JOURNAL OF BEIJING INSTITUTE OF TECHNOLOGY

Object Recognition Algorithm Based on an Improved Convolutional Neural Network

Zheyi Fan, Yu Song, Wei Li

2020, 29(2): 139-145. doi:10.15918/j.jbit1004-0579.19116

[Abstract](856) [PDF 662KB](595)

Abstract:
In order to accomplish the task of object recognition in natural scenes, a new object recognition algorithm based on an improved convolutional neural network (CNN) is proposed. First, candidate object windows are extracted from the original image. Then, candidate object windows are input into the improved CNN model to obtain deep features. Finally, the deep features are input into the Softmax and the confidence scores of classes are obtained. The candidate object window with the highest confidence score is selected as the object recognition result. Based on AlexNet, Inception V1 is introduced into the improved CNN and the fully connected layer is replaced by the average pooling layer, which widens the network and deepens the network at the same time. Experimental results show that the improved object recognition algorithm can obtain better recognition results in multiple natural scene images, and has a higher degree of accuracy than the classical algorithms in the field of object recognition.

Anisotropic Total Variation Regularization Based NAS-RIF Blind Restoration Method for OCT Image

Xuesong Fu, Jianlin Wang, Zhixiong Hu, Yongqi Guo, Kepeng Qiu, Rutong Wang

2020, 29(2): 146-157. doi:10.15918/j.jbit1004-0579.20014

[Abstract](607) [PDF 1663KB](539)

Abstract:
Based on anisotropic total variation regularization (ATVR), a nonnegativity and support constraints recursive inverse filtering(NAS-RIF) blind restoration method is proposed to enhance the quality of optical coherence tomography (OCT) image. First, ATVR is introduced into the cost function of NAS-RIF to improve the noise robustness and retain the details in the image. Since the split Bregman iterative is used to optimize the ATVR based cost function, the ATVR based NAS-RIF blind restoration method is then constructed. Furthermore, combined with the geometric nonlinear diffusion filter and the Poisson-distribution-based minimum error thresholding, the ATVR based NAS-RIF blind restoration method is used to realize the blind OCT image restoration. The experimental results demonstrate that the ATVR based NAS-RIF blind restoration method can successfully retain the details in the OCT images. In addition, the signal-to-noise ratio of the blind restored OCT images can be improved, along with the noise robustness.

2D Augmented Coprime Array Geometry Based on the Difference and Sum Coarray Concept

Guiyu Wang, Shun’an Zhong, Xiangnan Li, Xiaohua Wang, Shiwei Ren

2020, 29(2): 158-166. doi:10.15918/j.jbit1004-0579.19120

[Abstract](731) [PDF 967KB](492)

Abstract:
The concept of difference and sum (diff-sum) coarray has attracted a lot of attentions in the estimation of direction-of-arrival (DOA) for the past few years, due to its high degrees-of-freedom (DOFs). A vectorized conjugate augmented MUSIC (VCA-MUSIC) algorithm is applied to generate an equivalent signal model which contains the virtual sensor positions of both the difference and sum of the physical sensors in the two-dimensional (2D) arrays, by utilizing both the spatial and temporal information. Besides, an augmented 2D coprime array configuration is presented with the basis on the concept of difference and sum coarray. By compressing the inter-element spacing of one subarray and introducing the proper separation between the two subarrays of 2D coprime array, the redundancy between the difference coarray and the sum one can be reduced so that more virtual sensors in both coarrays can make contributions to the DOFs. As a result, a much larger consecutive area in the diff-sum coarray can be achieved, which can significantly increase the DOFs. Numerical simulations verify the superiority of the proposed array configuration.

Simplified Modeling and Mathematical Analysis in Cardiac Electric Field Propagation for Human and Animals

Ping Lang, Lei Han

2020, 29(2): 167-176. doi:10.15918/j.jbit1004-0579.19104

[Abstract](518) [PDF 951KB](523)

Abstract:
The different characteristics of cardiac electric field (CEF) radiation in humans and other animals are presented in this paper. Physical modeling and mathematical analysis are developed to comprehensively unveil the properties of CEF, based on typical heartbeat waveforms. Our numerical simulation results demonstrate that the frequency bandwidths and the cycle durations of CEF are different for healthy humans versus humans on the verge of death and for humans versus other animals. The results indicate that the present study may extensively contribute towards recognizing human beings or other animal targets quickly and accurately with CEF in dangerous situations or in other applications.

