Loading...
Welcome to Journal of Automotive Safety and Energy,

Current Issue

  • 2024, Vol. 15 No. 4 Published on:31 August 2024 Previous issue    Next issue
    Review, Progress and Prospects
    Overview of the combustion of ammonia-hydrogen internal combustion engines
    WANG Zhi, QI Yunliang, CHEN Qingchu, LI Jun
    2024, 15(4):  443-466.  doi:10.3969/j.issn.1674-8484.2024.04.001
    Abstract ( 391 )   HTML ( 821)   PDF (11029KB) ( 158 )  

    With the introduction of Chinese goals of “carbon peak” and “carbon neutrality”, the low-carbon and zero-carbon transition of internal combustion engines is imperative. Ammonia, as a zero-carbon fuel and a high-energy-density carrier for hydrogen, is a promising alternative fuel for achieving carbon neutrality in the near to mid-term. Developing ammonia-hydrogen combustion technology for high-power, zero-carbon internal combustion engines is of significant importance for global climate governance. This paper analyzes the potential of ammonia as a future green energy source and its practical applications in internal combustion engines. It reviews the latest advancements in ammonia-hydrogen engine combustion from the aspects of combustion modes, reaction kinetics of ammonia-hydrogen fuel, and fuel supply methods, comparing three combustion modes (spark ignition/homogeneous compression ignition/jet ignition), two ammonia supply methods (gaseous port injection/liquid ammonia direct injection), and two jet ignition methods (active/passive). A promising technology of ammonia-hydrogen synergy combustion based on online ammonia cracking to produce hydrogen from a single liquid ammonia fuel tank is proposed and discussed. Key technical and scientific issues to be addressed in spray, combustion, and nitrogen-based emission control are also pointed out. Research indicates that using a small amount of hydrogen (less than 3%) to ignite ammonia-air mixtures in ammonia-hydrogen engines can achieve stable combustion and high thermal efficiency while extending the lean limit. Ammonia-hydrogen synergy zero-carbon high-power internal combustion engines, as efficient and reliable application carriers for ammonia fuel, have broad application potential and value in heavy-duty vehicles, construction machinery, ocean-going vessels, and power generation. The development of ammonia-hydrogen engines can advance fundamental combustion theory and revitalize China's internal combustion engine industry.

    Figures and Tables | References | Related Articles | Metrics
    Automotive Safety
    Fault-tolerant control strategy for active steering of distributed driving electric vehicles under in-wheel motor failure
    LU Yanbo, LIANG Jinhao, YIN Guodong, FENG Jiwei, WANG Fanxun
    2024, 15(4):  467-476.  doi:10.3969/j.issn.1674-8484.2024.04.002
    Abstract ( 176 )   HTML ( 134)   PDF (6444KB) ( 299 )  

    An active steering fault-tolerant control strategy was proposed based on an event triggering mechanism to improve the stability of distributed driving electric vehicles when a single wheel or single side in-wheel motors failed. A fault state observer based on linear matrix inequalities was designed, and a fault-tolerant switching mechanism based on residual evaluation function was established. The distributed-model predictive cooperative control strategy was adopted based on Pareto optimization theory to realize the optimal distribution mechanism of fault-tolerant control for vehicle chassis active steering system and yaw moment control system. Simulation experiments were carried out on the hardware-in-loop platform based on the integration of the software Matlab/Simulink and the software CarSim. The results show that the yaw rate displacement tracking error is reduced by 14.9%, and the lateral displacement tracking error is reduced by 28.2%, in the case of constant deviation or variable gain fault of in-wheel motor under double lane change condition with high/low road adhesion coefficients. Therefore, the fault tolerance and the stability of distributed driving electric vehicles are realized under the failure of wheel hub motor.

    Figures and Tables | References | Related Articles | Metrics
    Effect of the parameter disturbance on a vehicle brake control and protection method proposed earlier
    ZOU Tiefang, JIANG Die, ZHOU Jing, YUAN Xiangting
    2024, 15(4):  477-483.  doi:10.3969/j.issn.1674-8484.2024.04.003
    Abstract ( 108 )   HTML ( 67)   PDF (1206KB) ( 74 )  

    Investigated the effectiveness of vehicle brake control protection methods proposed earlier by the author under parametric perturbations. MADYMO virtual simulation software was used to simulate 4 types of typical locomotive, and 3 parameters for analyzing the influence of disturbance on the braking control and protection method; The 3 parameters including the first contact time between human head and vehicle body, the braking system coordination time, and the road adhesion factor. The results show that the first contact time between the human head and the car body is 80~100 ms, which has the most significant impact on the reduction of weighted injury cost (WIC), but the parameter sensitivity ranking is different in different vehicle types. When the disturbance of road adhesion factor is 0.45~0.80, the damage of human-vehicle collision does not increase, but the decrease of WIC caused by ground is greater than zero, which can effectively reduce the damage of human-ground collision. When the disturbance interval of the coordination time parameter of the braking system is 80~350 ms, the braking control and protection method is still effective.

