Table of Content

25 September 2020, Volume 42 Issue 9 Previous Issue    Next Issue

Select

Safety Field-based Improved RRT* Algorithm for Path Planning of Intelligent Vehicle Zhu Bing, Han Jiayi, Zhao Jian, Liu Shuai, Deng Weiwen 2020, 42 (9):  1145-1150.  doi: 10.19562/j.chinasae.qcgc.2020.09.001 Abstract ( 364 )   PDF (2929KB) ( 473 )   Save Rapidly-exploring random tree (RRT) algorithm is a common algorithm for path planning of intelligent vehicle. But traditional RRT and RRT* algorithms have disadvantages of large path jitter, easy to fall into local region and low calculation efficiency. In view of these problems, an improved RRT* algorithm for the path planning of intelligent vehicle based on safety field and real vehicle driving data is proposed in this paper. Firstly, a safety field based on safety distance model is established, and the key parameters of the model are extracted through driving data acquisition test. On this basis, an improved RRT* algorithm with safety field guidance and angle constraint strategies is proposed. Finally, the algorithm is verified by simulation. The results show that the path planning method proposed can calculate the effective path meeting the curvature constraint of vehicle trajectory with faster search speed and higher success rate References | Related Articles | Metrics

Select

Study on Path Planning and Tracking Control for Intelligent Vehicle Based on RRT and MPC Zhou Wei, Guo Xuexun, Pei Xiaofei, Zhang Zhen, Yu Jiaxing 2020, 42 (9):  1151-1158.  doi: 10.19562/j.chinasae.qcgc.2020.09.002 Abstract ( 475 )   PDF (2210KB) ( 684 )   Save In order to analyze the mutual influence between real-time planning and tracking control of smart car, a new architecture of path planning and tracking control for intelligent vehicle is proposed based on improved rapidly-exploring random tree (RRT) algorithm and linear time-varying model predictive control (LTV-MPC) algorithm. Firstly, basic RRT algorithm is modified by target orientation, node pruning, curve fitting and optimal path selection to ensure the planned path meets the vehicle kinematic constraint requirements and approaches the optimal solution. Then, the stability control on the desired path of intelligent vehicle is achieved based on LTV-MPC algorithm. The results of hardware-in-the-loop simulation show that with a vehicle speed of 36 km/h, a planning step of 2 m and a planning cycle of 0.1 s, the lateral acceleration is less than 0.2g, meeting the requirements of safety and real-time performance. Finally, the effects of factors such as vehicle speed, planning step and planning cycle on real-time planning and stability tracking are analyzed References | Related Articles | Metrics

Select

Simulation and Application of Passenger Vehicle Drivability Based on GT-SUITE Software Zeng Hao, Zheng Guangyong, Zhang Caixia, Kong Wei 2020, 42 (9):  1159-1165.  doi: 10.19562/j.chinasae.qcgc.2020.09.003 Abstract ( 191 )   PDF (2286KB) ( 234 )   Save In order to evaluate the drivability in the early stage of vehicle performance development, GT-SUITE software is used in this paper to establish the passenger vehicle drivability simulation model with consideration of the engine transient response and torque control strategy. The simulation model is used to simulate some typical driving conditions, and the model is corrected by test data to ensure that the simulation error is within 10%. Taking a 6-gear automatic transmission gasoline vehicle as an example, according to the different requirements of three driving modes, the throttle characteristics and the shifting schedule are calibrated virtually respectively. Finally, the virtual calibration results are verified by experiments, and the error between the simulation and testis within 8%, which fully verifies the effectiveness and practicability of the virtual calibration of drivability References | Related Articles | Metrics

