall performance. The negative effect of rigidity can partly be compensated by adding conducting foam, while soft materials can profit from an increase of electrode area.

This is a pre-print for the submission to IEEE LISS 2020 Conference, Budapest, Hungary.

A document by Mohaimenul Azam Khan Raiaan. Click on the document to view its contents.

Pilot spoofing attack brings new challenges to physical layer secure transmission. However, this method will not work without any knowledge about the pilot sequence and active eavesdropping can be detected by constructing random pilot sequence. Intelligent reflecting surface (IRS), with the real-time programmable features for wireless channels, provides new possibilities for effective pilot spoofing. In this paper, the IRS is deployed near the legitimate users and the legitimate signal can always be passively reflected. Then the control strategy is embedded into the communication process under time-division duplex mode to assist eavesdroppers to conduct pilot spoofing. By setting different phase shifts at the IRS during the uplink phase and downlink phase, the channel reciprocity disappears, and thus secure beamforming vector is biased towards the eavesdropper. Furthermore, in order to overhear more information, the average secrecy rate minimization problem based on statistical channel state information is established by carefully designing the phase shifts, which is non-trivial to solve. With alternating optimization and Charnes-Cooper transformation technique, the original problem is transformed into convex form and a near optimal solution is achieved. Finally, simulation results show that our proposed scheme can pose serious secure threat without any energy footprint. What’s more, if the IRS is not utilized by the internal users properly, it will bring more threat.

This is the second part of the three part series paper where the novel idea of representing flyback transformer losses as loss functions using hyperbolic secant spline interpolation, and its use in determining flyback transformer core saturation and instability in the control feedback loop from the loss function using Lyapunov method has been presented with scientific proof and validity. In this paper, the comparative analysis of 36 different flyback transformers from 2W to 27W is being presented. These two-winding transformers are operated in continuous/discontinuous mode, switched from MOSFET controlled by microcontroller from 41 kHz to 97 kHz. Eighteen wattage’s which are used internationally at different voltage and current ratings have been analyzed on fifteen different cores - EPC13, EF12.6, EFD15, EE13, EE16, EEL16,EE19, EEL19,EFD20,EE25, E20/10/6, PR18x11,EEL22,RM6S/I, RM8/I.

Optimization problems for trajectory planning in autonomous vehicle racing are characterized by their nonlinearity and nonconvexity. Instead of solving these optimization problems, usually a convex approximation is solved instead to achieve a high update rate. We present a real-time-capable model predictive control (MPC) trajectory planner based on a nonlinear single-track vehicle model and Pacejka’s magic tire formula for autonomous vehicle racing. After formulating the general nonconvex trajectory optimization problem, we form a convex approximation using sequential convex programming (SCP). The state of the art convexifies track constraints using sequential linearization (SL), which is a method of relaxing the constraints. Solutions to the relaxed optimization problem are not guaranteed to be feasible in the nonconvex optimization problem. We propose sequential convex restriction (SCR) as a method to convexify track constraints. SCR guarantees that resulting solutions are feasible in the nonconvex optimization problem. We show recursive feasibility of solutions to the restricted optimization problem. The MPC is evaluated on a scaled version of the Hockenheimring racing track in simulation. The results show that an MPC using SCR yields faster lap times than an MPC using SL, while still being real-time capable.

Backscatter communication (BackCom) is a promising technique for achieving high spectrum efficiency and power efficiency in the future Internet of Things systems. The capacity of BackCom networks can be maximized by optimizing the backscatter time and the reflection coefficient (RC). However, system energy efficiency (EE) cannot be guaranteed usually. In this paper, we investigate the energy-efficient resource allocation problem of a non-orthogonal multiple access (NOMA)-based BackCom. Particularly, the base station (BS) transmits signals to two cellular users based on the NOMA protocol, meanwhile, a backscatter device backscatters the signals to users using the passive radio technology. The total EE of the considered system is maximized by jointly optimizing power allocation for each NOMA user and the RC of backscatter device where the decoding order and the quality of service (QoS) of each user are guaranteed. To solve such a non-convex problem, we develop an efficient iterative algorithm to obtain the optimal solutions by using Dinkelbach’s method and the quadratic transformation approach. Numerical results show that the proposed algorithm can significantly improve the system EE compared with the orthogonal multiple access (OMA) scheme and the NOMA system without backscatter devices.

