基于逆模型预测控制的拟人驾驶控制

doi:10.19562/j.chinasae.qcgc.2024.04.005

Abstract

Abstract:

In this paper， a human-like driving control based on inverse model predictive control is proposed， which realizes human-like driving by updating the weight coefficients of the cost function of the control module using the loss function of the real-time trajectory generated by the model predictive control and the driver's trajectory. The human-like driving control is constructed as a two-layer optimization problem. In the lower layer， real-time state trajectories are generated by solving a typical optimal control problem using model predictive control. The optimization objective function of the lower layer is then updated by minimizing the error between the generated real-time trajectories and those of human drivers in the upper layer. The auxiliary systems based on the differential Pontryagin's Maximum Principle are constructed to solve the gradient of the weight coefficients of the cost function for the real axis trajectory. The driver's driving data are collected from the real vehicle， imitated， and tested. The results show that the method proposed in this paper， compared with two types of inverse optimal control methods based on the virtual-time trajectory， reduces the maximum error with the real trajectory by 73.52% and 65.03% in the three test conditions， with the driving behavior more anthropomorphic and has the generalization performance.

Key words: self-driving, human-like driving, inverse optimal control

Hui Liu,Fawang Zhang,Shida Nie,Jingliang Duan,Congshuai Guo,Lingxiong Guo. Human-Like Driving Control Based on Inverse Model Predictive Control[J].Automotive Engineering, 2024, 46(4): 596-604.

Figures/Tables 14

符号/单位	释义	数值
$I z$ /（kg $· m 2$ ）	转动惯量	1 536.7
$k f$ /（N $· r a d - 1$ ）	前轴侧偏刚度	-128 916
$k r$ /（N $· r a d - 1)$	后轴侧偏刚度	-85 944
$l f$ /m	质心到前轴距离	1.06
$l r$ /m	质心到后轴距离	1.85
$m$ /kg	整车质量	2 000

算法 IMPC
1：	初始化权重系数 $θ$ 和学习率 $η θ$ ，给定真实驾驶轨迹 $ξ d = x 0 : M d$
2：	repeat
3：	前向过程：
4：	以式（3）产生长度为 $M$ 的实轴轨迹 $ξ θ$
5：	以式（5）计算上层损失
6：	反向过程：
7：	For $t = 1 : M$ ：
8：	通过式（12）计算 $? u t - 1 θ ? θ$
9：	通过式（14）计算 $? u t - 1 θ ? x t - 1 θ$
10：	通过式（8）计算 $d x t θ d θ$
11：	结束For循环
12：	通过式（7）计算 $d L d θ$ ，并通过式（6）更新 $θ$
13：	直到收敛

符号	释义	数值
$η θ$	学习率	5
$N$	预测时域	20
$M$	实轴轨迹步数	100
$K$	参数迭代次数	10 000

状态	Init	PDP	IKKT	IMPC
横向位置/m	0.41	1.08	0.58	0.13
纵向速度/ $(m · s - 1)$	0.56	0.39	0.45	0.10
横向速度/ $(m · s - 1)$	0.97	0.29	0.22	0.05
横摆角/ $r a d$	0.29	0.17	0.07	0.03
横摆角速度/ $(r a d · s - 1)$	0.65	0.19	0.11	0.03
前轮转角/ $r a d$	0.36	0.11	0.18	0.01
纵向加速度/ $(m · s - 2)$	0.96	0.44	0.40	0.40

状态	Init	PDP	IKKT	IMPC
横向位置/m	1.42	0.99	0.65	0.14
纵向速度/ $(m · s - 1)$	0.6	0.31	0.58	0.17
横向速度/ $(m · s - 1)$	0.71	0.21	0.19	0.03
横摆角/ $r a d$	0.22	0.12	0.07	0.02
横摆角速度/ $(r a d · s - 1)$	0.46	0.14	0.11	0.02
前轮转角/ $r a d$	0.32	0.11	0.16	0.01
纵向加速度/ $(m · s - 2)$	0.87	0.29	0.26	0.17

