Search

The feasibility of using chip-based RFID designs as wireless sensor tags open a wide range of application possibilities in the field of robotics. However, multi-step lithography manufacturing and/or MEMS techniques are often required for the industrial-grade fabrication of such sensors. In this paper, we present a simple, home-based, two-step fabrication process to produce chipless RF-based wireless sensors. We use an office-based inkjet printer to produce the antenna traces using silver conductive ink. Kirigami-inspired designs are used to produce four sensors responsive to various mechanical stimuli commonly used for DIY robotic projects (contact, compression, extension, and bend). We demonstrate sensing and wireless transmission of the detected mechanical stimuli through the proposed chipless sensor tags with reliable consistency. The tags can be replicated quickly with the inkjet-printing method. This paper also contains analyses on the effects of varying the dimensions and electrical parameters of the tags. The developed antennas in this work can be used as wireless mechano-responsive sensors for robotic applications.

Deep neural networks (DNNs) are becoming an integral part of most software systems. Previous work has shown that DNNs have bugs. Unfortunately, existing debugging techniques do not support localizing DNN bugs because of the lack of understanding of model behaviors. The entire DNN model appears as a black box. To address these problems, we propose an approach that automatically determines whether the model is buggy or not, and identifies the root causes. Our key insight is that historic trends in values propagated between layers can be analyzed to identify faults, and localize faults. To that end, we first enable dynamic analysis of deep learning applications: by converting it into an imperative representation and alternatively using a callback mechanism. Both mechanisms allows us to insert probes that enable dynamic analysis over the traces produced by the DNN while it is being trained on the training data. We then conduct dynamic analysis over the traces to identify the faulty layer that causes the error. We propose an algorithm for identifying root causes by capturing any numerical error and monitoring the model during training and finding the relevance of every layer on the DNN outcome. We have collected a benchmark containing 40 buggy models and patches that contain real errors in deep learning applications from Stack Overflow and GitHub. Our benchmark can be used to evaluate automated debugging tools and repair techniques. We have evaluated our approach using this DNN bug-and-patch benchmark, and the results showed that our approach is much more effective than the existing debugging approach used in the state of the practice Keras library. For 34 out of 40 cases, our approach was able to detect faults whereas the best debugging approach provided by Keras detected 32 out of 40 faults. Our approach was able to localize 21 out of 40 bugs whereas Keras did not localize any faults., Comment: Accepted at ICSE 2021

This paper focuses on the thermal design of the axial flux permanent magnet synchronous machine (AFPMSM). The performance of the spiral and parallel housing channels are compared in this paper. And an effective cooling enhanced method is employed to improve the temperature distribution of the disc rotor. The rectangular shapes of fins, which added to the internal surface of housing, are used to enhance heat convective transfer. This method can have significant influence in all types of AFPMSM with housing cooling, especially multi-rotor ones.

This paper studies the effect of rotor structure on the electromagnetic performances of a dual-rotor single-stator (DRSS) axial flux permanent magnet (AFPM) machine with segmented stator and concentrated winding. Consequent-pole PM (CPM) and Halbach array PM (HPM) are introduced in the rotor structure to improve the PM material utilization ratio and output torque, and the sensitivity analysis of some rotor geometrical structural parameters that have significant impact on the key performances is conducted. For the AFPM machine with SPM, CPM, CP-HPM and CP-IPM, several key electromagnetic characteristics including open-circuit airgap flux density, back-EMF, average torque, torque ripple, PM eddy current loss and PM demagnetization are compared under the same operating condition by using the finite-element method. It is demonstrated that compared with SPM, the AFPM with CPM has significant advantages in PM utilization and loss, although its performance in terms of torque and PM demagnetization is slightly sacrificed.

The relationship between the sales volume of differently efficiency vehicle and purchase subsidy was analyzed using the PVAR model based on panel data of 35 vehicles’ types of 6 enterprises in 33 months from 2016 to 2018. The study showed that purchase subsidies have a positive impact on the sales of high-efficiency vehicles continuously and the effect on low-efficiency vehicles changes from positive to negative over time. The result of variance decomposition also shows that subsidies contribute highly rate to the sales of high-efficiency vehicles. National subsidy policy should be adjusted to focus on high-efficiency vehicles and encourage the product with high technical level, such as long driving range and great battery system quality vehicles. And vehicles with inefficient power and poor quality should be phase out. Through these measures, the goals that purchase subsidies can promote enterprise’s core technical ability and the effect of the “quantity” to “quality” can be achieved.