SystolicB^–1Circuit in Galois Fields Based on a Quaternary Logic Technique

Haixia Wu, Yilong Bai, Tian Wang, Xiaoran Li, Long He

2020, 29(2): 177-183. doi:10.15918/j.jbit1004-0579.19119

[Abstract](573) [PDF 661KB](1017)

Abstract:
In order to improve the circuit complexity and reduce the long latency ofB^–1operations, a novelB^–1operation in Galois Field GF(2⁴) is presented and the corresponding systolic realization based on multiple-valued logic (MVL) is proposed. The systolic structure employs multiple-valued current mode (MVCM) by using dynamic source-coupled logic (SCL) to reduce the initial delay and the transistor and wire counts. The performance is evaluated by HSPICE simulation in 0.18 μm CMOS technology and a comparison is conducted between our proposed implementation and those reported in the literature. The initial delay and the sum of transistors and wires in our MVL design are about 43% and 13% lower, respectively, in comparison with other corresponding binary CMOS implementations. The systolic architecture proposed is simple, regular, and modular, well suited for very large scale integration (VLSI) implementations. The combination of MVCM circuits and relevant algorithms based on MVL seems to be a potential solution for high performance arithmetic operations in GF(2^k).

Research of Current Mode Atomic Force Microscopy (C-AFM) for Si/SiC Heterostructures on 6H-SiC(0001)

Song Feng, Lixun Song, Yuan Zang, Zheyan Tu, Lianbi Li

2020, 29(2): 184-189. doi:10.15918/j.jbit1004-0579.20009

[Abstract](669) [PDF 898KB](460)

Abstract:
Si/SiC heterostructures with different growth temperatures were prepared on 6H-SiC(0001) by LPCVD. Current mode atomic force microscopy and transmission electron microscopy were employed to investigate the electrical properties and crystalline structure of Si/SiC heterostructures. Face-centered cubic (FCC) on hexagonal close-packing (HCP) epitaxy of the Si(111)/SiC(0001) heterostructure was realized at 900℃. As the growth temperature increases to 1050℃ , the <110> preferred orientation of the Si film is observed. The Si films on 6H-SiC(0001) with different growth orientations consist of different distinctive crystalline grains: quasi-spherical grains with a general size of 20 μm, and columnar grains with a typical size of 7 μm×20 μm. The electrical properties are greatly influenced by the grain structures. The Si film with <110> orientation on SiC(0001) consists of columnar grains, which is more suitable for the fabrication of Si/SiC devices due to its low current fluctuation and relatively uniform current distribution.

Structural Damage Law of Semiconductor Bridge Detonator under Impact and Overload Environment

Dongxiao Fu, Rui Zhang, Hu Liu, Fang Li, Zhenhua Du, Hongliang Ma

2020, 29(2): 190-194. doi:10.15918/j.jbit1004-0579.19117

[Abstract](543) [PDF 475KB](440)

Abstract:
Aiming to know the requirement of penetrating the munition semiconductor bridge detonator under the impact overload environment, the impact overload simulation device and the structural finite element software ANSYS/AUTODYN are used to study the variation of the axial dimension, charge and the chip gap of the semiconductor bridge detonator under the impact overload environment. The typical semiconductor bridge detonator is affected by the acceleration, and the strain increases with the increase of the acceleration. The semiconductor bridge detonator shows axial compression, in which the size becomes smaller, and the structural deformation occurs at the output end of the semiconductor bridge detonator. The typical semiconductor bridge detonator is elastically deformed when the acceleration is less than 40 000g. When the acceleration is more than 40 000g, the semiconductor bridge detonator housing is plastically deformed. The gap between the drug column and the chip is divided into three stages with the increase of the acceleration. Initially, with the increase of the acceleration, the gap rises rapidly until the acceleration reaches 43 000g, and when the gap reaches the maximum, the gap decreases rapidly with the increase of the acceleration. When the acceleration reaches 57 000g, the gap tends to be 0 μm in the initial state, and then the gap does not change with the acceleration to keep tending to 0 μm.