状态	Init	PDP	IKKT	IMPC
横向位置/m	1.53	1.06	0.69	0.19
纵向速度/ $(m · s - 1)$	1.05	0.88	0.80	0.20
横向速度/ $(m · s - 1)$	1.52	1.14	0.98	0.57
横摆角/ $r a d$	2.18	1.53	1.29	0.45
横摆角速度/ $(r a d · s - 1)$	0.96	0.75	0.65	0.38
前轮转角/ $r a d$	0.72	0.44	0.37	0.24
纵向加速度/（ $m · s - 2)$	0.99	0.97	1.05	0.39

References 31

1	XU D， DING Z， HE X， et al. Learning from naturalistic driving data for human-like autonomous highway driving［J］. IEEE Transactions on Intelligent Transportation Systems， 2020， 22（12）： 7341-7354.
2	EMUNA R， BOROWSKY A， BIESS A. Deep reinforcement learning for human-like driving policies in collision avoidance tasks of self-driving cars［J］. arXiv.2006.04218.2020.
3	HECKER S， DAI D， VAN GOOL L. Learning accurate， comfortable and human-like driving［J］. arXiv.1903.10995.2019.
4	SAMA K， MORALES Y， LIU H， et al. Extracting human-like driving behaviors from expert driver data using deep learning［J］. IEEE Transactions on Vehicular Technology， 2020， 69（9）： 9315-9329.
5	LIANG Y， LI Y， YU Y， et al. Path-following control of autonomous vehicles considering coupling effects and multi-source system uncertainties［J］. Automotive Innovation， 2021， 4（3）： 284-300.
6	GUO N， ZHANG X， ZOU Y. Real-time predictive control of path following to stabilize autonomous electric vehicles under extreme drive conditions［J］. Automotive Innovation， 2022， 5（4）： 453-470.
7	LI S E. Reinforcement learning for sequential decision and optimal control［M］. Springer， 2023.
8	WANG W， ZHANG Y， GAO J， et al. GOPS： a general optimal control problem solver for autonomous driving and industrial control applications［J］. Communications in Transportation Research， 2023， 3： 100096.
9	HECKER S， DAI D， LINIGER A， et al. Learning accurate and human-like driving using semantic maps and attention［C］. 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems （IROS）. IEEE， 2020： 2346-2353.
10	CODEVILLA F， SANTANA E， LÓPEZ A M， et al. Exploring the limitations of behavior cloning for autonomous driving［C］. Proceedings of the IEEE/CVF International Conference on Computer Vision， 2019： 9329-9338.
11	GU T， DOLAN J M. Toward human-like motion planning in urban environments［C］. 2014 IEEE Intelligent Vehicles Symposium Proceedings. IEEE， 2014： 350-355.
12	HANG P， LV C， XING Y， et al. Human-like decision making for autonomous driving： a noncooperative game theoretic approach［J］. IEEE Transactions on Intelligent Transportation Systems， 2020， 22（4）： 2076-2087.
13	SAMA K， MORALES Y， LIU H， et al. Extracting human-like driving behaviors from expert driver data using deep learning［J］. IEEE Transactions on Vehicular Technology， 2020， 69（9）： 9315-9329.
14	宋晓琳，盛鑫，曹昊天，等.基于模仿学习和强化学习的智能车辆换道行为决策［J］. 汽车工程， 2021， 43（1）： 59-67.
	SONG X L， SHENG X， CAO H T， et al. Lane‑change behavior decision‑making of intelligent vehicle based on imitation learning and reinforcement learning［J］. Automotive Engineering， 2021， 43（1）： 59-67.
15	XU Y， GAO W， HSU D. Receding horizon inverse reinforcement learning［J］. Advances in Neural Information Processing Systems， 2022， 35： 27880-27892.
16	LIU J， BOYLE L N， BANERJEE A G. An inverse reinforcement learning approach for customizing automated lane change systems［J］. IEEE Transactions on Vehicular Technology， 2022， 71（9）： 9261-9271.
17	HO J， ERMON S. Generative adversarial imitation learning［J］. Advances in Neural Information Processing Systems， 2016， 29.
18	KUDERER M， GULATI S， BURGARD W. Learning driving styles for autonomous vehicles from demonstration［C］. 2015 IEEE International Conference on Robotics and Automation （ICRA）， 2015.
