As biometric systems become widely deployed, the security of identification systems can be easily attacked by various spoof materials. This paper contributes to finding a reliable and practical anti-spoofing method using Convolutional Neural Networks (CNNs) based on the types of loss functions and optimizers. The types of CNNs used in this paper include AlexNet, VGGNet, and ResNet. By using various loss functions including Cross-Entropy, Center Loss, Cosine Proximity, and Hinge Loss, and various loss optimizers which include Adam, SGD, RMSProp, Adadelta, Adagrad, and Nadam, we obtained significant performance changes. We realize that choosing the correct loss function for each model is crucial since different loss functions lead to different errors on the same evaluation. By using a subset of the Livdet 2017 database, we validate our approach to compare the generalization power. It is important to note that we use a subset of LiveDet and the database is the same across all training and testing for each model. This way, we can compare the performance, in terms of generalization, for the unseen data across all different models. The best CNN (AlexNet) with the appropriate loss function and optimizers result in more than 3% of performance gain over the other CNN models with the default loss function and optimizer. In addition to the highest generalization performance, this paper also contains the models with high accuracy associated with parameters and mean average error rates to find the model that consumes the least memory and computation time for training and testing. Although AlexNet has less complexity over other CNN models, it is proven to be very efficient. For practical anti-spoofing systems, the deployed version should use a small amount of memory and should run very fast with high anti-spoofing performance. For our deployed version on smartphones, additional processing steps, such as quantization and pruning algorithms, have been applied in our final model.
The most common quadrature amplitude modulator (QAM) applies two Mach-Zehnder Modulators (MZM) and one phase shifter to generate high order modulation format. The bias of MZM changes over time due to temperature, vibration, and aging factors. The change in the biasing causes distortion to the generated QAM signal which leads to deterioration of bit error rate (BER) performance. Therefore, it is critical to be able to lock MZM’s Q point to the required operating point for good performance. We propose a technique for automatic bias control (ABC) of QAM transmitter using BER measurements and gradient descent optimization algorithm. The proposed technique is attractive because it uses the pertinent metric, BER, which compensates for bias drifting independently from other system variations such as laser source output power. The proposed scheme performance and its operating principles are simulated using OptiSystem simulation software for 4-QAM and 16-QAM transmitters.
The toxicity of lead associated with the lifecycle of perovskite solar cells (PSCs( is a serious concern which may prove to be a major hurdle in the path toward their commercialization. The current proposed lead-free PSCs including Ag(I), Bi(III), Sb(III), Ti(IV), Ge(II), and Sn(II) low-toxicity cations are still plagued with the critical issues of poor stability and low efficiency. This is mainly because of their chemical stability. In the present research, utilization of all inorganic CsSnGeI3 based materials offers the advantages to enhance resistance of device to degradation, reduce the cost of cells, and minimize the carrier recombination. The presence of inorganic halide perovskite improves the photovoltaic parameters of PCSs via improved surface coverage and stability. The inverted structure of simulated devices using a 1D simulator like solar cell capacitance simulator (SCAPS) version 3308 involves TCOHTL/Perovskite/ETL/Au contact layer. PEDOT:PSS, PCBM, and CsSnGeI3 used as hole transporting layer (HTL), electron transporting layer (ETL), and perovskite absorber layer in the inverted structure for the first time. The holes are injected from highly stable and air tolerant Sn0.5Ge0.5I3 perovskite composition to HTM and electrons from the perovskite to ETL. Simulation results revealed a great dependence of power conversion efficiency (PCE) on the thickness and defect density of perovskite layer. Here the effect of an increase in operating temperature from 300 K to 400 K on the performance of CsSnGeI3 based perovskite devices is investigated. Comparison between simulated CsSnGeI3 based PCSs and similar real testified devices with spiro-OMeTAD as HTL showed that the extraction of carriers at the interfaces of perovskite absorber depends on the energy level mismatches between perovskite and HTL/ETL. We believe that optimization results reported here represent a critical avenue for fabricating the stable, low-cost, efficient, and eco-friendly all-inorganic Cs-Sn-Ge based lead-free perovskite devices.
It is necessary to test to see if optical devices such as camera, night vision devices are working properly. Therefore, a precision biaxial rotary system (gimbal) is required for mounting Unit Under Test, UUT. The Gimbal systems can be utilized for precise positioning of the UUT; hence, optical test can be performed with high accuracy. The weight of UUT, which is placed outside the axis of rotation, causes an off-axis moment to the mounting armature. The off-axis moment can act against the direction of movement for some orientation, thus the electrical motor, which rotates the gimbal axis, has to apply higher level of torque to guide and stabilize the system. Moreover, UUT and its mounting fixture to the gimbal can be changed, which causes change in applied resistance moment to the gimbals electrical motor. In this study, a preloaded spring is added to the gimbal system for minimizing applied off axis moment with the help of four bar mechanism. Two different possible methods for preloading spring are introduced and system optimization is performed to eliminate all moment which is created by off axis weight.