Select

Vehicle Speed Control of Electric Vehicle Driving Robot Based on Inverse Control Strategy Model Chu Cancan, Wang Dong, Zhang Weigong, Xu Tong 2020, 42 (9):  1166-1173.  doi: 10.19562/j.chinasae.qcgc.2020.09.004 Abstract ( 174 )   PDF (1594KB) ( 252 )   Save For meeting the requirements of driving robot in using accelerator pedal travel to achieve the speed following of electric vehicle, a vehicle speed control scheme based on inverse control strategy model is proposed. The scheme introduces the model for vehicle control strategy on the basis of the speed control circuit of traditional longitudinal kinetics, and transforms the original vehicle speed control by electric motor torque into that by accelerator pedal travel, and hence reduces the risk of retrofitting the original vehicle to driving robot and thus enhances robot applicability. The results of dynamometer test show that compared with manual driving, the control method proposed effectively increases the accuracy of vehicle speed tracking with an error not more than±1 km/h, meeting the national standard for electric vehicle test References | Related Articles | Metrics

Select

Prediction of Continued Driving Range of Battery Electric Vehicle Based on Map Information and Cyclic SVR Model Tian Huixin, Li Xiaoyu, Liu Fang 2020, 42 (9):  1174-1182.  doi: 10.19562/j.chinasae.qcgc.2020.09.005 Abstract ( 209 )   PDF (1454KB) ( 265 )   Save In view of the situation that the prediction accuracy of continued driving mileage is hard to enhanced due to the unknown driving conditions in the future, a prediction method for the driving mileage of battery electric vehicle based on map information and cyclic SVR model is proposed. The method predicts the future driving conditions according to map information, and takes the corresponding parameters of working conditions as the input of SVR model to calculate the SOC decline value and the remaining SOC value of the unit mileage in that working condition. The process is iterated repeatedly until the SOC value returns to zero, then the number of SVR runs is the remaining driving range. Finally, a simulation is carried out using ADVISOR based on actual driving data, and the results show that the method has a high prediction accuracy of continued driving mileage. References | Related Articles | Metrics

Select

Study of DPF Carbon Load Prediction Model Based on Flow Resistance Shi Xiuyong, Jiang Degang, Liang Yunfang, Liang Pengfei 2020, 42 (9):  1183-1188.  doi: 10.19562/j.chinasae.qcgc.2020.09.006 Abstract ( 176 )   PDF (1728KB) ( 263 )   Save Accurate prediction of carbon load is the key to the application of diesel particulate filter (DPF) technology. Firstly, the major factors influencing DPF differential pressure value are analyzed theoretically in this study. It is found that the differential pressure is primarily affected by exhaust gas volume flow rate when the physical structure and internal carbon load of DPF are given. Therefore, the flow resistance value is proposed to be used to characterize the correlation between the differential pressure and exhaust volume flow under specific carbon load. Secondly, a one-dimensional model of DPF is built using GT-Power software, simulating the influence of different carbon load in DPF on flow resistance. The analysis shows that the differential pressure value fluctuates greatly under transient conditions due to the influence of engine running conditions, but the flow resistance value keeps relatively stable and is positively correlated with the carbon load in DPF. Finally, the DPF simulation model is verified by engine bench test, which are in good agreement, with a maximum error of 10% References | Related Articles | Metrics

Select

Capacity Prediction Method of Lithium-ion Battery Under Random Discharge Condition Sun Daoming, Yu Xiaoli 2020, 42 (9):  1189-1196.  doi: 10.19562/j.chinasae.qcgc.2020.09.007 Abstract ( 200 )   PDF (1289KB) ( 313 )   Save For the problem of low accuracy of lithium-ion battery capacity prediction, a seeking optimization algorithm-support vector machine (SOA-SVM) based capacity prediction method is proposed. By analyzing the random discharge process of lithium-ion battery, two indicators, the mean and standard error of random discharge capacity reflecting the capacity change of lithium-ion battery are constructed which are used as the feature parameters for capacity prediction. The principle component analysis is used to analyze the correlation between the feature parameters and extract the principle components. Based on the first principle component and the capacity data of part of tested batteries, SOA is used to optimize hyper-parameters of SVM and train the model. The optimized model combined with the first principle component date of other batteries is adopted to predict the capacity of lithium-ion batteries. The prediction results show that the proposed capacity prediction method has high prediction accuracy References | Related Articles | Metrics

