Welcome to Journal of Beijing Institute of Technology
Convolutional Neural Network Based on Spatial Pyramid for Image Classification
Gaihua Wang, Meng Lü, Tao Li, Guoliang Yuan, Wenzhou Liu
2018, 27(4): 630-636.   doi: 10.15918/j.jbit1004-0579.17140
[Abstract](342) [PDF 359KB](311)
A novel convolutional neural network based on spatial pyramid for image classification is proposed. The network exploits image features with spatial pyramid representation. First, it extracts global features from an original image, and then different layers of grids are utilized to extract feature maps from different convolutional layers. Inspired by the spatial pyramid, the new network contains two parts, one of which is just like a standard convolutional neural network, composing of alternating convolutions and subsampling layers. But those convolution layers would be averagely pooled by the grid way to obtain feature maps, and then concatenated into a feature vector individually. Finally, those vectors are sequentially concatenated into a total feature vector as the last feature to the fully connection layer. This generated feature vector derives benefits from the classic and previous convolution layer, while the size of the grid adjusting the weight of the feature maps improves the recognition efficiency of the network. Experimental results demonstrate that this model improves the accuracy and applicability compared with the traditional model.
Inverse Kinematics Analysis of a 7-DOF Space Manipulator for Trajectory Design
Pengfei Xin, Jili Rong, Yongtai Yang
2017, 26(3): 285-291.   doi: 10.15918/j.jbit1004-0579.201726.0301
[Abstract](846) [PDF 1415KB](1093)
To solve the inverse kinematics problem for redundant degrees of freedom (DOFs) manipulators has been and still continues to be quite challenging in the field of robotics. Aiming at trajectory planning for a 7-DOF space manipulator system, joint rotation trajectories are obtained from predetermined motion trajectories and poses of the end effector in Cartesian space based on the proposed generalized inverse kinematics method. A minimum norm method is employed to choose the best trajectory among available trajectories. Numerical simulations with the 7-DOF manipulator show that the proposed method can achieve the planned trajectory and pose under the circumstances of minimum angular velocities. Moreover, trajectory results from the proposed kinematics model and inverse kinematics method has the advantages of simple modelling, low computation cost, easy to solve and plan trajectory conveniently. The smooth and continuous joint rotation functions obtained from the proposed method are suitable for practical engineering applications.
Design and implementation of path planning algorithm for vehicle parking
YANG Yi, QU Xin, ZHU Hao, ZHANG Lu, LI Xing-he
2016, 25(4): 502-511.   doi: 10.15918/j.jbit1004-0579.201625.0408
[Abstract](749) [PDF 2062KB](1309)
Parking is an important and indispensable skill for drivers. With rapid urban development, the automatic parking assistant system is one of the key components in future automobiles. Path planning is always essential for solving parking problems. In this paper, a path planning method is proposed for parking using straight lines and circular curves of different radius without collisions with obstacles. The parking process is divided into two steps in which the vehicle reaches the goal state through the intermediate state from the initial state. The intermediate state will be selected from the intermediate state set with a certain criterion in order to avoid obstacles. Similarly, an appropriate goal state will be selected based on the size of the parking lot. In addition, an automatic parking system is built, which effectively achieves the parking lot perception, path planning and performs parking processes in the environment with obstacles. The result of simulations and experiments demonstrates the feasibility and practicality of the proposed method and the automatic parking system.
Combined Transmission Interference Spectrum of No Core Fiber and BP Neural Network for Concentration Sensing Research
Fang Wang, Heng Lu, Yunpeng Li, Yufang Liu
2017, 26(2): 267-275.   doi: 10.15918/j.jbit1004-0579.201726.0217
[Abstract](404) [PDF 1660KB](441)
To investigate wavelength response of the no core fiber(NCF)interference spectrum to concentration, a three-layer back propagation(BP) neural network model was established to optimize the concentration sensing data. In this method,the measured wavelength and the corresponding concentration were trained by a BP neural network, so that the accuracy of the measurement system was optimized. The wavelength was used as the training set and got into the input layer of the three layer BP network model which is used as the input value of the network, and the corresponding actual concentration value was used as the output value of the network, and the optimal network structure was trained. This paper discovers a preferable correlation between the predicted value and the actual value, where the former is approximately equal to the latter. The correlation coefficients of the measured and predicted values for a sucrose concentration were 1.000 89 and 1.003 94; similarly, correlations of 0.999 51 and 1.018 8 for a glucose concentration were observed. The results demonstrate that the BP neural network can improve the prediction accuracy of the nonlinear relationship between the interference spectral data and the concentration in NCF sensing systems.