To improve the registration accuracy of a source point cloud and template point cloud when the initial relative deflection angle is too large, a PointNetLK algorithm combined with an oriented bounding box (PointNetLK-OBB) is proposed. In this algorithm, the OBB of a 3D point cloud is used to represent the macro feature of source and template point clouds. Under the guidance of the iterative closest point algorithm, the OBB of the source and template point clouds is aligned, and a mirror symmetry effect is produced between them. According to the fitting degree of the source and template point clouds, the mirror symmetry plane is detected, and the optimal rotation and translation of the source point cloud is obtained to complete the 3D point cloud registration task. To verify the effectiveness of the proposed algorithm, a comparative experiment was performed using the publicly available ModelNet40 dataset. The experimental results demonstrate that, compared with PointNetLK, PointNetLK-OBB improves the registration accuracy of the source and template point clouds when the initial relative deflection angle is too large, and the sensitivity of the initial relative position between the source point cloud and template point cloud is reduced. The primary contribution of this paper is the use of PointNetLK to avoid the non-convex problem of traditional point cloud registration and leveraging the regularity of the OBB to avoid the local optimization problem in the PointNetLK context.
In this study, the commercial Computational Fluid Dynamics (CFD), ANSYS-Fluent, has been used to optimize the marine propeller with the design of experiment (DOE) method. At the initial stage, different propeller parameters ware selected for the three different levels. The four characteristics factors are: no. of the blade, camber value, pitch delta & chord at the hub. Then, CAD modelling is performed by considering the selected factor and level. In this investigation, a total of 9 test models are simulated with the Reynolds-Averaged Navier-Stokes (RANS) equations. The standard, realizable
The time coordination of overcurrent relays (OCR) in a power distribution network is of great importance, as it reduces the power outages by avoiding the mal-operation of the backup relays. For this, the optimum value of the time multiplier setting (TMS) of OCRs should be chosen. The problem of determining the optimum value of TMS of OCRs in power distribution networks is formulated as a constrained optimization problem. The objective is to find the optimum value of TMS of OCRs to minimize the time of operation of relays under the constraint of maintaining the coordination of relays. A power distribution network can have a combination of numerical and electromechanical relays. The TMS of numerical relays can be set to any real value (which satisfies the constraints of the problem), whereas the TMS of electromechanical relays can be set in fixed step (0 to 1 in steps of 0.05). The main contribution of this paper is a formulation of the problem as a mixed-integer linear programming (MILP) problem and application of Gomory's cutting plane method to find the optimum value of TMS of OCRs. The TMS of electromechanical relays are taken as integers in the range 1 to 20 in the step of 1, and these values are mapped to 0.05 to 1 in the step of 0.05. The results obtained are compared with those obtained using a simplex method and its variants. It has been shown that the mixed-integer linear programming method outperforms the simplex method (and its variants) in the case of a system having a combination of numerical and electromechanical relays.
To ensure reliability in miniaturized devices or processes with increased heat fluxes, very efficient cooling methods have to be employed in order to cope with small available cooling surfaces. To address this problem, a certain type of evaporator/heat exchanger was developed: It is called a swirl evaporator due to its flow characteristic. The swirl evaporator consists of a concentrically eroded screw geometry in which a capillary tube is guided, which is inserted into a pocket hole in components with high heat load. The liquid refrigerant R32 is sprayed through the capillary tube to the end face of the blind hole and is sucked off against the injection direction in the screw geometry. Its inner diameter is between one and three millimeters. The refrigerant is sprayed into the pocket hole via a small tube aligned in the center of the bore hole and is sucked off on the front side of the hole against the direction of injection. The refrigerant is sucked off in a helical geometry (twisted flow) so that it is accelerated against the hot wall (centrifugal acceleration). This results in an increase in the critical heat flux of up to 40%. In this way, more heat can be dissipated on the same surface/available installation space. This enables a wide range of technical applications. To optimize the design for the needs in various fields of industry, like the internal tool cooling when machining nickel base alloys like Inconel 718, a correlation-based model of the swirl-evaporator was developed. The model is separated into 3 subgroups with overall 5 regimes. The pressure drop and heat transfer are calculated separately. An approach to determine the locality of phase change in the capillary and the swirl was implemented. A test stand has been developed to verify the simulation.