Select

Cooling Airflow Simulation and Heat Dissipation Improvement of Engine Compartment Li Tiantian, Zhao Lanping, Wang Jianxin, Zhu Zhijun, Zhang Jun, Zhang Hao 2020, 42 (9):  1197-1205.  doi: 10.19562/j.chinasae.qcgc.2020.09.008 Abstract ( 167 )   PDF (11461KB) ( 136 )   Save In order to reduce cooling drag, a numerical simulation on the air flow of original engine compartment is conducted. Based on the results of simulation, different improvement schemes are proposed from three aspects: changing the flow direction of cooling air, reducing the energy loss of air flow passing through front-end components and enhancing radiator performance, on which wind tunnel tests are carried out for verification. The results show that after front-end plate separator is added the flow rate of intake air increases by 15%, and cooling drag lowers by 5 counts. As for the heat dissipation within engine compartment, by combining simulation and test, several measures are taken, such as adding baffle plate behind steering box to intensify the local air flow, adding flow guiding shroud behind fan for guiding hot airflow and adding flow guiding device at the intake port of engine to supply extra cooling air flow. As a result, the heat dissipation of steering box and transmission mounts are improved with the cooling air flow in front-end unaffected References | Related Articles | Metrics

Select

Study on Spectrum Smoothness for Vehicle Interior Wind Noise Evaluation Shen Zhe, Wang Yigang, Yang Zhigang, He Yinzhi, Peng Liqi 2020, 42 (9):  1206-1210.  doi: 10.19562/j.chinasae.qcgc.2020.09.009 Abstract ( 187 )   PDF (1024KB) ( 269 )   Save While the existing evaluation index for vehicle interior wind noise can only reflect macro-weighting of each frequency component with a lack of judgment on partial spectrum problems, it needs to be perfected. By analyzing wind tunnel test data of 23 different mass-produced vehicles, the positive correlation between vehicle type grade and interior wind noise performance is verified,and the spectrum regularity of superior samples is summarized. The results illustrate that leak noise is the main cause of roughness in narrow-range of medium-high frequency band, which should be avoided as far as possible. By adding up the difference between the moving-average spectrum and original spectrum, a method to obtain the quantitative index of spectral roughness is proposed. The results show that spectrum smoothness is closely related to interior wind noise performance and especially sensitive to leak noise, which can be used to evaluate the vehicle interior wind noise level References | Related Articles | Metrics

Select

Research on Aerodynamic Characteristics of Rear Spoiler in an Electric SUV Li Xianjin, Li Zhelong 2020, 42 (9):  1211-1215.  doi: 10.19562/j.chinasae.qcgc.2020.09.010 Abstract ( 175 )   PDF (1571KB) ( 307 )   Save The fuel economy and stability of vehicle in high-speed driving is directly related to the aerodynamic drag and lift of the vehicle. Compared with traditional fuel vehicles, reducing aerodynamic drag is much more important for electric vehicle in increasing driving mileage and lowering energy consumption. In this paper, Reynolds time-average scheme is adopted to conduct a vehicle external flow field simulation on a battery electric SUV at a speed of 120 km/h, with the results of drag and lift coefficients compared with that of the wind-tunnel test of clay model of the same scale. The common Realizable k-ε turbulent model is used to perform a simulation optimization on the rear spoiler of that SUV to study the effects of the different declined angles of the upper surface of spoiler on the drag and lift coefficients of vehicle. Furthermore, on the basis of best angle obtained, the optimal state of rear spoiler is determined by the comparatively analysis on five different types of through holes on spoiler. The results of final validation test show that the aerodynamic drag coefficient reduces by 3.9%, while aerodynamic lift coefficient increases, but in acceptable range References | Related Articles | Metrics