    Figures and Tables | References | Related Articles | Metrics
    Simulation experiments for the passenger injuries with different sitting positions in autonomous vehicle collisions
    WU Hequan, LI Yihui, LIU Jin, ZHANG Shenao
    2024, 15(4):  484-491.  doi:10.3969/j.issn.1674-8484.2024.04.004
    Abstract ( 173 )   HTML ( 73)   PDF (2733KB) ( 59 )  

    Frontal crash simulation experiments were made with a comparative analysis of the kinematics and loads of its occupants, to investigate the crash safety of occupants in autonomous vehicles with different sitting positions such as representative occupant attitudes (including variations in head, upper limbs and torso). The results show that in the frontal collision simulation experiment, the comprehensive injury risk of leaning passengers is higher than that of other postures. Compared to the conversational posture, the leaning posture resulted in a 45.5% increase in chest compression, more rib fractures, and a 48.8% increase in visceral organ strain. The left leaning posture significantly reduces the risk of injury, with a 49.1% decrease in head injury risk and a 14.4% decrease in chest injury risk. Therefore, the occupant injury can be reduced by adjusting the occupant posture before the vehicle collision, so as to improve the collision safety of autonomous vehicle.

    Figures and Tables | References | Related Articles | Metrics
    Research on aerodynamic characteristics of driving platooning in wind and rain coupled environment
    XU Jianmin, YANG Wei, GONG Xiaoyan, WU Song, LI Luonan
    2024, 15(4):  492-502.  doi:10.3969/j.issn.1674-8484.2024.04.005
    Abstract ( 103 )   HTML ( 64)   PDF (3625KB) ( 65 )  

    In order to explore the aerodynamic load characteristics and flow field characteristics of the driving queue under wind and rain environment, this paper established a driving queue model under wind and rain coupling environment based on Euler-Lagrange method, considering the interaction between raindrops and air flow, and verified the accuracy of the model by comparing the test data. Then the effects of rainfall, lateral wind speed and vehicle spacing on the aerodynamic load characteristics and flow field characteristics of vehicles in the queue were studied. The results show that the force and torque on vehicles in the driving queue increase continuously with the increase of rainfall and lateral wind speed. When the rainfall is 50~250 mm/h, the resistance of the first vehicle in the queue increases from 27.21 to 28.49 N. When the lateral wind is 4.4~15.5 m/s, the lateral force of the first vehicle increases from 101 to 554.3 N. With the increase of queue spacing, the aerodynamic resistance of vehicles increases at first and then decreases, the lateral force and roll moment increase continuously, and the yaw moment decreases continuously. When the queue spacing is 0.25 to 0.75 times, the lateral force of the tail car (the fifth car) increases from 137.6 to 174.74 N, and the yaw moment decreases from 39.29 to 12.47 Nm.

    Figures and Tables | References | Related Articles | Metrics
    Typical accident scene extraction and accident factor association rules analysis for truck-two-wheeler collisions
    NIE Jin, ZHANG Xiang, ZHANG Yue, WANG Bingyu, YI Xiangxian, ZHOU Wu
    2024, 15(4):  503-510.  doi:10.3969/j.issn.1674-8484.2024.04.006
    Abstract ( 125 )   HTML ( 68)   PDF (1069KB) ( 98 )  

    Typical accident scenarios of truck-two-wheeler collisions were extracted based on video information. The correlation among accident factors within these scenarios was explored using association rules analysis to elucidate accident characteristics that were not typically described by typical accidents. Accident information was extracted from 210 cases of truck-two-wheeler accidents with accompanying video information sourced from the internet. These cases were then subjected to K-modes clustering to obtain typical accident scenarios. Subsequently, association rule mining was employed to analyze the degree of association among accident factors within each accident scenario. The results show that truck-two-wheeler accidents can be categorized into four typical scenarios, namely two types of intersection accidents, one straight road accident, and one T-junction accident. In intersection accidents, there is a high association between accident features and the presence of visual obstacle. The accident featuring a “truck turning left” will cause high accident risk. In straight road accidents, there is an association between truck braking for avoidance and the presence of traffic signals on the road. In T-junction accidents, there is an association between medium-sized truck steering for avoidance and two-wheeler riders being crushed. These research findings can provide reference to safety measures and safety testing scenarios for two-wheeler riders.