The direction of arrival (DoA) estimation is the crucial aspect of the radar technologies for detecting and dividing several signal sources. In this scenario, the antenna array output modeling involves numerous parameters including noise samples, signal waveform, signal directions, signal number, and signal to noise ratio (SNR), and thereby the methods of the DoA estimation rely heavily on the generalization characteristic for establishing a large number of the training data sets. Hence, we have analogously represented the two different optimization models of the DoA estimation; (1) the implementation of the decision directed acyclic graph (DDAG) for the multiclass least-squares support vector machine (LS-SVM), and (2) the optimization method of the deep neural network (DNN) radial basis function (RBF). We have rigorously verified that the LS-SVM DDAG algorithm is capable of accurately classifying DoAs for the three classes. However, the accuracy and robustness of the DoA estimation are still highly sensitive to technological imperfections of the antenna arrays such as non-ideal array design and manufacture, array implementation, mutual coupling effect, and background radiation and thereby the method may fail in representing high precision for the DoA estimation. Therefore, this work has a further contribution on developing the DNN-RBF model for the DoA estimation for overcoming the limitations of the non-parametric and data-driven methods in terms of array imperfection and generalization. The numerical results of implementing the DNN-RBF model have confirmed the better performance of the DoA estimation compared with the LS-SVM algorithm. Consequently, we have analogously evaluated the performance of utilizing the two aforementioned optimization methods for the DoA estimation using the concept of the mean squared error (MSE).
Maneuver decision-making plays a critical role in high-performance intelligent driving. This paper proposes a risk assessment-based decision-making network (RADMN) to address the problem of driving strategy for the commercial vehicle. RADMN integrates two networks, aiming at identifying the risk degree of collision and rollover and providing decisions to ensure the effectiveness and reliability of driving strategy. In the risk assessment module, risk degrees of the backward collision, forward collision and rollover are quantified for hazard recognition. In the decision module, a deep reinforcement learning based on multi-objective optimization (DRL-MOO) algorithm is designed, which comprehensively considers the risk degree and motion states of each traffic participant. To evaluate the performance of the proposed framework, Prescan/Simulink joint simulation was conducted in highway scenarios. Experimental results validate the effectiveness and reliability of the proposed RADMN. The output driving strategy can guarantee the safety and provide key technical support for the realization of autonomous driving of commercial vehicles.
Estimating the 6D pose of objects is a core step for robot bin-picking tasks. The problem is that various objects are usually randomly stacked with heavy occlusion in real applications. In this work, we propose a method to regress 6D poses by predicting three points for each object in the 3D point cloud through deep learning. To solve the ambiguity of symmetric pose, we propose a labeling method to help the network converge better. Based on the predicted pose, an iterative method is employed for pose optimization. In real-world experiments, our method outperforms the classical approach in both precision and recall.
It is known that using Coanda aerosurfaces can drastically augment the lift forces when applied to an Unmanned Aerial Vehicle (UAV) platform. However, Coanda saucer UAVs, which commonly use a dish-like, radially-extending structure, have shown no significant increases in thrust/lift force and therefore have never been commercially successful: the additional thrust/lift generated by the Coanda surface diminishes since the airstreams emerging from the rotor compartment expand radially causing serious loss of momentums and therefore a net loss of total thrust/lift. To overcome this technical weakness, we propose to examine a Coanda surface of straight, cylindrical design and optimize its geometry for highest thrust/lift utilizing computational fluid dynamics software ANSYS Fluent®. The results of this study reveal that a Coanda UAV configured with 4 sides of straight, cylindrical Coanda surface achieve an overall 45% increase in lift compared to conventional Coanda Saucer UAV configurations. This venture integrates with an ongoing research project where a Coanda prototype is being assembled. Additionally, a custom thrust-stand has been constructed for thrust/lift measurement.
In recent years, maintenance optimization has attracted special attention due to the growth of industrial systems complexity. Maintenance costs are high for many systems, and preventive maintenance is effective when it increases operations' reliability and safety at a reduced cost. The novelty of this research is to consider general repair in the modeling of multi-unit series systems and solve the maintenance problem for such systems using the semi-Markov decision process (SMDP) framework. We propose an opportunistic maintenance policy for a series system composed of two main units. Unit 1, which is more expensive than unit 2, is subjected to condition monitoring, and its deterioration is modeled using a gamma process. Unit 1 hazard rate is estimated by the proportional hazards model (PHM), and two hazard rate control limits are considered as the thresholds of maintenance interventions for unit 1. Maintenance is performed on unit 2, considering an age control limit. The objective is to find the optimal control limits and minimize the long-run expected average cost per unit time. The proposed algorithm is applied to a numerical example to compare the effectiveness of the proposed policy (policy Ⅰ) with policy Ⅱ, which is similar to policy Ⅰ, but instead of general repair, replacement is performed. Results show that policy Ⅰ leads to lower average cost compared with policy Ⅱ.