Select

Optimization on Parameters of Active Safety Airbag by Using NSGA-Ⅱ Ge Ruhai, Cui Yizhong, Hong Liang, Xiao Xuan 2020, 42 (9):  1216-1223.  doi: 10.19562/j.chinasae.qcgc.2020.09.011 Abstract ( 162 )   PDF (1998KB) ( 281 )   Save The occupant restraint system of school bus is the key to protecting the safety of child occupant. The effects of control parameters such as the length of belt, the gas mass flow rate, the installation position, the opening degree of vent hole, the initial pressure of airbag, the initial pressure of exhaust and the distance between seats on the protection results of child occupant are studied in this paper. By using sensitivity analysis, the key parameters for children occupant protection are selected, i.e. the belt length, the vent hole opening degree, the exhaust initial pressure and the distance between seats. Based on the design of experiment with Latin hypercube sampling and polynomial response surface model, the surrogate models for weighted injury criteria (WIC), neck injury criterion Nij and key airbag parameters are constructed. A multi-objective optimization on WIC and Nij is conducted by using NSGA-II. The results show that with a belt length of 0.205 m, a vent hole opening degree of 200%, an exhaust initial pressure of 1.15×105 Pa and an inter-seat distance of 0.65 m, WIC and Nij simultaneously get a small value, reducing by 60.75% and 60.94% respectively, achieving the goal of reducing the neck injury as much as possible while enhancing the overall protection effects of child occupant References | Related Articles | Metrics

Select

Anti-Rollover Control of Bus Based on Nonlinear Disturbance Estimation Shi Qiujun, Li Jing 2020, 42 (9):  1224-1231.  doi: 10.19562/j.chinasae.qcgc.2020.09.012 Abstract ( 150 )   PDF (2195KB) ( 238 )   Save In the anti-rollover control of bus, there are various unknown nonlinear disturbances and parameter perturbations in actual vehicle system modeling process, so it is difficult to establish an accurate vehicle model, and there is problem of big chattering in standard sliding mode control (SMC). The RBF-ADSMC (radial basis function-adaptive sliding mode control, RBF-ADSMC) algorithm is proposed in this paper. Firstly, the radial basis function (RBF) neural network controller is used to estimate various unknown disturbance items and parameter perturbation items in vehicle modeling process. Then, the RBF neural network is used to adaptively adjust the key parameters of the standard SMC. Finally, the electronically controlled pneumatic hardware in the loop test bench is built, and the control algorithm is verified on the hardware in the loop test bench. The test results show that the RBF-ADSMC algorithm has good control effect and can meet the bus rollover control requirements. Compared with the SMC algorithm, the RBF-ADSMC algorithm can reduce the roll angle and lateral acceleration of the bus and improve the anti-rollover control effect of the bus References | Related Articles | Metrics

Select

Comparative Analysis on Pedestrian Lower Limb Models FlexPLI, FlexPLIUBM and aPLI Hu Shuaishuai, Lü Xiaojiang, Wang Pengxiang, Wang Miao, Gu Pengyun , Zhang Pengju 2020, 42 (9):  1232-1239.  doi: 10.19562/j.chinasae.qcgc.2020.09.013 Abstract ( 173 )   PDF (3950KB) ( 267 )   Save European project SENIORS proposed a novel pedestrian lower limb model FlexPLIUBM, and JARI and JAMA in Japan proposed model aPLI. These two models show better biomechanical responses, including motion and force responses.A series of basic analyses are conducted on the currently used model FlexPLI and the above-mentioned two models, which may be applied in the future, in this paper. Firstly, some key parameters, including total mass, segment mass, position of mass center, structure, stiffness and dimension of lower limb models FlexPLI, FlexPLIUBM and aPLI are analyzed and compared. Then based on these key parameters, kinematics and mechanics analyses on crash process are carried out with the rules of and differences between evaluation indicators obtained. Finally, tests are performed to verify analyzed conclusions.All the comparative analyses and conclusions on pedestrian lower limb models FlexPLI, FlexPLIUBM and aPLI can provide guidance for the development of vehicle front bumper. References | Related Articles | Metrics

