مجله مهندسی برق و الکترونیک ایران
سال بیستم شماره 1 (بهار 1402)

Microelectromechanical thermal sensors are one of the most accurate and important tools for measuring the direction and velocity of an acoustic wave and winds. Detection of wind direction and speed in different ranges has different applications such as meteorology, wind power plants, gas flow measurement in smart building and gas consumption of power plants. In this paper, a four wires sensor is designed and built to measure the velocity and direction of a fluid simultaneously. Comsol Multi-physics software has been used for simulation and design. To increase the sensitivity of the sensor, the all of sensor elements are fabricated on a silicon membrane to reduce the heat transfer of the structure with the sensor body. To detect the direction, a four-wire and orthogonal structure has been used, which provides the ability to measure the direction up to 5 degrees. A test set-up was designed and built to characterize this sensor, and the sensor was tested up to speeds of 5 meters per second, which is the range of industrial flow-meters, wind and gasometer of smart building. By changing the distance of the wires, this sensor is able to measure speeds of 30m/s.

Memory resistor or memristor is already fabricated successfully using current nano dimension technology. Based on its unique hysteresis, the amount of resistance remains constant over time, controlled by the time, the amplitude, and the polarity of the applied voltage. The unique hysteretic current-voltage characteristic in the memristor causes this element to act as a non-volatile resistive memory, to store the information till the next perturbation. In this paper, a bridge consisting of four memristors is introduced as a synapse to connect the two neuron cells, leaky integrate-and-fire, and Hodgkin-Huxley neuron cells. The use of memristor bridge synapse in the proposed architecture solves one of the major problems, regarding nonvolatile weight storage in analog neural network implementations. By changing different values of each memristor, the synaptic weight will be programmable. The current-voltage characteristics and their weight variations will be examined in the behavior of neuron spikes.

Today, electrical stimulation as a treatment method by creating muscle contraction by creating internal electrical stimulation is a method to reduce muscle pain. In addition, it is used to treat osteoarthritis. The aim of this study was to study, design and implement optimized hardware for electrical muscle stimulators to reduce pain and treat osteoarthritis. Here, the design of the optimal system with much lower implementation costs than the existing work is discussed. Among the various methods for making this device, the cost of these devices is very high and it is not possible to provide it by patients. Here, this problem can be solved by building a low-cost and optimal circuit that has the necessary properties for electrical stimulation of muscles. In this structure, using two potentiometers, the power and frequency of pulses applied to the body can be controlled in accordance with various applications from muscle stimulation to the treatment of osteoarthritis. The use of transistors with proper flow capability and safe device is suitable for tasks that require the application of stronger pulses. According to the proposed structure, the frequency of stimuli is from the range of Hz to kHz, which can be considered for a wider range of treatments in this basin. Thus, the application of signals with lower frequency and higher current causes a greater depth of penetration in the body and is suitable for the treatment of osteoarthritis and further stimulation of damaged cartilage. The proposed circuit has been tested on different people and the desired result has been received from the experiments.

In this paper, a new method for evaluating CMOS image sensors based on computer modeling and analysis is introduced. Image sensors are composed of different parts, each of which has a specific effect on image quality. Circuits of image sensors can be evaluated and analyzed using circuit simulators or theoretically, but these methods cannot help to produce the final image. In order to produce the image and evaluate it, it is necessary that the image sensor chip is made and the image produced by the chip is practically tested in the laboratory. Using the proposed method, it is possible to model any block of the sensor and examine the effect of that block on the image without making a chip. In this paper, the effect of source-follower as a pixel readout circuit on the final image has been investigated. The simulation results showed that the source-follower circuit with a 20% attenuation and a 1.15% non-linearity effect on the pixel signal can cause a significant error in the generated image. The related circuits are designed using 0.18µm CMOS technology and simulated by CADENCE software with SPECTRE simulator and also MATLAB software is used to model it.