In this study, the mechanical properties of alginate hydrogel material for self-standing 3D scaffold architecture with proper shape fidelity are investigated. In-lab built 3D bio-printer extrusion-based technology is utilized to fabricate 3D alginate scaffold constructs. The pressure, needle speed and stage speed are varied using a computer-controlled system. The experimental result indicates that the concentration of alginate solution, calcium chloride (CaCl2) cross-linking concentration and cross-linking ratios lead to the formation of alginate hydrogel with various gelation states. Besides, the gelling conditions, such as cross-linking reaction time and temperature also have a significant effect on the mechanical properties of alginate hydrogel. Various experimental tests such as the material gelation, the material spreading and the printability test for filament collapse as well as the swelling test were conducted to evaluate the fabricated 3D scaffold constructs. The result indicates that the fabricated 3D scaffold from composition of 3.5% wt alginate solution, that is prepared in DI water and 1% wt CaCl2 solution with cross-linking ratios of 7:3 show good printability and sustain good shape fidelity for more than 20 days, compared to alginate hydrogel that is prepared in a phosphate buffered saline (PBS). The fabricated self-standing 3D scaffold constructs measured 30 mm × 30 mm and consisted of 4 layers (n = 4) show good pore geometry and clear grid structure after printing. In addition, the percentage change of swelling degree exhibits high swelling capability with respect to time. The swelling test shows that the geometry of 3D alginate-scaffold construct and of the macro-pore are rarely changed, which indicates the capability of holding the shape fidelity during the incubation period. This study demonstrated that the mechanical and physical properties of alginate hydrogel could be tuned for a 3D bio-printing extrusion-based system to fabricate self-standing 3D scaffold soft structures. This 3D bioengineered scaffold provides a natural microenvironment present in the extracellular matrix of the tissue, which could be seeded with the biological cells to generate the desired 3D live tissue model for in vitro and in vivo tissue engineering applications.
In this paper, we describe how Bayesian inferential reasoning will contributes in obtaining a well-satisfied prediction for Distributed Constraint Optimization Problems (DCOPs) with uncertainties. We also demonstrate how DCOPs could be merged to multi-agent knowledge understand and prediction (i.e. Situation Awareness). The DCOPs functions were merged with Bayesian Belief Network (BBN) in the form of situation, awareness, and utility nodes. We describe how the uncertainties can be represented to the BBN and make an effective prediction using the expectation-maximization algorithm or conjugate gradient descent algorithm. The idea of variable prediction using Bayesian inference may reduce the number of variables in agents’ sampling domain and also allow missing variables estimations. Experiment results proved that the BBN perform compelling predictions with samples containing uncertainties than the perfect samples. That is, Bayesian inference can help in handling uncertainties and dynamism of DCOPs, which is the current issue in the DCOPs community. We show how Bayesian inference could be formalized with Distributed Situation Awareness (DSA) using uncertain and missing agents’ data. The whole framework was tested on multi-UAV mission for forest fire searching. Future work focuses on augmenting existing architecture to deal with dynamic DCOPs algorithms and multi-agent information merging.
The purpose of this article is to optimize the Equivalent Electric Circuit Model (EECM) of different orders to obtain greater precision in the modeling of Li-ion battery packs. Optimization includes considering circuits based on 1RC, 2RC and 3RC networks, with a dependent voltage source and a series resistor. The parameters are obtained experimentally using tests in the time domain and in the frequency domain. Due to the high non-linearity of the behavior of the battery pack, Genetic Algorithm (GA) was used to solve and optimize the parameters of each EECM considered (1RC, 2RC and 3RC). The objective of the estimation is to minimize the mean square error between the measured impedance in the real battery pack and those generated by the simulation of different proposed circuit models. The results have been verified by comparing the Nyquist graphs of the estimation of the complex impedance of the pack. As a result of the optimization, the 2RC and 3RC circuit alternatives are considered as viable to represent the battery behavior. These battery pack models are experimentally validated using a hardware-in-the-loop (HIL) simulation platform that reproduces the well-known New York City cycle (NYCC) and Federal Test Procedure (FTP) driving cycles for electric vehicles. The results show that using GA optimization allows obtaining EECs with 2RC or 3RC networks, with high precision to represent the dynamic behavior of a battery pack in vehicular applications.