This paper presents a quadrilateral mesh element based magnetic equivalent circuit framework (MECF) for an axial flux permanent magnet synchronous machine (AFPMSM). Firstly, the modelling process of quadrilateral mesh element is introduced. Then, the axial and circumferential air gap flux densities of the AFPMSM are calculated by two kinds of air gap mesh methods respectively. The proposed MECF method can not only consider the material nonlinearity, leakage flux and rotor movement, but also predict the locations where magnetic saturation occurs and small fluctuations of the air gap flux densities. Finally, the MECF results are compared with finite element analysis (FEA).

A novel high step-up converter is presented for Distributed PV Generation Systems, which combines with a hybrid asymmetrical voltage multiplier cell (HAVMC) and a coupled inductor. HAVMC is made up of an inductor, three diodes and three capacitors. Compared with the classical voltage multiplier cell, two couple capacitors in the cell operate in asynchronous manner to charging or discharging when main switch is turned on or turned off in a switching cycle. A novel passive lossless clamped circuit (PLCC) which is composed of a diode and two capacitors is combined into the novel converter to absorb the energy stored in leakage inductor of coupled inductor, allow a continuous input current, decrease the current ripple of the input source and realize the switch turn-on naturally. Furthermore, the reverse-recovery problem of diodes is alleviated owing to the leakage inductor. Its operating principles are analyzed in detail. Then, the analysis and performance have been fully validated experimentally on a 46V/430V prototype circuit at 50 kHz switching frequency.

This paper focuses on the multi-objective optimization of an air-cored axial flux permanent magnet synchronous machine (AFPMSM) with segmented PMs. Support vector machine based on radial basis kernel function is combined with orthogonal design to create the surrogate model between output performance and design parameters of a coreless AFPMSM with segmented PMs efficiently, and genetic algorithm is used to obtain the global optimal solutions to decrease the torque ripple and increase the output torque. The effectiveness of the proposed method is verified by finite element analysis.

Black swan is prevalent in emerging markets, leading to affect long-term returns. These outliers have stronger power to determine the performance. This paper from the perspective of big data by using these 9 emerging markets from the belt and road countries and about 60000 daily returns presents that the small outliers have a significant massive effect on long-term performance. Compared with some Asia countries, the Central and Eastern Europe countries' analysis results have some irreplaceable advantages, such as low risk, stable returns, and diversified advantages. From the data analysis, we draw the conclusion that black swans are caused by a cause which means it can be predicted in the forward investment market. For the long-term investment of emerging markets, in good and bad times, maintaining a certain degree of passivity and diversification of investment plays a crucial role.

This paper presents a detailed research of using dual-skew magnet topology in a Yokeless and segmented armature (YASA) machine for cogging torque minimization. The results show that compared with conventional skew magnet, dual-skew magnet has a better performance in minimization cogging torque while having low impact on back-EMF.

A novel strategy of model predictive direct power control (MPDPC) for PWM rectifier is proposed in this paper. Previous studies have proposed one-vector-based and two-vector-based MPDPC methods for PWM rectifier, which achieve good steady-state performance and quick dynamic responses. However, due to the limited switching states in the three-phase two-level PWM Rectifier, the sampling frequency should be high to obtain the satisfactory control performance. Besides, the switching frequency is variable, and this makes it difficult to design the filtering inductance. To deal with those problems, while maintaining all the advantages of the conventional MPDPC, this paper presents a new control strategy. The proposed MPDPC is based on an optimal application of three voltage vectors in a symmetrical way, which is known as symmetrical 3+3 vectors' sequence. The simulation and experimental results prove that, compared to conventional MPDPC, the proposed MPDPC achieves the steady-state harmonic spectrum at lower level and remains quick dynamic responses.