Experimental and Numerical Study on the Gas Explosion in Urban Regulator Station

Zhenyi Liu, Yuanyuan Ma, Yuan Ren, Mingzhi Li, Pengliang Li, Song Wan

2020, 29(2): 195-208. doi:10.15918/j.jbit1004-0579.19115

[Abstract](620) [PDF 5884KB](428)

Abstract:
Regulator station is an important part in the urban gas transmission and distribution system. Once gas explosion occurs, the real explosion process and consequences of methane gas explosion in the regulator station were not revealed systematically. In this study, a full-scale experiment was carried out to simulate the regulator station explosion process, and some numerical simulations with a commercial CFD software called FLACS were conducted to analyze the effect of ignition and vent conditions on the blast overpressure and flame propagation. The experimental results demonstrated that the peak overpressure increased as the distance from the vent increased within a certain distance. And the maximum overpressure appeared 3 m away from the door, which was about 36.6 kPa. It was found that the pressure-time rising curves obtained from the simulation are basically the same as the ones from the experiment, however, the time of reaching the peak pressure was much shorter. The numerical simulation results show that the peak overpressures show an increase trend as the ignition height decreased and the vent relief pressure increased. It indicates that the damage and peak overpressure of gas explosion could be well predicted by FLACS in different styles of regulator station. In addition, the results help us to understand the internal mechanism and development process of gas explosion better. It also offers technical support for the safety protection of the urban regulator station.

Numerical Simulation of Elastic Stress Wave Refraction at Air-Solid Interfaces

Ruchao Shi, Xiaowang Sun

2020, 29(2): 209-221. doi:10.15918/j.jbit1004-0579.19092

[Abstract](545) [PDF 1792KB](457)

Abstract:
Elastic wave refraction at the air-solid interface and wave propagations in the vicinity of the air-solid interface are numerically studied. The modified ghost fluid method (MGFM) and isobaric fix methods are combined to solve the fluid and solid statuses at the air-solid interface and construct a continuous boundary condition for the air-solid interface. The states in the ghost domain are evaluated by the MGFM-algorithm. The solid governing equations are solved with second order spatial discretization. Numerical tests verify the correctness of the presented continuous boundary condition and show that the combined method is convergent in the vicinity of the air-solid interface. The 3D numerical results by the combined method are close to those of the Arbitrary-Lagrangian-Eulerian (ALE) method. The combined method is robust when applied for multi-dimensional problems. A compression stress wave impacting on the air-solid interface result in a compression wave in air. A tension stress wave impacting on the air-solid interface result in an expansion wave in air.

Radiator Optimization Design for Planar Motors Based on Parametric Components

Ming Zhang, Qiaomu Xu, Rong Cheng, Leijie Wang, Xin Li

2020, 29(2): 222-231. doi:10.15918/j.jbit1004-0579.20016

[Abstract](541) [PDF 834KB](400)

Abstract:
Focusing on the design problem of high-performance radiators for planar motors in the wafer stage of the lithography machine, a thermal-fluid coupling optimization scheme based on parametric solid components was proposed. The mapping method between component parameters and pseudo-density values was established. An analytical solution for the sensitivity of pseudo-density to component parameters was given. The conjugate heat transfer function with the shallow channel approximation term was solved through the pseudo-density information. In the optimization example, circular components were selected, and the position and the size of solid components were chosen as design variables. In order to eliminate calculation errors caused by pseudo-density, an optimized pseudo-density field was converted into the result based on parametric components. Compared to the reference motor radiator, the average surface temperature rise of the optimized water-cooling motor radiator is reduced by 22.4%, which verifies the feasibility and effectiveness of the proposed method.

Characteristics Extraction of Vehicle State Information Based on Entropy Calculation

Zepeng Gao, Zheng Liu, Sizhong Chen, Hongbin Ren, Zechao Li, Yong Chen

2020, 29(2): 232-240. doi:10.15918/j.jbit1004-0579.19118

[Abstract](674) [PDF 1327KB](606)

Abstract:
A method of extracting and detecting vehicle stability state characteristics based on entropy is proposed. The vehicle’s longitudinal and lateral dynamics models are established for complex driving and maneuver conditions. The corresponding state observer is designed by adopting the moving horizon estimation algorithm, which realizes the observation of the vehicle stability state considering the global state information. Meanwhile, the Shannon entropy is modified to approximate entropy, and the approximate entropy value of the observed vehicle state is calculated. Furthermore, the optimal controller is designed to further validate the reliability of the entropy value as the reference of control system. Simulation results demonstrate that this method can quickly detect the instability state of the system during the process of vehicle driving, which provides a reference for risk prediction and active control.