This paper proposes a new structure for digital address decoders based on flip-flops with application in wake-up signal generators of wireless networks nodes. Such nodes equipped with this device can be utilized in Internet of Things applications where the nodes are dependent on environment energy harvesting to survive for a long time. Different parts in these wireless nodes should have an electrical circuit design with low energy consumption and our proposed address decoder used in the wake-up signal receiver can improve such systems. The proposed structure as a digital address decoder uses power much less than conventional digital address decoders. We showed that our device consumes power even more efficiently in systems with longer address lengths and higher input data rates. To reduce power consumption, the proposed structure uses a sequential architecture, comparator blocks, activation units and sleep mode design for flip-flops. Flip-flops' propagation delay is used to calculate the circuit total response delay which was conversely related to the power consumption of the design. Static power of the device was limited by determining the essential active MOSFETs of the whole design and putting the others in the sleep mode. Simulation of the both proposed and conventional structures were carried out in a 32nm Cmos library in Hspice and showed that for a structure with an address length of 64 bits and 100 kbps input data rate, we will have more than 90% reduction in power consumption compared to the state-of-the-art structures. Comparing the results with similar tasks and adding a telecommunication receiver segment promises a 50% reduction in the total power consumption of the wake-up receiver of a complete IOT device.

Cloud computing gives a large quantity of processing possibilities and heterogeneous resources, meeting the prerequisites of numerous applications at diverse levels. Therefore, resource allocation is vital in cloud computing. Resource allocation is a technique that resources such as CPU, RAM, and disk in cloud data centers are divided among cloud users. The resource utilization, cloud service provider profit, and user satisfaction are the common objectives of the proposed algorithms. In this study, the algorithm is based on fuzzy logic and tries to achieve better results than other resource allocation algorithms by using meta-heuristic methods. After designing the preliminary fuzzy inference system (FIS), the use of meta-heuristic algorithms is ended up tuning the FIS parameters. By adjusting the parameters, membership functions are improved and finally a trained FIS is delivered. PSO has been used to train the fuzzy system. The fuzzy resource allocation algorithm responded 95% of the requests and achieved 61.61% of the resource efficiency, while the fuzzy-PSO algorithm answered 96% of the requests and improved the resource utilization by 0.5%. The results have shown that the application of PSO improves fuzzy resource allocation efficiency. More requests are answered and it increases resource utilization.

There are many obstacles in quantum circuits implementation with large scales, so distributed quantum systems are appropriate solution for these quantum circuits. Therefore, reducing the number of quantum teleportation leads to improve the cost of implementing a quantum circuit. The minimum number of teleportations can be considered as a measure of the efficiency of distributed quantum systems. Due to the complexity of reducing teleportation costs in multi-partition quantum computing, a two-level optimization is proposed. In the first level, using graph partition methods, the quantum circuit is divided into k partitions. Secondly, using a series of proposed rules and functions, the number of teleportations for the partitioned quantum circuit is improved. In previous work, only optimization at the first level has been conisdered for the distributed quantum circuits. The implementation results show that the cost of teleportation in the proposed algorithm has been improved compared to previous work on benchmark circuits.

The goal of recommender system is to provide desired items for users. One of the main challenges affecting the performance of recommendation systems is the cold-start problem that is occurred as a result of lack of information about a user/item. In this article, first we will present an approach, uses social streams such as Twitter to create a behavioral profile, then user profiles are clustering with machine learning techniques. Based on this clustering, predictions are made using machine learning techniques such as the Random Forest algorithm (RF) and the Gradient Boosting algorithm (GB). Therefore, the user is not required to provide any kind of data explicitly anymore. As a result of this method, cold start problem will decrease among users' social networks. Because the system uses this data to create user profiles, this will be an input for recommender systems. Numerous experiments have been performed in this field and compared to some new cold start algorithms; very satisfactory results have been obtained. In this paper, we have concluded that the clustering process greatly increases the performance accuracy of the models and reduces the average absolute error, and also the Gradient Boosting algorithm has a better performance than the Random Forest algorithm.

The issue of classification is still a topic of discussion in many current articles. Most of the models presented in the articles suffer from a lack of explanation for a reason comprehensible to humans. One way to create explainability is to separate the weights of the network into positive and negative parts based on the prototype. The positive part represents the weights of the correct class and the negative part represents the weights that are incorrectly assigned to that class. This network is called the winner-takes-all network based on the positive and negative Euclidean distance or ± ED-WTA. In this paper, using the kernel expansions and achieving local explainability, higher accuracy has been achieved in this field through nonlinear modeling. Methods in this paper will be presented to improve the temporal and spatial space of the algorithm. The article also uses the Nystrom method to approximate kernel to scale the algorithm against large datasets. Using this single-layer network and the Gaussian kernel function, 98.01% accuracy is obtained on the test data on the MNIST dataset, and it also explains the reasons of decision well with the input data using kernel expansions. Explainability has also been investigated on two classes FERET dataset.