    Figures and Tables | References | Related Articles | Metrics
    Diagnosis of residual bidirectional LSTM automotive motor bearings with attention mechanism
    JIANG Jian, WANG Ping
    2024, 15(4):  511-519.  doi:10.3969/j.issn.1674-8484.2024.04.007
    Abstract ( 90 )   HTML ( 5)   PDF (1879KB) ( 446 )  

    In order to ensure the safe driving of vehicles and accurately diagnose and monitor motor bearing faults, this paper proposed an automotive motor bearing fault diagnosis method based on residual bidirectional Long Short-Term Memory (LSTM) network with attention mechanism. The feature extraction module combined LSTM groups that move in both forward and backward directions to fully perceive the fault features of automotive motor bearings. The signal diagnosis module adopted a residual bidirectional LSTM architecture and combined the local enhanced attention mechanism to optimize the weights and obtain the hidden state quantity. Global average pooling (GAP) method and SoftMax model are used in fault classification module to effectively detect faults. The results show that the detection accuracy of this method for automotive motor bearing fault detection reaches 93.1%. Under the condition of only 30 training samples, the accuracy reaches 66.3%. When the signal-to-noise ratio of the test set decreases from 10 dB to 2 dB, the accuracy only drops by 8.5%. Therefore, the proposed method has higher accuracy and stronger robustness.

    Figures and Tables | References | Related Articles | Metrics
    Automotive Energy Efficiency and Environment Protection
    Microscopic characteristics of liquid ammonia spray at different ambient pressures and ambient temperatures
    LIU Xiao, YAO Xiaoxin, WANG Ze, TANG Chenglong
    2024, 15(4):  520-525.  doi:10.3969/j.issn.1674-8484.2024.04.008
    Abstract ( 85 )   HTML ( 3)   PDF (1379KB) ( 332 )  

    Microscopic characteristics were studied for liquid ammonia spray experimentally to realize low-carbon and zero-carbon goals with liquid ammonia as fuel of internal combustion engine. The liquid ammonia spray at different ambient pressures (100, 200, 400, 600 kPa) and ambient temperatures (293, 313, 333 K) were experimentally investigated by using the Particle/Droplet Image Analysis (PDIA) method and using a single hole injector with the injection pressure of 1 MPa. Some microscopic parameters were obtained, such as the particle density function, cumulative distribution function, the Sault Mean Diameter (SMD), and the droplet number density. The results show that the liquid ammonia spray has a good atomization effect and have large amount of small particles under low ambient pressure; The particle density function and cumulative distribution function move to a larger droplet sizes, the number of larger droplets increase and the droplet density decrease; The particle density function, cumulative distribution function and SMD hardly changed with the ambient temperature, but the liquid ammonia spray evaporation is enhanced at high ambient temperature.

    Figures and Tables | References | Related Articles | Metrics
    Influence of anisotropic conductive expanded graphite bipolar plates on performance of proton exchange membrane fuel cell
    DING Yujie, GAN Quanquan, SHAO Yangbin, XU Liangfei, LI Jianqiu, OUYANG Minggao
    2024, 15(4):  526-535.  doi:10.3969/j.issn.1674-8484.2024.04.009
    Abstract ( 94 )   HTML ( 3)   PDF (1750KB) ( 37 )  

    Established a three-dimensional, two-phase, non-isothermal fuel cell single channel model to investigate the influence of anisotropic structure of Expanded Graphite Bipolar Plates (EGBPs) on thermal management and the performances of proton exchange membrane fuel cell (PEMFC). The distribution of temperature, the current density and the membrane water content under four main material orientations were compared. The coupling effect between EGBPs heat transfer characteristics and the outperformance were revealed. The results show that increasing the conductivity along proton-transport direction kz from the 5 W·m-1·K-1 for a traditional structure to the 280 W·m-1·K-1, enhances the voltage by 22 mV under 2.2 A cm-2. The conductivity along gas-flow direction ky plays an important role in heat dissipation. Increasing ky and kz to 280 W·m-1·K-1, or accomplishing isotropic conductivity (kx = ky = kz = 20 W·m-1·K-1), reduces the Membrane-Electrode-Assembly (MEA) core-area peak-temperature by 2 ℃. Therefore, enhancing the gas-flow and proton-transport direction conductivities and accomplishing isotropic structure are one of the future main development goal of expanded graphite bipolar plates.