19	KALAKRISHNAN M， PASTOR P， RIGHETTI L， et al. Learning objective functions for manipulation［C］. 2013 IEEE International Conference on Robotics and Automation. IEEE， 2013： 1331-1336.
20	LEVINE S， KOLTUN V. Continuous inverse optimal control with locally optimal examples［J］. arXiv， 2012.
21	ENGLERT P， VIEN N A， TOUSSAINT M. Inverse KKT： learning cost functions of manipulation tasks from demonstrations：［J］. The International Journal of Robotics Research， 2017： 1474-1488.
22	JIN W， KULIC D， LIN J F-S， et al. Inverse optimal control for multiphase cost functions［J］. IEEE Transactions on Robotics， 2019： 1387-1398.
23	JIN W， KULIC D， MOU S， et al. Inverse optimal control from incomplete trajectory observations［J］. The International Journal of Robotics Research， 2018.
24	LIANG Z， JIN W， MOU S. An iterative method for inverse optimal control［C］. 2022 13th Asian Control Conference （ASCC）， 2022.
25	JIN W， WANG Z， YANG Z， et al. Pontryagin differentiable programming： an end-to-end learning and control framework［J］. Advances in Neural Information Processing Systems， 2020， 33： 7979-7992.
26	GUO L， JIA Y. Inverse model predictive control （IMPC） based modeling and prediction of human-driven vehicles in mixed traffic［J］. IEEE Transactions on Intelligent Vehicles， 2020， 6（3）： 501-512.
27	GE Q， SUN Q， LI S， et al. Numerically stable dynamic bicycle model for discrete-time control［C］. 2021 IEEE Intelligent Vehicles Symposium Workshops （IV Workshops）. IEEE， 2021： 128-134.
28	LI S， LI K， RAJAMANI R， et al. Model predictive multi-objective vehicular adaptive cruise control［J］. IEEE Transactions on Control Systems Technology， 2011， 19（3）：556-566.
29	LI S， JIA Z， LI K， et al. Fast online computation of a model predictive controller and its application to fuel economy-oriented adaptive cruise control［J］. IEEE Transactions on Intelligent Transportation Systems， 2015， 16（3）：1199-1209.
30	赵菲，王建，张天雷，等. 云控场景下车辆队列的模型预测控制方法［J］. 汽车工程， 2022， 44（2）： 179-189.
	ZHAO F， WANG J， ZHANG T L， et al. Model predictive control method for vehicle platoon under cloud control scenes［J］. Automotive Engineering， 2022， 44（2）： 179-189.
31	王明，唐小林，杨凯，等. 考虑预测风险的自动驾驶车辆运动规划方法［J］. 汽车工程， 2023， 45（8）： 1362-1372.
	WANG M， TANG X L， YANG K， et al. A motion planning method for autonomous vehicles considering prediction risk［J］. Automotive Engineering， 2023， 45（8）： 1362-1372.

[1]	Tian Yuan,Xuan Zhao,Rui Liu,Qiang Yu,Xichan Zhu,Shu Wang. An Analysis on Longitudinal Driving Characteristics in Urban Intersection Based on Natural Driving Data [J]. Automotive Engineering, 2022, 44(6): 821-830.
[2]	Dongjian Song,Bing Zhu,Jian Zhao,Jiayi Han,Yanchen Liu. Human-Like Behavior Decision-Making of Intelligent Vehicles Based on Driving Behavior Generation Mechanism [J]. Automotive Engineering, 2022, 44(12): 1797-1808.
[3]	Xiong Ying, Mao Xuesong. Research on the Method of Navigating Autonomous Driving Vehicle Through Expressway Toll Region [J]. Automotive Engineering, 2020, 42(9): 1263-1269.
[4]	Liao Jiacai, Cao Libo, Xia Jiahao,Zhang Xiao & Wu Qiang. Multi-target Trajectory Tracking Based on Kernelized Correlation Filtering and Motion Model [J]. Automotive Engineering, 2019, 41(10): 1179-1188.

Human-Like Driving Control Based on Inverse Model Predictive Control

RichHTML

PDF (PC)

Like

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 14

References 31

Related Articles 4

Metrics

Comments

Recommended 0