One of the challenges to analyze, test and debug Android apps is that the potential execution orders of callbacks are missing from the apps' source code, however, bugs, vulnerabilities and refactoring transformations have been found to be related to callback sequences. Existing work on control flow analysis of Android apps focuses on analyzing GUI behaviors. Our observation is that orthogonal to GUI, the Android API methods also play an important role in determining the order of callbacks, and previously, the APIs were modeled manually in an ad-hoc way. This paper presents a complementary solution of constructing program paths for Android apps. We proposed a specification called Predicate Callback Summary (PCS) that represents the callback control flow information (including callback sequences as well as the conditions the callbacks are invoked) in Android API methods. Our experiments show we can compute PCSs with reasonable accuracy and scalability and use them to build inter-callback control flow graphs for apps. Our detailed experimental data are available at: http://www.cs.iastate.edu/~weile/toolsdata/SummarizeAndroidFramework/lithium.html.

Bugs that cause reduced battery life in smartphones have been classified as energy bugs. The majority of previous research on energy bugs focused on the Android platform. The goal of this work is to identify the similarities and differences of energy bugs for the Android, iOS and Windows platforms, based on which, we can then determine whether the techniques of managing energy bugs developed for one platform can be potentially applicable in other platforms. Our results show that applications in Android and Windows share similar root causes for leading to battery drain, but energy bugs in iOS applications are produced differently. Our conclusions are drawn based on a comparison of power models for the three platforms and an analysis of the bugs from 6 applications common to Android, iOS and Windows.

Many gas reservoirs with high H2S content have been found in China, such as Luojiazhai Gas Field and Puguang Gas Field, and they play an important role in China's natural gas industry. However, there are a series of serious problems in production of these gas reservoirs, such as H2S poison, strong corrosion, and sulfur deposition. Therefore, it is important to gas flowing theory study on gas reservoirs with high H2S content. This paper discusses sour gas flow and sulfur deposition. Based on the experimental work and aerodynamics gas-solid flow theory, three-dimensional multi-component mathematical model for sour gas reservoir is developed. In the model, factors of affecting sulfur particle deposition has been studied, including gas flow rate, initial H2S concentration and formation permeability.

This paper presents a blind method for estimating both the insertion loss and crosstalk channel in multi-user Digital Subscriber Line (DSL) systems. Based on cyclostationary properties of Discrete Multi-tone (DMT) signal, we select appropriate cyclic suffix length of DMT signal in the transmitter side such that the receiver reserves the insertion loss channel information of the discussed line and crosstalk channels information from other lines on the chosen tone, while the insertion loss channel effects from other lines and the crosstalk effect on the current line from other users on the chosen tone are all zeros. Then the insertion loss and crosstalk channels can respectively be estimated from received data on each line. Computer simulation results based on real measured data verify that the proposed method can estimate both the insertion loss and crosstalk channels. Furthermore, compared to the previous works, the new method has lower estimation errors and is easy to be implemented.

For the purpose of overcoming the influence of process nonlinearities absolutely and improving the load-following capability, aiming at the typical model of boiler-turbine coordinate system and based on nonlinear method of backstepping, some dummy variables are introduced to construct the Lyapunov function of error signals. A nonlinear coordinated controller for boiler-turbine units is designed. The dynamic performance, disturbance rejection, and robustness, are proved by 3 simulation tests in different operating points.

In order to speed up the VQ encoding process, it is beneficial to firstly estimate how large the Euclidean distance is between the input vector and a candidate codeword by using appropriate low dimensional features of a vector instead of an immediate Euclidean distance computation. The sum, the partial sums and the partial variances of a vector are used together as the appropriate features in this paper. Utilizing these features between the input vector and a candidate codeword, we obtain some inequalities of eliminating unnecessary codewords. First the inequality based on the sum is used for terminating the searching process. Then, using other inequalities based on subvector sum and variance, more unnecessary codewords are eliminated. During the search process for code, the proposed algorithm can reject a lot of codewords, while introducing no extra distortion compared to the conventional full search algorithm. Experimental results show that the reduction of computation is outstanding in the computational complexity and the number of distortion calculations against full search method.

For improving the load adaptability of main steam temperature control system, a load scheduling control structure is presented for the power generation unit. In this structure, the steam flow is selected as the scheduling parameter and four sets of local-controllers are embedded which are corresponding to four typical conditions. Each local-controller is designed and tuned by the cascade internal model control theory. The smooth switching among local-controllers is realized by the membership functions of fuzzy subsets. Because of the high order of outer-loop controllers, Taylor series expanded is adopted to reduce them to real PID. The presented control structure keeps the advantages of traditional cascade system: disturbance rejection, frequency increment, etc. Simulation tests show that the load scheduling control structure has good dynamic performance and can work well in a wide range of load.