    Figures and Tables | References | Related Articles | Metrics
    Chemical mechanism and N2O generation analysis of ammonia-methanol combustion in low proportions of methanol blends
    JIN Guangjie, LU Mingfei, LIU Xing, LONG Wuqiang, WANG Peng
    2024, 15(4):  536-544.  doi:10.3969/j.issn.1674-8484.2024.04.010
    Abstract ( 134 )   HTML ( 6)   PDF (4621KB) ( 347 )  

    In order to investigate the benefits of ammonia fuel in reducing greenhouse gas (GHG) emissions, and overcome the shortcomings of high ammonia ignition energy and slow laminar flame speed, a chemical reactor network (CRN) model of the combustion of methanol-ammonia mixtures was developed using Chemkin to analyze the component evolution, product generation rate and sensitivity of low proportion methanol (0~10%) and ammonia combustion. Chemical reaction kinetics was used to analyze the N2O generation/reduction mechanism. The results show that pure ammonia combustion at equivalence ratios of 1.1-1.2 can concurrently decrease emissions of NO, unburned NH3, and N2O, resulting in a reduction of over 70% in GHG emissions compared to pure methanol combustion. Ammonia combustion remains beneficial for GHG emission control when the equivalence ratio exceeds 0.9. Methanol addition effectively controls N2O generation, further mitigating GHG emissions. Under rich combustion conditions, methanol addition promotes H generation, leading to increased N2O consumption. However, under lean combustion conditions, methanol enhances the suppression of NH production, resulting in reduced N2O generation. Moreover, increasing pressure and temperature can effectively reduce GHG emissions, although the mechanisms vary.

    Figures and Tables | References | Related Articles | Metrics
    Performance analysis and structural parameter optimization of baffle channel in PEM fuel cell based on 3D model
    ZHANG Luo, PU Dongyi, HU Song, FAN Zhijun, CHEN Dongfang, XU Xiaoming
    2024, 15(4):  545-552.  doi:10.3969/j.issn.1674-8484.2024.04.011
    Abstract ( 179 )   HTML ( 3)   PDF (1304KB) ( 66 )  

    Against the backdrop of rapid advancements in hydrogen energy technology, the flow channel design of proton exchange membrane fuel cells (PEMFCs), as a key application, has been subjected to intensive study. This research developed a full-scale, single flow channel model of a PEMFC based on actual dimensions and incorporated baffle structures of varying shapes within the model to investigate, through simulation analysis, the impact of the geometric parameters of these baffles on cell performance. The results show that the presence of baffle structures significantly enhances power output, with increases in current density ranging from 10% to 31.25% and power boosts between 3% and 9.2%. Notably, a positive correlation exists between gas flow velocity and power output with the height of the baffles, with the performance of rectangular baffles at a height of 0.2 mm being optimal. This demonstrates that an appropriate baffle height can effectively improve PEMFC efficiency by enhancing mass transfer effects and optimizing the distribution of the working medium within the fuel cell, ameliorating water flooding in the latter half of the flow channel caused by water accumulation, thereby further augmenting cell power. Compared to the length of baffles, their height emerges as the primary direction for optimization. The study provides theoretical support for the optimization of full-scale PEMFC flow channels and offers guidance for enhancing the efficiency of hydrogen energy utilization.

    Figures and Tables | References | Related Articles | Metrics
    Testing and statistics of the heavy-duty vehicle chassis dynamometer fuel consumption measurement
    TONG Chang, REN Shuojin, LI Tengteng, LIU Dong, LIU Lin, ZHANG Peng, LIU Zhiwei, JING Xiaojun, YANG Zhengjun
    2024, 15(4):  553-560.  doi:10.3969/j.issn.1674-8484.2024.04.012
    Abstract ( 121 )   HTML ( 6)   PDF (1324KB) ( 146 )  

    The impact of vehicle light-mass on 100 kmt volume fuel consumption were investigated to implement the strategy of China Ⅳ according to the current fuel consumption standard for heavy-duty vehicles in China mainland. The chassis dynamometer certification data of 1 408 heavy-duty vehicles meeting China Ⅲ fuel consumption standard were statistically analyzed. The results show that the largest deviation between the average fuel consumption and the China Ⅳ limits is from tractors, which is about 8.0%. The deviation between the average fuel consumption and the China Ⅳ limits of trucks is around 3.0%. The urban condition is the main strategy of fuel consumption optimization for trucks with the Gross Vehicle Mass (GVM) exceeding 20 t. The fuel volume consumption per 100 kmt is negatively correlated with the vehicle payload. With the increasing of GVM, the fuel volume consumption per 100 kmt of trucks and dump trucks shows a downward trend, while the fuel volume consumption per 100 kmt of tractors is basically unchanged, which is nearly 1.0 L. Therefore, using vehicle light-mass technology is beneficial for reducing 100 kmt volume fuel consumption.

