Classification and Identification of Risky Driving Behavior Based on Hybrid Strategy Improved ASO-LSSVM

doi:10.12141/j.issn.1000-565X.230753

Abstract

Abstract:

This paper aims to solve the problem of slow convergence rate and large error of existing intelligent algorithms in the process of optimizing support vector machine to identify risky driving behavior. Firstly, Tent mapping was used to replace the random setting of population initialization of ASO algorithm to increase the diversity and quality of atomic population. Secondly, the hybrid mechanism of dimension-by-dimension pinhole imaging reverse learning and Cauchy mutation was used to improve the diversity of preferred positions of atomic individuals and overcome the problem that ASO algorithm is easy to fall into local optimum and premature convergence. Finally, the adaptive variable spiral search strategy was introduced to improve the atomic individual position update process, so as to improve the global search ability of ASO algorithm, realize the effective balance between global search and local development, and alleviate the problem that ASO algorithm is easy to fall into local optimum and lack of convergence accuracy. Taking the vehicle trajectory data of the exit ramp of Shanghai North Cross Channel as the input, the study used the hybrid strategy to improve the ASO algorithm so as to optimize the LSSVM parameters. And it constructed the classification and identification model of the risk driving behavior of the expressway exit ramp based on IASO-LSSVM. Numerical simulation results show that the average value, standard deviation, best fitness and worst fitness of the numerical simulation results of the IASO algorithm in 12 benchmark test functions are closer to the best optimization value. Compared with ASO-LSSVM and LSSVM, the accuracy, precision, recall and F₁ value of risk driving behavior classification and identification results of IASO-LSSVM model increased by 11.5~24.5, 14.1~29.0,15.1~28.6, 14.7~31.2 percentage points respectively, and the error range was the smallest in different types of risky driving behavior identification results. The accuracy and convergence rate of IASO algorithm are better than those of ASO algorithm, and the IASO-LSSVM model can be used for accurate identification of different types of risk driving behavior, which can provide data support and theoretical basis for reasonable discrimination of vehicle driving trajectory state and formulation of early warning and prevention measures of risk driving behavior.

Key words: urban transport, expressway exit ramp, risk driving behavior classification and identification, atom search optimization, hybrid strategy, LSSVM

CLC Number:

U 491.1

HE Qingling, PEI Yulong, DONG Chuntong, et al. Classification and Identification of Risky Driving Behavior Based on Hybrid Strategy Improved ASO-LSSVM[J]. Journal of South China University of Technology(Natural Science Edition), 2024, 52(9): 131-141.

Figures/Tables 9

Fig.1

Fig.2

Table 1

Simulation results of benchmarking functions"

