Surrounding vehicle recognition and information map construction technology in automatic driving

doi:10.3969/j.issn.1674-8484.2022.01.013

Abstract

Abstract:

A method of vehicle recognition and map construction during automatic driving was proposed based on the fusion of camera image and lidar information to improve satisfactory success rate of single sensor recognition of surrounding vehicles and engineering practicability. The joint calibration and time registration of the camera and the lidar were carried out. The vehicle shadows and lidar detection information in the image were extracted, and the coordinate correlation features between the two were obtained, and whether there was a vehicle nearby was determined according to its threshold range. The camera image information was fused with the detection information of lidar, and the information map of the surrounding vehicles was constructed, and the experimental verification was carried out. The results show that the accuracy of vehicle recognition is 4.25% higher than that of (D-S) evidence theory by using the feature of coordinate correlation degree. The vehicle information map in this paper can realize overtaking track driving based on third-order Bessel curve on the premise of safety.

Key words: automatic pilot, vehicle identification, map construction, information fusion, time registration

CLC Number:

U471.15

LI Maoyue, LV Hongyu, HE Xiangmei, XU Guangqi, YU Wei. Surrounding vehicle recognition and information map construction technology in automatic driving[J]. Journal of Automotive Safety and Energy, 2022, 13(1): 131-141.

Figures/Tables 26

References 16

[1]	Herrero-Huerta M, Rodriguez Gonzalvez P, Rainey K M. Yield prediction by machine learning from UAS-based multi-sensor data fusion in soybean[J]. Plant Methods, 2020,16:1-16.
[2]	SHAN Pengfei, LÜ Hui, YU Liming, et al. A multisensor data fusion method for ball screw fault diagnosis based on convolutional neural network with selected channels[J]. IEEE Sensors J, 2020,20(14):7896-7905.
[3]	XU Liu, ZHOU Baoding, HUANG Panpan, et al. Kalman filter-based data fusion of Wi-Fi RRT and PDR for indoor localization[J]. IEEE Sensors J, 2021,21(6):8479-8490.
[4]	DU Hao, WEI Wang, XU Chaowen, et al. Real-time onboard 3D state estimation of an unmanned aerial vehicle in multi-environments using multi-sensor data fusion[J]. Sensors, 2020,20(3):919.
[5]	LIU Yahui, FAN Xiaoqian, LÜ Chen, et al. An innovative information fusion method with adaptive Kalman filter for integrated INS/GPS navigation of autonomous vehicles[J]. Mech Syst Signal Procg, 2018,100:605-616.
[6]	杨永旭, 高自凡, 朱辉, 等. 一种基于决策距离测量与D-S证据理论结合的多源冲突证据信息融合算法[J]. 兰州文理学院学报, 2018,32(6):62-67.
	YANG Yongxu, GAO Zifan, ZHU Hui, et al. A multi-source conflict evidence information fusion algorithm based on decision distance measurement and D-S evidence theory[J]. J Lanzhou Univ Arts Sci, 2018,32(6):62-67. (in Chinese)
[7]	Asvadi A, Premebida C, Peixoto P, et al. 3D Lidar-based static and moving obstacle detection in driving environments: An approach based on voxels and multi-region ground planes[J]. Robot Autonomous Syst, 2016,83:299-311.
[8]	Alonso J D, Vidal E R, Rotter A, et al. Lane-change decision aid system based on motion-driven Vehicle Tracking[J]. IEEE Trans Vehi Tech, 2008,57(5):2736-2746.
[9]	Ochman M. Hybrid approach to road detection in front of the vehicle[J]. IFAC-PapersOnLine, 2019,52(8):245-250.
[10]	FANG Qiongyao, LIANG Juihung, LU Chaoshan, et al. A real-time visual-based front-mounted vehicle collision warning system[C]// 2013 IEEE Symp Compu Intell Vehi Transp Syst (CIVTS), 2013: 1-8.
[11]	SHAN Zhenyu, ZHU Qianqian, ZHAO Danna. Vehicle collision risk estimation based on RGB-D camera for urban road[J]. Mulimedia Syst, 2017,23(1):119-127.
[12]	李茂月, 吕虹毓, 王飞, 等. 基于机器视觉的智能车辆鲁棒车道线识别方法[J]. 中国机械工程, 2021,32(2):242-251.
	LI Maoyue, LÜ Hongyu, WANG Fei, et al. Robust lane line recognition method for intelligent vehicles based on machine vision[J]. Chin Mech Eng, 2021,32(2):242-251. (in Chinese)
[13]	李文立, 郭凯红. D-S证据理论合成规则及冲突问题[J]. 系统工程理论与实践, 2010,30(8):1422-1432.
	LI Wenli, GUO Kaihong. Rules and conflicts of D-S evidence theory[J]. Syst Engi Theory Practice, 2010,30(8):1422-1432. (in Chinese)
[14]	刘放, 王衍学. 基于多域特征与改进D-S证据理论的齿轮故障智能诊断方法[J]. 机械传动, 2019,43(9):159-165.
	LIU Fang, WANG Yanxue. Intelligent diagnosis method of gear faults based on multi-domain characteristics and improved D-S evidence theory[J]. Mech Transmission, 2019,43(9):159-165. (in Chinese)
[15]	徐孙庆, 耿俊豹, 魏曙寰, 等一种改进的D-S冲突证据合成方法[J]. 火力与指挥控制, 2019,44(10):84-88.
	XU Sunqing, GENG Junbao, WEI Shuhuan, et al. An improved D-S conflict evidence synjournal method[J]. Fire Command Control, 2019,44(10):84-88. (In Chinese)
[16]	李进, 陈杰平, 徐朝胜, 等. 基于动态图像阈值的智能车辆路径导航[J]. 农业机械学报, 2013,44(4):39-44.
	LI Jin, CHEN Jieping, XU Zhaosheng, et al. Intelligent vehicle path navigation based on dynamic image threshold[J]. J Agric Mach, 2013,44(4):39-44. (in Chinese)