Select

Multi-performance Integrated Optimal Shift Schedule for Ramp Driving Lu Han, Yin Xiaofeng, Chen Kexu, Wu Xiaohua, Liang Yiming, Liu Yang 2020, 42 (9):  1240-1247.  doi: 10.19562/j.chinasae.qcgc.2020.09.014 Abstract ( 170 )   PDF (1624KB) ( 305 )   Save For ramp adaptation problem of shift schedule, the approach for multi-performance integrated optimal shift schedule (MPIOSS) for ramp driving of stepped automated transmission vehicle has been investigated. Firstly, the influence of ramp on shifting process is analyzed, and a method to determine the gear range according to the relation between the driving force under various throttle opening and slope resistance is put forward. Secondly, an integrated performance evaluation function is constructed by using the linear weighting method of ideal point and square sum to realize the overall optimization of vehicle power performance and fuel economy under the premise of reflecting the driver's performance expectation. Then, an optimization approach for MPIOSS for ramp driving based on genetic algorithm is proposed, which takes the integrated performance evaluation function as the optimization objective and considers the constraints of ramp driving. The proposed optimization approach is applied to formulate the MPIOSS shift schedules for a 5-speed automated manual transmission (AMT). Finally, the multi-performance integrated optimal shift schedules considering ramp and those ignoring ramp are verified via simulation under fixed throttle and fixed slope, fixed throttle and changing slope, changing throttle and fixed slope, and changing throttle and changing slope conditions. The results show that under the four kinds of ramp driving, the multi-performance integrated optimal ramp shift schedule can eliminate unexpected shift operation and improve the ride comfort. Under the modified HWFET-MTN driving cycle, the fuel economy of the vehicle with multi-performance integrated optimal ramp shift schedules is only slightly lower than that of the vehicle with conventional shift schedule, with a reduction of about 0.8%. Furthermore, power performance simulation under fixed slope shows the proposed MPIOSS for ramp driving can reflect the driver's performance demand intention References | Related Articles | Metrics

Select

Vehicle Classification Method in Multi-climates Based on Modified LSTM-AdaBoost Algorithm Li Da, Zhang Zhaosheng, Liu Peng, Wang Zhenpo, Dong Haotian 2020, 42 (9):  1248-1255.  doi: 10.19562/j.chinasae.qcgc.2020.09.015 Abstract ( 167 )   PDF (1407KB) ( 319 )   Save In view of the poor results of existing domestic and oversea vehicle classification schemes and relatively significant effects of climate on them, a multi-climate vehicle classification method based on modified LSTM-AdaBoost (long short-term memory neural network-Adaptive boosting) algorithm is proposed, and a “multi-layer grid method” is also put forward to accurately determine the hyperparameters of LSTM. Firstly, the geomagnetic vehicle detection system and vehicle classification method are established. Then the results of vehicle classification based on modified LSTM-AdaBoost are analyzed, and the classification accuracies of different vehicle classification methods and different climates are compared. The results show that compared with K-nearest neighbor and BP neural network algorithms for classification, the proposed method has higher accuracy with a highest classification accuracy of 92.2%. Among three climates of torrential rain, haze and fine day, the classification accuracy in torrential rain is lowest, but the difference is rather small, 3.9 percentage points at most References | Related Articles | Metrics

Select

Traffic Sign Recognition Based on Improved Cascade Convolution Neural Network Wang Hai, Wang Kuan, Cai Yingfeng, Liu Ze, Chen Long 2020, 42 (9):  1256-1262.  doi: 10.19562/j.chinasae.qcgc.2020.09.016 Abstract ( 289 )   PDF (2488KB) ( 318 )   Save The detection and recognition of traffic signs in automatic driving scene is very important. In order to improve the accuracy of traffic sign detection in the natural scene, this paper proposes an improved traffic sign recognition algorithm based on Cascade-RCNN. Firstly, the deep feature information of FPN module is fused into the shallow feature layer for the special task of small targets such as traffic signs. Secondly, the evaluation index IoU of the target detection task is improved by introducing in the direct evaluation index GIoU of the target detection task to guide the positioning task, which improves the detection accuracy. Finally, the algorithm is verified by experiments in GTSDB, a German traffic sign data set. When the network extraction is based on ResNet101 features, the mAP can reach 98.8%. The experimental results show that the proposed algorithm is effective and has superior engineering practical value References | Related Articles | Metrics