Structure Optimization and Performance Analysis of a Multiple Radial Magnetorheological Valve
Guoliang Hu, Fang Zhong, Haiyun Zhang, Ruqi Ding
2017, 26(4): 458-467.   doi: 10.15918/j.jbit1004-0579.201726.0405
[Abstract](494) [PDF 3971KB](545)
Due to the controllable and reversible properties of the smart magnetorheological (MR) fluid, a novel multiple radial MR valve was developed. The fluid flow channels of the proposed MR valve were mainly composed of two annular fluid flow channels, four radial fluid flow channels and three centric pipe fluid flow channels. The working principle of the multiple radial MR valve was introduced in detail, and the structure optimization design was carried out using ANSYS software to obtain the optimal structure parameters. Moreover, the optimized MR valve was compared with pre-optimized MR valve in terms of their magnetic flux density of radial fluid resistance gap and performance of pressure drop. The experimental test rig was set up to investigate the performance of pressure drop of the proposed MR valve under different currents applied and different loading cases. The results show that the pressure drop between the inlet and outlet port could reach 5.77 MPa at the applied current of 0.8 A. Furthermore, the experimental results also indicate that the loading cases had no effect on the performance of pressure drop.
Effect of Radial Resistance Gap on the Pressure Drop of a Compact Annular-Radial-Orifice Flow Magnetorheological Valve
Guoliang Hu, Jiawei Zhang, Mingke Liao, Ruqi Ding
2018, 27(4): 535-546.   doi: 10.15918/j.jbit1004-0579.17125
[Abstract](336) [PDF 7750KB](291)
A compact annular-radial-orifice flow magnetorheological (MR) valve was developed to investigate the effects of radial resistance gap on pressure drop. The fluid flow paths of this proposed MR valve consist of a single annular flow channel, a single radial flow channel and an orifice flow channel through structure design. The finite element modelling and simulation analysis of the MR valve was carried out using ANSYS/Emag software to investigate the changes of the magnetic flux density and yield stress along the fluid flow paths under the four different radial resistance gaps. Moreover, the experimental tests were also conducted to evaluate the pressure drop, showing that the proposed MR valve has significantly improved its pressure drop at 0.5 mm width of the radial resistance gap when the annular resistance gap is fixed at 1 mm.
Multiple mobile-obstacle avoidance algorithm for redundant manipulator
GUAN Xiao-qing, HAN Bao-ling, LIANG Guan-hao, CHANG Qing
2016, 25(1): 71-76.   doi: 10.15918/j.jbit1004-0579.201625.0111
[Abstract](547) [PDF 687KB](645)
In order to overcome the shortcomings of the previous obstacle avoidance algorithms, an obstacle avoidance algorithm applicable to multiple mobile obstacles was proposed. The minimum prediction distance between obstacles and a manipulator was obtained according to the states of obstacles and transformed to escape velocity of the corresponding link of the manipulator. The escape velocity was introduced to the gradient projection method to obtain the joint velocity of the manipulator so as to complete the obstacle avoidance trajectory planning. A 7-DOF manipulator was used in the simulation, and the results verified the effectiveness of the algorithm.
Co-simulation of a quadruped robot's mechanical and hydraulic systems based on ADAMS and AMESim
HAN Bao-ling, SI Shi-cai, LUO Qing-sheng, XIAO Da-hua, NIU Kai
2016, 25(2): 218-224.   doi: 10.15918/j.jbit1004-0579.201625.0209
[Abstract](391) [PDF 671KB](565)
In order to observe the change and fluctuation in flow and pressure of a hydraulic quadruped robot's hydraulic system when the robot walks on trot gait, a co-simulation method based on ADAMS and AMESim is proposed. Firstly, the change rule in each swing angle of the hydraulic quadruped robot's four legs is analyzed and converted to the displacement change of the hydraulic cylinder by calculating their geometric relationship. Secondly, the robot's dynamic model is built in ADAMS and its hydraulic and control system models are built in AMESim. The displacement change of the hydraulic cylinder in the hydraulic system is used as the driving function of the dynamics model in ADAMS, and the driving force of the dynamics model is used as the loads of the hydraulic system in AMESim. By introducing the PID closed-loop control in the control system, the co-simulation between hydraulic system and mechanical system is implemented. Finally, the curve of hydraulic cylinders' loads, flow and pressure are analyzed and the results show that they fluctuate highly in accordance with the real situation. The study provides data support for the development of a hydraulic quadruped robot's physical prototype.
Adaptive leader follower control for multiple quadrotors via multiple surfaces control
ABBAS Rabah, WU Qing-he
2016, 25(4): 526-532.   doi: 10.15918/j.jbit1004-0579.201625.0411
[Abstract](407) [PDF 340KB](727)
Based on the multiple surface and fixed undirected communication topology, the adaptive leader follower control for multiple quadrotors is discussed. Our approach is based on leader follower architecture. Multiple surface control (MSC) is used to design consensus controller to make multiple quadrotors construct a formation during flying with the presence of uncertainty item caused by the ground effect during landing or taking off. Simulation results are presented to validate the effectiveness of the proposed controller.