One of the most important concerns of a data miner is always to have accurate and error-free data. Data that does not contain human errors and whose records are full and contain correct data. In this paper, a new learning model based on an extreme learning machine neural network is proposed for outlier detection. The function of neural networks depends on various parameters such as the structure of them, initial weights, number of hidden layer neurons, and learning rate. Quantum computing is a new method of information processing based on quantum mechanics, the concepts of which are also used today in applications of artificial intelligence. In the proposed method, the neural network of the extreme learning machine is improved using the concept of the quantum fuzzy clustering c-Means. This clustering helps to find the optimal weight of the input layer connections to the hidden layer of the neural network. It also allows network architecture to be constructively constructed in the hidden layer and improves learning. The performance of the proposed method in terms of accuracy, correct positive rate and correct negative rate shows the superiority of the proposed method in detecting outlier data compared to other methods.

Matrix completion problem has gathered a lot of attention in recent years. In the matrix completion problem, the goal is to recover a low-rank matrix from a subset of its entries. The graph matrix completion was introduced based on the fact that the relation between rows (or columns) of a matrix can be modeled as a graph structure. The graph matrix completion problem is formulated by adding the graph total variation term to the objective function of matrix completion problem. However; in practice, the observed data is noisy and contains outliers. Outlier data is defined as part of the observed that are different than other parts and are not consistent with the data structure. In this paper, we apply graph total variation based on the directed Laplacian and propose a new method for graph matrix completion. We introduce a new method called GMCO-DL for the case where both noise and outliers exist in observations. Simulation results show outstanding results for the proposed method in terms of estimation error.

In this paper, a Fredkin gate using photonic-crystal around the 1550 nm telecommunication wavelength is modelled. The proposed structure has a compact footprint compared with previous works of research groups. The proposed structure is based on air holes on silicon substrate. Silicon substrate and air holes have 466 nm and 177 nm radius, respectively. For analyse of the structure Finite-Difference-Time-Domain (FDTD) is applied. To access of nonlinear effects, a glass ring with refractive index of 1.4 and 10-14m2/w is used. To benchmark of the model, response time and contrast ratio is investigated. Numerical results show, in the best case, raise time of 0.21 ps, fall time of 0.02 ps as well as 4.31 contrast ratio can be obtained.

In this paper, a Dual-Bit Parallel Adder (DBPA) based on minority function using Carbon-Nanotube Field-Effect Transistor (CNFET) is proposed. The possibility of having several threshold voltage (Vt) levels by CNFETs leading to wide use of them in designing of digital circuits. The main goal of designing proposed DBPA is to reduce critical path delay in adder circuits. The proposed design positively effects the speed and power consumption parameters through data path shortening.  In order to evaluate the proposed design, several simulations are performed using Synopsys HSPICE in 32nmCMOS and 32nm CNFET technologies. Proposed DBPA circuit with five other two-bit adder circuits implemented using five different full adder cells in power consumption, delay and Power Delay Product (PDP) parameters has been compared. To evaluate the performance of different designs in larger circuits, 4-bit Ripple Carry Adder (RCA) and 8-bit RCA have been simulated. Based on obtained results, the proposed design is faster than other designs due to the data path shortening. The results of the simulations confirm the higher efficiency of the proposed design with respect to other designs.

In energy supply applications for low-power sensors, there are cases where energy should be transmitted from a low-power battery to an output stage load capacitor. This paper presents an adiabatic charging circuit with a parallel switches approach that connects to a low-power battery and charges the load capacitor using a buck converter which operates in continuous conduction mode (CCM). A gate controller of parallel switches (GCPS) is used to increase the duty-cycle of the input signal of buck converter switches, which controls the inductor current and charges the load capacitor in 256 steps. The proposed parallel switches approach is used to improve the energy efficiency under different powers transmission and comprises a current sensor, comparator, R-S latch, parallel pMOS switches, and parallel nMOS switches. For high powers transfer, the large switches are paralleled with small switches and result in conduction losses reduction. Also, for low powers transfer, the total switching losses are significantly reduced by turning off the large PMOS/NMOS switches during the charging operation. The proposed circuit was designed and simulated in a 0.18μm CMOS technology. The efficiency of the proposed circuit during the capacitor charge-up cycle is more than 80% for average input powers between 0.5mW to 30mW.