    Figures and Tables | References | Related Articles | Metrics
    Research on fuel consumption characteristics of light vehicles based on chassis dynamometer test
    LAN Yunfei, GUO Shuwen, YANG Bohan, ZENG Lewei, LUO Mingjing, WANG Fengbin, LI Xintong, ZHENG Xuan
    2024, 15(4):  561-568.  doi:10.3969/j.issn.1674-8484.2024.04.013
    Abstract ( 120 )   HTML ( 2)   PDF (4136KB) ( 157 )  

    A detailed analysis was conducted on the fuel consumption measurement values of different light-duty gasoline vehicles to explore the factors affecting fuel consumption of light gasoline vehicles and provide experimental data support for the development of real-time fuel consumption data correction methods for on-board diagnostic systems (OBD), thereby promoting refined control and management of motor vehicle fuel consumption. The light-duty gasoline vehicles (LDGVs), complying with emission standards phases V and VIb (China V and VIb), were tested under both cold- and hot-start conditions using the Worldwide Light-duty Test Cycle (WLTC). Transient operating parameters of each engine and each entire vehicle were read by an OBD system during the testing. Subsequently, the fuel consumption of the tested LDGVs at different starting temperatures and speeds were calculated using the carbon balance method and the OBD air fuel ratio method. The effects of temperature and speed on fuel consumption of different LDGVs were also analyzed. The interpolation method was applied to compensate for the data loss during the OBD parameter reading process, enhancing the linear correlation coefficients in instantaneous fuel consumption between the two calculation methods. Furthermore, the causes of errors in OBD fuel consumption calculation were investigated. The results showed that the fuel consumption of China VIb LDGVs is less affected by the starting temperature than China V, with cold start fuel consumption only 2.02%~2.27% higher than hot start. The LDGVs equipped with gasoline particulate filter (GPF) have a fuel consumption increase of approximately 0.48 L per 100 kilometers compared to vehicles without GPFs, and the increase in fuel consumption caused by GPF is related to the driving speed. The repeated start stop of the engine in the low-speed range will result in fuel consumption reaching 1.46 times that of the entire stage. Uncertainties of the measured fuel consumption in OBD mainly occur in the deceleration stage of the vehicle.

    Figures and Tables | References | Related Articles | Metrics
    Adaptive optimization control strategy for PEM fuel cell tram considering durability
    GAO Fengyang, LIU Jia, HAN Guopeng, QI Fengxu, LIU Qingyin
    2024, 15(4):  569-578.  doi:10.3969/j.issn.1674-8484.2024.04.014
    Abstract ( 108 )   HTML ( 5)   PDF (1898KB) ( 80 )  

    Aiming at the life decay of proton exchange membrane (PEM) fuel cells in fuel cell/Li battery/supercapacitor hybrid energy storage streetcars during dynamic loading, start-stop cycle, idling cycle, and high-power operation, an energy management strategy combining the Pontryagin's minimum principle(PMP) of very small value and durability was proposed. The start-stop control strategy controlled the start and stop of the fuel cell, effectively reducing the number of fuel cell starts. The joint cost function was constructed using the overall hydrogen consumption of the energy storage system as the economic cost and the fuel cell performance degradation index as the durability cost. The online adaptation of the covariance variable with the real-time change of charge state was realized under satisfying the desired conditions. The proposed strategy was simulated and compared with the traditional minimal value strategy and state machine strategy. The results show that the proposed strategy under the urban cycling condition reduces the peak current by 33.2% and hydrogen consumption by 12.50% compared with the traditional Pontryagin strategy fuel cell; the proposed strategy under the suburban cycling condition reduces the peak current by 21.88% and hydrogen consumption by 40.39% compared with the traditional Pontryagin strategy. The proposed management strategy has good adaptability under different working conditions, solves the shortcomings of traditional PMP that can only be applied offline, controls the number of start-stop at a low level, and has the ability to extend the life of the fuel cell.