测试函数	指标	IASO	ASO	GWO	PSO	GA
f₁	Best	2.9×10^-268	6.6×10^-30	2.0×10^-18	1.5×10²	3.4
	Worst	4.1×10^-180	8.9×10^-27	1.1×10^-16	5.9×10²	5.5×10
	Average	1.4×10^-181	1.1×10^-27	3.2×10^-17	3.4×10²	1.7×10
	Standard	0	2.0×10^-27	3.4×10^-17	1.0×10²	1.2×10
f₂	Best	1.9×10^-274	1.1×10³	3.2×10^-9	2.0×10³	8.7×10²
	Worst	4.7×10^-177	3.9×10³	1.5×10^-4	3.3×10⁴	5.7×10³
	Average	1.6×10^-178	2.3×10³	9.9×10^-6	1.0×10⁴	2.3×10³
	Standard	0	8.0×10²	2.9×10^-5	6.7×10³	1.0×10³
f₃	Best	2.8×10	2.6×10	2.4×10	1.3×10³	2.6×10²
	Worst	2.9×10	2.9×10	4.5×10²	4.6×10⁴	5.7×10³
	Average	2.8×10	2.7×10	6.3×10	1.5×10⁴	1.2×10³
	Standard	1.3×10	6.7×10^-1	8.2×10	1.2×10⁴	1.1×10³
f₄	Best	8.3×10^-6	3.6×10^-2	5.7×10^-4	4.3×10^-2	1.3×10^-1
	Worst	1.4×10^-3	1.6×10^-1	7.3×10^-3	2.7	1.1
	Average	3.8×10^-4	8.1×10^-2	2.2×10^-3	3.2×10^-1	5.8×10^-1
	Standard	4.0×10^-4	3.2×10^-2	1.4×10^-3	4.9×10^-1	3.1×10^-1
f₅	Best	-1.3×10⁴	-8.3×10³	-7.4×10³	-9.3×10³	-1.3×10⁴
	Worst	-4.4×10³	-5.9×10³	-3.3×10³	-5.8×10³	-5.7×10³
	Average	-1.0×10⁴	-6.9×10³	-6.1×10³	-7.5×10³	-7.2×10³
	Standard	2.3×10³	4.9×10²	8.2×10²	9.6×10²	2.6×10³
f₆	Best	0	2.4×10¹	5.7×10^-14	1.4×10²	1.8×10
	Worst	0	5.6×10¹	1.1×10	2.6×10²	5.8×10
	Average	0	3.8×10¹	2.8	1.9×10²	3.3×10
	Standard	0	8.6	3.8	3.0×10	9.7×10
f₇	Best	4.4×10^-16	7.5×10^-14	6.3×10^-10	3.7	1.0×10
	Worst	4.4×10^-16	1.6×10^-13	1.5	8.6	2.0×10
	Average	4.4×10^-16	1.0×10^-13	1.6×10^-1	5.6	2.0×10
	Standard	0	2.2×10^-14	4.2×10^-1	1.0	1.8
f₈	Best	0	0	0	1.7	1.1
	Worst	0	6.6×10^-2	2.7×10^-2	9.1	7.2
	Average	0	5.5×10^-3	3.2×10^-3	4.2	2.6
	Standard	0	1.3×10^-2	7.0×10^-3	1.9	1.1
f₉	Best	3.1×10^-4	6.1×10^-4	3.1×10^-4	8.0×10^-4	3.1×10^-4
	Worst	4.8×10^-4	5.5×10^-3	2.0×10^-2	2.5×10^-2	2.0×10^-2
	Average	3.5×10^-4	1.1×10^-3	3.9×10^-3	9.0×10^-3	1.2×10^-3
	Standard	4.8×10^-5	8.6×10^-4	7.5×10^-3	9.2×10^-3	3.6×10^-3
f₁₀	Best	-3.3	-3.3	-3.3	-3.3	-3.3
	Worst	-3.3	-3.2	-3.1	-1.6	-3.0
	Average	-3.3	-3.3	-3.2	-3.1	-3.1
	Standard	1.4×10^-15	6.1×10^-2	8.1×10^-2	3.2×10^-1	8.4×10^-2
f₁₁	Best	-1.0×10	-1.0×10	-1.0×10	-1.0×10	-6.2
	Worst	-5.1	-2.7	-2.7	-2.6	-3.9
	Average	-9.1	-8.3	-9.1	-8.1	-4.8
	Standard	2.1	3.2	2.3	2.7	5.5×10^-1
f₁₂	Best	-1.0×10	-1.0×10	-1.0×10	-1.0×10	-1.0×10
	Worst	-4.8	-3.7	-5.1	-5.1	-4.1
	Average	-1.0×10	-8.8	-1.0×10	-1.0×10	-5.5
	Standard	1.0	2.5	9.6×10^-1	1.3	1.7