This paper presents a novel spatiotemporal optimization approach for maximizing the output power of an ocean current turbine (OCT) under uncertain ocean velocities. In order to determine output power, ocean velocities and the power consumed and generated by an OCT system are modeled. The stochastic behavior of ocean velocities is a function of time and location, which is modeled as a Gaussian process. The power of the OCT system is composed of three parts, including generated power, power for maintaining the system at an operating depth, and power consumed for changing the water depth to reach the maximum power. Two different algorithms, including model predictive control (MPC) as a model-based method and reinforcement learning (RL) as a learning-based method, are proposed to design the optimization structure, and comparative studies are presented. On one hand, the MPC based controller is faster in finding the optimal water depth, while the RL is also computationally feasible considering the required time for changing operating depth. On the other hand, the cumulative energy production of the RL algorithm is higher than the MPC method, which verifies that the learning-based RL algorithm can provide a better solution to address the uncertainties in renewable energy systems. Results verify the efficiency of both presented methods in maximizing the total power of an OCT system, where the total harnessed energy after 200 hours shows an over 18% increase compared to the baseline.

The given literature focuses on developing a Smart Factory model based on Cloud and Edge computing used to develop Transportation Management System(TMS) using a iFogSim wrapper. Cloud computing identifies data centres for users and offer computer system services on-demand, including data storage and processing power, without direct active user management. In the smart industry, several devices are connected together across the internet, where vast volumes of data are collected during the entire process of output. Thus, to handle this data smart factory based on cloud and edge computing is used. The intelligent cloud-based factory offers some facility like large scale analysis of data. Concepts like fog and edge computing play a significant role in extending data storage and network capacities in the cloud that addresses some challenges, such as over-full bandwidth and latency. The literature also focuses on the implementation of TMS using the iFogSim Simulator. The simulator provides efficient execution of TMS by showing the amount of resources used which gives an idea regarding optimum use of resources. All types of data related to TMS is obtained at cloud by using smart factory like object location, time taken and energy consumption. To implement the TMS we have created a topology which displays various devices connected to the cloud which gives necessary information regarding the ongoing transportation simulation.

This paper applies a Spatiotemporal Graph Convolutional Recurrent Neural Network which is a tight combination of a Graph Neural Network (GNN) to a Recurrent Neural Network (RNN) architecture for air pollution forecasting in long-term for the entire city. Our model can effectively learn the spatial and temporal features of the air pollution data and its influential factors (e.g. weather, traffic, external areas) at the time. Our method achieves better performance than a state-of-the-art ConvLSTM model in air pollution forecasting and a hybrid GNN-based model that separates GNN and RNN in discrete layers.

The COVID-19 crisis is profoundly influencing the global economic framework due to restrictive measures adopted by governments worldwide. Finding real-time data to correctly quantify this impact is very significant but not as straightforward. Nevertheless, an analysis of the power demand profiles provides insight into the overall economic trends. To accurately assess the change in energy consumption patterns, in this work we employ a multi-layer feed-forward neural network that calculates an estimation of the aggregated power demand in the north of Italy, (i.e, in one of the European areas that were most affected by the pandemics) in the absence of the COVID-19 emergency. After assessing the forecasting model reliability, we compare the estimation with the ground truth data to quantify the variation in power consumption. Moreover, we correlate this variation with the change in mobility behaviors during the lockdown period by employing the Google mobility report data. From this unexpected and unprecedented situation, we obtain some intuition regarding the power system macro-structure and its relation with the overall people’s mobility. Postprint accepted for publication in the proceedings of the 2020 AEIT International Annual Conference (AEIT). How to cite: P. Scarabaggio, M. La Scala, R. Carli and M. Dotoli, ”Analyzing the Effects of COVID-19 Pandemic on the Energy Demand: the Case of Northern Italy,” 2020 AEIT International Annual Conference (AEIT), 2020. DOI: https://doi.org/10.23919/AEIT50178.2020.9241136  © 2020 AEIT 978-8-8872-3747-4. Personal use of this material is permitted. Permission 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.

Quality assurance (QA) is an important task in manufacturing to assess whether products meet their specifications. However, QA might be expensive, time-consuming, or incomplete. This paper presents a solution for predictive analytics in QA based on machine sensor values during production while employing specialized machine-learning models for classification in a controlled environment. Furthermore, we present lessons learned while implementing this model, which helps to reduce complexity in further industrial applications. The paper’s outcome proves that the developed model was able to predict product quality, as well as to identify the correlation between machine-status and faulty product occurrence.

Finely tuned deep generative adversarial networks (GANs) can generate high-quality (HQ) images. However, the discriminator in GAN is only able to distinguish true or fake images. Moreover, numerous synthesized images from GANs are imperfect, and we can not reconstruct those images via GANs. In this paper, we revisit pre-trained GANs and offer a self-supervised method to quickly transform GAN’s discriminators into encoders. We reuse parameters of the GAN’s discriminator and replace its output layer, so that it can be transformed into an encoder and output reformed latent vectors. The transformation makes the pre-trained GAN a symmetrical architecture and allows for better performance. Based on the method, GANs can be made to reconstruct synthesized images via encoders. Compared to synthesized images, these reconstructions can maintain or even attain higher quality.