    Figures and Tables | References | Related Articles | Metrics
    Intelligent Driving and Intelligent Transportation
    An obstacle avoidance path planning algorithm for autonomous buses based on tracking error observation and target measurement error observation
    LI Yulong, XIE Hui, SONG Kang
    2024, 15(4):  579-590.  doi:10.3969/j.issn.1674-8484.2024.04.015
    Abstract ( 112 )   HTML ( 3)   PDF (2637KB) ( 69 )  

    An obstacle avoidance path planning algorithm based on tracking error observation and target measurement error observation was proposed to enhance the obstacle avoidance safety of autonomous buses, aiming at the risk of collision caused by the uncertainty of tracking control and target measurement. The mean-variance model was utilized to describe the tracking error, and the Gaussian process regression was introduced to quantify target measurement error. A probability distribution model for the lateral spatial position was established, providing theoretical support for solving the safe lateral offset value. An obstacle avoidance path planning was implemented through the Bézier path model, and a comprehensive evaluation system was established to assess the performance of the obstacle avoidance planning. In the simulation environment, the influence of uncertain factors such as tracking error and measurement error on obstacle avoidance quality was explored, and the real vehicle data simulation verification was carried out on the test road of unmanned intelligent bus in Tianjin University, and the effect of the three algorithms was evaluated. The results show that the obstacle avoidance safety is improved by 17.61% in actual driving, and it has good robustness. The proposed average time of the algorithm is 7.43 ms, which has good real-time performance.

    Figures and Tables | References | Related Articles | Metrics
    Semantic segmentation of real-time LiDAR point clouds based on multi-scale self-attention
    ZHANG Chen, LIU Chang, ZHAO Jin, WANG Guangwei, XU Qing
    2024, 15(4):  591-601.  doi:10.3969/j.issn.1674-8484.2024.04.016
    Abstract ( 123 )   HTML ( 5)   PDF (2891KB) ( 99 )  

    A real-time point cloud semantic segmentation method was proposed for mobile robot platforms through digital experiments, to enhance segmentation accuracy within the constraints of in-vehicle computing resources. The approach used a projection-based LiDAR technique, projecting the 3-D point cloud onto a spherical image and applying 2-D convolution. The approach integrated the multi-head self-attention (MHSA) mechanism, adapting the Transformer, a software semantic segmentation, architecture into convolution operations to build a multi-scale self-attention (MSSA) framework. The results show that on the NVIDIA JETSON AGX Xavier computing platform, the proposed method achieves a high segmentation accuracy with the mean ratio of Intersection to Union (mIoU) being 63.9%, and a fast detection speed of 41 frame/s, compared to state-of-the-art methods like the CENet, the FIDNet, and the PolarNet, therefore, demonstrating the effectiveness of the mobile robot platforms.

    Figures and Tables | References | Related Articles | Metrics
    Prediction of driver visual distraction duration under a human-machine interaction mode with full-touch-screen
    WANG Chang, NIU Jin, WANG Yifei, ZHANG Yali, MA Wanliang
    2024, 15(4):  602-609.  doi:10.3969/j.issn.1674-8484.2024.04.017
    Abstract ( 131 )   HTML ( 6)   PDF (1671KB) ( 70 )  

    A visual distraction duration prediction model considering the driver's distraction style was established to accurately predict the driver's visual distraction duration in the full touch-screen human-computer interaction mode. Based on the multi-speed visual distraction data collected from the real car road test, a set of basic operation units suitable for full touch screen vehicle-mounted devices was constructed based on the Keystroke Level Model (KLM), and a Random Forest Model (RF) for visual distraction duration prediction with the number of each basic operation units and vehicle speed as input features was established. The Self-Organizing Map (SOM) Algorithm was used to cluster drivers into three types of distraction styles: Cautious, normal and aggressive to characterize the differences in drivers' visual distraction characteristics, and the parameters were added to the original model as input features to realize optimization. The results show that the optimized model has the best prediction accuracy. The mean square error, mean absolute error and determination coefficient R2 of the test set are 2.414 8 s2, 1.037 1 s and 0.934 5, respectively, which are 30.52%, 11.8% and 3.18% higher than those of the original model. The model performance is significantly better than the linear regression model and XGBoost model. The results can be used to assist in-vehicle driver assistance systems in timely warning or implementing intervention to reduce the risk of rear-end collision caused by distraction, and provide guidance for the design of interactive interfaces.

    Figures and Tables | References | Related Articles | Metrics