Image Classification Based on the Fusion of Complementary Features
Huilin Gao, Wenjie Chen
2017, 26(2): 197-205.   doi: 10.15918/j.jbit1004-0579.201726.0208
[Abstract](656) [PDF 3583KB](617)
Image classification based on bag-of-words (BOW) has a broad application prospect in pattern recognition field but the shortcomings such as single feature and low classification accuracy are apparent. To deal with this problem, this paper proposes to combine two ingredients:(i) Three features with functions of mutual complementation are adopted to describe the images, including pyramid histogram of words (PHOW), pyramid histogram of color (PHOC) and pyramid histogram of orientated gradients (PHOG). (ii) An adaptive feature-weight adjusted image categorization algorithm based on the SVM and the decision level fusion of multiple features are employed. Experiments are carried out on the Caltech 101 database, which confirms the validity of the proposed approach. The experimental results show that the classification accuracy rate of the proposed method is improved by 7%-14% higher than that of the traditional BOW methods. With full utilization of global, local and spatial information, the algorithm is much more complete and flexible to describe the feature information of the image through the multi-feature fusion and the pyramid structure composed by image spatial multi-resolution decomposition. Significant improvements to the classification accuracy are achieved as the result.
Analysis of vehicle powertrain dynamic performance
LUO Guo-liang, ZHANG Fu-jun, YUAN Hao-jie
2013, 22(2): 171-178.  
[Abstract](4415) [PDF 803KB](12)
To study the vehicle dynamic characteristics under typical cycle conditions, a steady-state simulation model of the engine in GT-Power is established and verified with engine bench test data. A dynamic model of the engine is then established. A co-simulation with the engine dynamic model in GT-Power and the vehicle transmission model in AMESim is conducted based on the technology of HLA/RTI. The parameter changes of vehicle powertrain in the accelerating process of 0-32.km/h, and vehicle typical cycle conditions are studied. The influence laws of the typical parameters influencing vehicle dynamic characteristics are obtained, and a new approach of improving vehicle dynamic characteristics is proposed. The results show that the vehicle powertrain dynamic model can simulate, analyze and predict dynamic changes of vehicle in actual operating conditions and guide powertrain matching and optimization.
Determining the characteristics of the initial fixed-focus of laser beam riding guidance information field
MEI Yue-song, YU Jian-qiao, YU Tian-peng, LI Jing-xu
2013, 22(1): 6-11.  
[Abstract](4441) [PDF 345KB](8)
Determining characteristics of the initial fixed-focus are a key technique for the design of laser beam riding guidance missile. Through analyzing the effects of the initial cone information field in guidance and control system of missile, those effects are considered as an approximate lead compensation network. The position and time of guidance spot initial fixed-focus can be designed. This method is applied to determine the characteristics of the initial fixed-focus of laser information field for a laser beam guidance gun-launched missile. The results of design and simulation show that the initial cone information field apparently speeds up convergence on the initial trajectory, and provides a guarantee for the implementation of minimum range index of a missile system.
Roundness error evaluation by minimum zone circle via microscope inspection
JIANG Li, ZHANG Zhi-jing, WU Wei-ren, JIN Xin, JIE De-gang
2013, 22(2): 185-190.  
[Abstract](3276) [PDF 353KB](5)
Utilizing the convex hull theory, a novel minimum zone circle (MZC) method, named improved minimum zone circle (IMZC) was developed in this paper. There were three steps for IMZC to evaluate the roundness error. Firstly, with the convex hull algorithm, data points on the circle contour were categorized into two sets to determine two concentric circles which contained all points of the contour. Secondly, vertexes of the minimum circumscribed circle and the maximum inscribed circle were found out from the previously determined two sets, and then four tangent points for determining the two concentric circles were also found out. Lastly, according to the evaluation using the MZC method, the roundness error was figured out. In this paper, IMZC was used to evaluate roundness errors of some micro parts. The evaluation results showed that the measurement precision using the IMZC method was higher than the least squared circle (LSC) method for the same set of data points, and IMZC had the same accuracy as the traditional MZC but dramatically shortened computation time. The computation time of IMZC was 6.89% of the traditional MZC.
Numerical simulation of non-contact explosion by door breaching explosive
JIANG Fei, HAN Feng, WANG Jian-zhong
2013, 22(1): 1-5.  
[Abstract](4667) [PDF 1809KB](8)
Aiming at the estimation of personal injury attached by counter-terrorist door breaching explosive blast wave, according to the actual scene, four typical application space models of counter-terrorist door breaching explosives are established, and numerical simulation of air-blast wave propagation by non-contact explosion counter-terrorist door breaching explosive are achieved. The research results show that the overpressure behind the target door is attenuated deeply through the burglary resistant safety door, and the propagation of blast wave and the damage effect under different space conditions are obviously different.
Submission & Review




Indexed by EI;Scopus

Eidtor-in-chief:Ran Tao,Beijing Institute of Technology,Beijing,China