Table 1

Fig.3

Table 2

Table 3

Fig.4

Fig.5

Table 4

References 18

1	TUBTIMHIN S， LAOHASIRIWONG W， PITAKSANURAT S，et al ．Clustering of road traffic injuries during the 7-day Songkran Holiday，Thailand：a spatial analysis［J］．Kathmandu University Medical Journal，2019，67（3）：184-189．
2	陈永恒，熊帅，白乔文，等．快速路入口匝道车辆汇入位置的分布特征分析［J］．华南理工大学学报（自然科学版），2018，46（11）：125-131．
	CHEN Yongheng， XIONG Shuai， BAI Qiaowen，et al ．Analysis of the distribution characteristics of on-ramp vehicle merging positions on expressway［J］．Journal of South China University of Technology（Natural Science Edition），2018，46（11）：125-131．
3	郭淼，赵晓华，姚莹，等．基于驾驶行为和交通运行状态的事故风险研究［J］．华南理工大学学报（自然科学版），2022，50（9）：29-38．
	GUO Miao， ZHAO Xiaohua， YAO Ying，et al ．Accident risk research based on driving behavior and traffic operation status［J］．Journal of South China University of Technology （Natural Science Edition），2022，50（9）：29-38．
4	CHEN Y， WANG K， LU J J ．Feature selection for driving style and skill clustering using naturalistic driving data and driving behavior questionnaire［J］．Accident Analysis & Prevention，2023，185（6）：1-16．
5	MA L， QU S， SONG L，et al ．Human-like car-following modeling based on online driving style recognition［J］．Electronic Research Archive，2023，31（6）：3264-3290．
6	PENG J， SHAO Y ．Intelligent method for identifying driving risk based on V2V multisource big data［J］．Complexity，2018（1）：1-9．
7	ZHU S， LI C， FANG K，et al ．An optimized algorithm for dangerous driving behavior identification based on unbalanced data［J］．Electronics，2022，11（10）：57-72．
8	JAHANGIRI A， BERARDI V J， MACHIANI S G ．Application of real field connected vehicle data for aggressive driving identification on horizontal curves［J］．IEEE Transactions on Intelligent Transportation Systems，2017，19（7）：2316-2324．
9	WEI L， LIANG L， LEI T，et al ．On-board unit （OBU）-supported longitudinal driving behavior monitoring using machine learning approaches［J］．Sensors，2023，23（15）：67-75．
10	SHANGGUAN Q， FU T， WANG J，et al ．A proactive lane-changing risk prediction framework considering driving intention recognition and different lane-changing patterns［J］．Accident Analysis & Prevention，2022，164（1）：1-14.
11	SUN Q， WANG C， FU R，et al ．Lane change strategy analysis and recognition for intelligent driving systems based on random forest［J］．Expert Systems with Applications，2021，186（10）：1-14.
12	CHEN S， YAO H， QIAO F，et al ．Vehicles driving behavior recognition based on transfer learning［J］．Expert Systems with Applications，2023，213（12）：119-254．
13	ZHAO D， ZHAO S ．Sparse least squares support vector machine based methods for vehicle driving behavior recognition［J］．Proceedings of the Institution of Mechanical Engineers，Part D：Journal of Automobile Engineering，2024，238（6）：1392-1404.
14	RAVI C， TIGGA A， REDDY G T，et al ．Driver identification using optimized deep learning model in smart transportation［J］．ACM Transactions on Internet Technology，2022，22（4）：1-17．
15	ALJOHANI A A ．Real-time driver distraction recognition：a hybrid genetic deep network based approach［J］．Alexandria Engineering Journal，2023，66（2）：377-389．
16	ZHAO W， WANG L， ZHANG Z ．Atom search optimization and its application to solve a hydrogeologic parameter estimation problem［J］．Knowledge-Based Systems，2019，163（1）：283-304．
17	GAO Z M， ZHAO J， ZHANG Y J ．Review of chaotic mapping enabled nature-inspired algorithms［J］．Mathematical Biosciences and Engineering，2022，19（8）：8215-8258．
18	余修武，黄露平，刘永，等．融合柯西折射反向学习和变螺旋策略的WSN象群定位算法［J］．控制与决策，2022，37（12）：3183-3189．
	YU Xiu-wu， HUANG Lu-ping， LIU Yong，et al ．Cauchy refraction opposition-based learning and variable helix mechanism of elephant herding localization algorithm in WSN［J］．Control and Decision，2022，37（12）：3183-3189．

函数类别	平均耗时/s		IASO与ASO平均耗时的比值/%
函数类别	ASO	IASO	IASO与ASO平均耗时的比值/%
均值	0.948	0.944	99.6
单峰基准测试函数	1.014	1.004	99.0
多峰基准测试函数	0.868	0.882	101.6
复合基准测试函数	0.963	0.945	98.1

算法名称	MAE	排名
IASO	192.2	1
ASO	538.6	2
GWO	665.3	3
PSO	730.1	4
GA	2 540.8	5

类别	精确率/%			召回率/%			F₁值/%
类别	IASO-LSSVM	ASO-LSSVM	LSSVM	IASO-LSSVM	ASO-LSSVM	LSSVM	IASO-LSSVM	ASO-LSSVM	LSSVM
整体	98.5	84.4	69.5	98.0	82.9	69.4	98.2	83.5	67.0
正常驾驶	96.0	87.6	50.0	96.8	93.0	65.4	96.4	90.2	56.7
缓速行驶	99.2	89.7	96.2	100.0	91.2	52.9	99.6	90.4	68.3
近距跟驰	100.0	91.7	97.0	100.0	97.1	97.0	100.0	94.3	97.0
急加减速	96.3	58.2	9.0	91.2	53.3	33.3	93.7	55.6	14.2
频繁换道	100.0	85.4	83.3	100.0	72.9	76.9	100.0	78.7	80.0
强行换道	99.2	93.6	81.7	100.0	90.0	90.6	99.6	91.8	85.9

[1]	WENG Jiancheng, WU Mingzhu, WEI Ruicong, et al. Speed Prediction for Road Around Large Scale Activities Venues Considering Multiple Factors Synergism [J]. Journal of South China University of Technology(Natural Science Edition), 2024, 52(8): 34-44.
[2]	TANG Jinjun, REN Maoxin, LI Zhitao, et al. Optimization of Metro Feeder Bus Routes Based on Surrogate-assisted NSGA-II Algorithm [J]. Journal of South China University of Technology(Natural Science Edition), 2024, 52(11): 95-105.
[3]	CHEN Xiaohong, TIAN Tiantian, ZHANG Hua. Activity Resilience of Urban Residents: Empirical Evidence Based on the Recovery of Rail Transit Ridership During the COVID-19 Pandemic [J]. Journal of South China University of Technology(Natural Science Edition), 2023, 51(7): 109-119.
[4]	Zhou Jia- zhong Zhang Dian- ye. Behavior Model of Urban Trip Chains in Multi- Mode Transportation Network [J]. Journal of South China University of Technology (Natural Science Edition), 2014, 42(2): 125-131.