相机位置	Δx/mm	Δy/mm	Δz/mm	α/(°)	β/(°)	γ/(°)
前置	30	0	-13	12.22	2.18	-24.52
左置	-15	-60	-41	2.47	-5.94	-24.14
右置	-15	-60	-41	2.00	2.49	-24.79
后置	-140	0	-42.5	11.93	2.38	-25.00

相机位置	Δx/mm	Δy/mm	Δz/mm	α/(°)	β/(°)	γ/(°)
前置	30	0	-13	12.22	2.18	-24.52
左置	-15	-60	-41	2.47	-5.94	-24.14
右置	-15	-60	-41	2.00	2.49	-24.79
后置	-140	0	-42.5	11.93	2.38	-25.00

图像帧序号	横坐标				计算结果
图像帧序号	x_C1	x_C1	x_R1	x_R1	Δx₁	Δx₁	ε₁	ε₂
7	92	135	103	151	11	16	0.74	0.67
13	98	138	113	147	15	9	0.62	0.71
19	278	366	285	374	7	8	0.92	0.91
28	492	547	518	563	26	16	0.53	0.64
41	472	515	497	528	25	13	0.41	0.58
57	501	542	523	557	22	15	0.46	0.56
76	488	527	504	535	16	8	0.59	0.74
95	273	358	289	375	16	17	0.81	0.80
117	81	129	95	145	14	16	0.71	0.68
128	104	145	125	157	21	12	0.49	0.63

图像帧序号	横坐标				计算结果
图像帧序号	x_C1	x_C1	x_R1	x_R1	Δx₁	Δx₁	ε₁	ε₂
7	92	135	103	151	11	16	0.74	0.67
13	98	138	113	147	15	9	0.62	0.71
19	278	366	285	374	7	8	0.92	0.91
28	492	547	518	563	26	16	0.53	0.64
41	472	515	497	528	25	13	0.41	0.58
57	501	542	523	557	22	15	0.46	0.56
76	488	527	504	535	16	8	0.59	0.74
95	273	358	289	375	16	17	0.81	0.80
117	81	129	95	145	14	16	0.71	0.68
128	104	145	125	157	21	12	0.49	0.63

图像帧	横坐标						计算结果 1			计算结果 2
图像帧	x_C3	x_C4	x_C5	x_R3	x_R4	x_R5	Δx₃	Δx₄	Δx₅	σ₁	σ₂	σ₃	σ₄
12	412	457	537	433	481	559	21	24	22	0.53	0.50	0.70	0.72
23	445	483	571	462	514	591	17	31	20	0.55	0.40	0.65	0.74
37	83	158	196	112	164	209	29	6	13	0.61	0.88	0.84	0.71
43	124	173	207	146	182	215	22	9	8	0.55	0.75	0.74	0.76
73	93	161	205	122	185	220	29	24	15	0.57	0.62	0.45	0.57
97	136	214	263	172	241	275	36	27	12	0.54	0.61	0.45	0.65
115	463	502	594	483	531	608	20	29	14	0.49	0.40	0.68	0.82
134	439	485	547	461	514	572	22	29	25	0.52	0.45	0.53	0.57