Select

Research on the Method of Navigating Autonomous Driving Vehicle Through Expressway Toll Region Xiong Ying, Mao Xuesong 2020, 42 (9):  1263-1269.  doi: 10.19562/j.chinasae.qcgc.2020.09.017 Abstract ( 173 )   PDF (1942KB) ( 285 )   Save In expressway toll region, the traditional method of path planning and intelligent decision-making cannot realize safe driving of autonomous driving vehicle due to randomness of surrounding vehicles behavior. To solve the problem, a structured control model, named as decision tree, is proposed, which controls the vehicle speed to make it drives through the toll region along the predefined path safely. Firstly, the scene parameters of the expressway toll region, the method for predefining the path between the toll window and of the ramp, and the vehicle model are given. On this basis, a structured control model of the decision tree is proposed to control the speed to realize safe driving of the vehicle. Meanwhile, the pure tracking method is adopted for navigating the vehicle driving along the predefined path. Finally, random road environment is constructed by computer simulation for verifying the safety of the decision tree control model, the continuity of vehicle speed and feasibility of acceleration and deceleration. The results show that autonomous driving vehicle can pass through the expressway toll region safely with vehicle speed controlled by the decision tree control model under the condition that surrounding vehicles obey traffic rules. In addition, the vehicle speed, acceleration and deceleration comply with the traffic rules, and are within the executable range References | Related Articles | Metrics

Select

Vertical Load Measurement of Automotive Intelligent Tire Huang Xiaojing, Zhang Feng, Zhang Shiwen, Zhengqi WU, Wei Sheng, Wang Feng 2020, 42 (9):  1270-1276.  doi: 10.19562/j.chinasae.qcgc.2020.09.018 Abstract ( 187 )   PDF (1596KB) ( 390 )   Save A scheme of the data acquisition of tire vertical load and the design of signal processing system is presented in this paper based on embedded platform. Specifically, MEMS sensors are embedded inside the tires to obtain their basic real-time data of radial acceleration, pressure and temperature, from which characteristic information is extracted to represent tire vertical load. Then bench test and real vehicle test are conducted respectively to verify the measurement method of vertical load of intelligent tire and characteristics extraction algorithm. The results show that the vertical load measurement method of intelligent tire is feasible,with its accuracy meeting the requirements of engineering application References | Related Articles | Metrics

Select

Control of Home Intelligent Vehicle Based on Udwadia-Kalaba Theory Zhao Yating, Sun Hao, Chen Xiaolong, Wu Qilin 2020, 42 (9):  1277-1283.  doi: 10.19562/j.chinasae.qcgc.2020.09.019 Abstract ( 148 )   PDF (1840KB) ( 216 )   Save For the modeling and control of complex electromechanical system with nonholonomic constraints such as intelligent vehicles, a classification modeling and control method based on the Udwadia-Kalaba theory and the Udwadia control framework is proposed. The method consists of two parts, i.e. passive constraint and servo constraint. The former solves the problem of nonholonomic constraints in the system model based on Udwadia-Kalaba theory. The latter solves the problem of approximate trajectory tracking control based on Udwadia control. The method is applied to the home intelligent vehicle system, and an accurate dynamic model and control system is established. The effectiveness of the classification modeling and control method is verified by the test and simulation analysis of the motor References | Related Articles | Metrics

Select

Motor Fault Diagnosis and Failure Control for Distributed Four-wheel Drive Electric Vehicles Zhu Shaopeng, Fu Qitao, Huang Xiaoyan, Wang Zhiwei, Yang Xinghao 2020, 42 (9):  1284-1291.  doi: 10.19562/j.chinasae.qcgc.2020.09.020 Abstract ( 221 )   PDF (2059KB) ( 542 )   Save In this paper, a fault diagnosis and failure control strategy based on motor failure control gain is proposed for the soft and hard faults of motor in distributed four-wheel drive electric vehicles. A fault diagnosis module and a failure mode judgment and driving force redistribution module are designed, which are combined with vehicle direct yaw moment control to achieve the secondary distribution control of driving force based on fault diagnosis. By MATLAB/Simulink and CarSim joint simulation and RCP real vehicle test, the effectiveness of the fault diagnosis and failure control strategy proposed is verified. The proposed control strategy can make sufficient use of the motor redundancy characteristics of distributed four-wheel drive electric vehicle to ensure the safe driving of vehicle under the various common failure conditions of motor References | Related Articles | Metrics