Research on Measurement Method of Muffler Performance under High Temperature and High-Speed Airflow Conditions

Liping Xie, Chihua Lu, Zhien Liu, Yawei Zhu

2020, 29(2): 241-250. doi:10.15918/j.jbit1004-0579.19127

[Abstract](615) [PDF 892KB](440)

Abstract:
Ignoring the influence of flow velocity and flow temperature on muffling performance, performance tests were conducted without airflow in the development phase of the muffler which accounts to the difficulty of obtaining a perfect match between the actual noise reduction effect and the design goal. Based on the two-load test theory, a set of high-temperature and high-speed airflow simulation measurement devices for the muffler has been built. In order to avoid the impact of high-temperature and high-speed airflow on the sensor, a high temperature resistant sensor holder has been designed for the test rig. The sound pressure has been measured in the pipe by using the lead-out measurement. In addition, a variable impedance load is placed at the end of the test tube to realize the switch between two different impedance loads by the wave handle of the variable impedance load. A sound source correction method is proposed to decrease the random fluctuation of the spectral characteristics of the output noise signal, which is caused by the acoustic impedance variation at the connection between the transition pipeline and the combined sound source system. Finally, an acoustic software has been used to calculate the transmission loss of the muffler in the presence of high temperature airflow. In comparing the experimental measurements and the simulation results, the small difference shows that: the bench not only can effectively simulate high-temperature and flow velocity environment of the engine but also accurately test the transmission loss of the muffler.

Novel Design Method of COG Measurement System Via Supporting Reaction Method

Qirong Zhang, Xin Jin, Kai Zhou, Zhongqing Zhang, Dongmei Liu

2020, 29(2): 251-259. doi:10.15918/j.jbit1004-0579.20011

[Abstract](526) [PDF 1033KB](476)

Abstract:
Center of gravity (COG) is an important parameter of projectiles and rockets, for which supporting reaction method(or support reaction method) is an important COG measurement method. Based on this supporting reaction method a novel design method is proposed to determine the key design parameters of the COG measurement system. The method can quantitatively analyze the influence of the design parameters on the COG accuracy, in the measurement system designed with supporting reaction method. Using the principle of static balance, the error propagation theory, and the system accuracy analysis method, the equal-range required sensor precision (RSP) surface and non-equal-range required sensor pair precision (RSPP) adapted surface are adopted. The influence of random errors(like sensor accuracy and distance calibration accuracy) is analyzed. The selection strategy of equal-range and non-equal-range sensors is chosen, and then the recommended calibration accuracy values are obtained. For the quality detection accuracy of ±0.6 kg and the axial COG detection accuracy of ±1.5 mm, the RSP surface is drawn by the proposed method, and the force sensor with ±0.23 kg detection accuracy is selected. The experimental verification meets the accuracy requirements and verifies the effectiveness of the proposed design method for the system parameters of the COG measurement equipment.

Design of a 70 MPa Two-Way Proportional Cartridge Valve for Large-Size Hydraulic Forging Press

Jing Yao, Yuxin Yin, Zhaosheng Dong, Yuantong He

2020, 29(2): 260-272. doi:10.15918/j.jbit1004-0579.20006

[Abstract](732) [PDF 1718KB](486)

Abstract:
For an ultra-high-pressure hydraulic transmission system of a large-size hydraulic forging press (LHFP), a 70 MPa two-way proportional cartridge valve has been developed to improve the power weight ratio of the hydraulic forging press. In this study, a nominal diameter 25 mm (DN25) cartridge valve is taken as the research object. A longer concentric cylindrical annular gap is set to effectively prevent the ultra-high-pressure oil from flowing to the pilot stage and a seated valve structure is set to form the linear sealing zone in the closing state of the main valve port. Electric-displacement feedback is adopted to realize precise control of the main valve port flow and the features of this valve are investigated. In order to verify the strength and static and dynamic characteristics, the finite element model and a simulation model of the valve proposed above are built. There is a little deformation which does not affect the main valve spool movement, and the main valve port flow meets the design demands. Then, the prototype of DN25 70TPCV is manufactured and a ultra-high-pressure experimental platform is developed. The experimental results show that the DN25 70TPCV designed in this study has the advantage of fast response, high control precision, and low leakage, which can meet the requirements of LHFPs.

2020 Vol. 29, No. 2