Nowadays, LCL filters utilization to suppress the grid-connected inverter switching harmonics is expanding and various methods such as capacitor current active damping scheme could be investigated to damp out these harmonics. However, the computational delays including control system, and PWM switching delays have significant effect on damping system stability. In most of LCL filter parameter designing schemes, the try and error methods are used, and the complicated calculations are needed for system stability considering grid impedance variation and delay effect. In this paper, at first a systematic method is provided for filter parameters designing by which better dynamic response and lower filter volume size are resulted and then, the step-by-step design procedure of grid-connected converter controller parameters with LCL filter by reducing the computational delay in inner (capacitor current feedback)- and outer control loop (grid current) is proposed for LCL filter resonance damping which is stable in case of severe variations in grid impedance. The simulation results by MATLAB/Simulink illustrate appropriate system operation.

In the training phase of learning algorithms, it is always important to have a suitable training data set. The presence of outliers, noise data, and inappropriate data always affects the performance of existing algorithms. The active learning method (ALM) is one of the powerful tools in soft computing inspired by the computation of the human brain. The operation of this algorithm is completely based on simple calculations and tries to model a Multi-Input Single-Output (MISO) system as a set of Single-Input Single-Output (SISO) subsystems and breaks a complex problem into several simpler problems. Each of the subsystems is then modeled by an Ink Drop Spread (IDS) plane. In this paper, to improve the performance of ALM, with changes in how it works, data called negative data has been used. In the ALM, it is possible to use negative data in the training phase by using the IDS operator. By doing so, a policy similar to the reward and punishment policy is in reinforcement learning methods is used. To investigate the accuracy of the proposed algorithm, some simulations in control have been done on an inverted pendulum system and the FVU value is 0.0143. The simulations results confirm the proper performance and increase the computational power of the proposed method compared to the existing method.

Multi-agent systems are systems in which several agents accomplish a mission in a cooperative manner. In this paper, a novel idea for the construction of a movable runway platform based on multi-agent systems is presented. It is assumed that an aerial agent (quadrotor) decides to make an emergency landing due to reasons such as a decrease in energy level or technical failure, while there is no suitable platform for landing or the agent is not able to reach another place for landing. The goal is to create a specific formation pattern by a group of mobile robots that are smaller than the aerial agent, so that it can land on them. Each of the mobile robots is equipped with flat grid landing pad in its upper part so that when the robots approach each other and form a specific geometric arrangement (such as a pentagon), the landing pads are connected to each other and create a wider platform for landing. Cooperative control of agents (from aerial agent control to mobile robot formation control) is the main goal of this article, which is presented in different dynamic models under directed graph along with proof of Lyapunov-based stability analysis.

Aerodynamic load simulators generate the required time varying load to test the actuator’s performance in the laboratory. Electric Load Simulator (ELS) as one of variety of the dynamic load simulators should follows the rotation of the Under Test Actuator (UTA) and applies the desired torque to UTA’s rotor at the same time. In such a situation, a very large torque is imposed to the ELS from the UTA. Recently, the Permanent Magnet Synchronous Motor (PMSM) has been used for the ELS. In practice, the composition of PMSM and its drive is an unknown nonlinear plant. Model-based ELS design methodologies use a simplified model for PMSM and its drive. In this paper, a model-free method is used for ELS controller design. Two applicable and novel experiments are set up for identifying the ELS and disturbance models. Then, the identified models are used for a robust controller design based on quantitative feedback theory. The proposed method uses the ELS as a black-box and eliminates the large disturbance only based on the estimated model of the unknown nonlinear model. The control effort limitation is considered in the design stage and simultaneously a large bandwidth is achieved for the non-minimum phase system. The robust performance of the controller is evaluated on the ELS coupled to UTA.