熊璐 , 杨兴 , 卓桂荣 , 等 . 无人驾驶车辆的运动控制发展现状综述 [J ] . 机械工程学报 , 2020 , 56 ( 10 ): 127 - 143 . DOI: 10.3901/JME.2020.10.127 http://doi.org/10.3901/JME.2020.10.127 回顾无人驾驶车辆的运动控制问题。从系统模型、控制方法以及控制结构等角度切入，分别在纵向运动控制、路径跟踪控制和轨迹跟踪控制三个层面对国内外的研究进展进行综述，并提出对无人驾驶车辆运动控制技术的发展展望。当前运动控制研究多集中于常规工况，为实现无人驾驶车辆在处理人类驾驶员认为具有挑战性或缺乏操纵能力的复杂动态场景下的潜力，运动控制研究须从常规工况向极限工况拓展，但是极限工况下车辆的非线性和多维运动耦合特征显著增强，对系统建模以及算法的自适应性和鲁棒性的要求进一步提高。同时，为应对复杂场景下的多目标协调优化问题，考虑环境不确定性的运动规划与控制集成设计需要深入研究。增加执行器手段可以提升极限工况下车辆的侧向响应速度和控制裕度，但是冗余异构执行器的控制分配研究仍有待突破。运动控制的实现依赖于路面附着系数、质心侧偏角等信息输入，因此基于多源传感信息融合的关键状态与参数估计问题亟需解决。此外，将机器学习应用到车辆运动控制领域也是一个重要的发展方向。

XIONG L , YANG X , ZHUO G R , et al. Review on motion control of autonomous vehicles [J ] . Journal of Mechanical Engineering , 2020 , 56 ( 10 ): 127 - 143 . (in Chinese) DOI: 10.3901/JME.2020.10.127 http://doi.org/10.3901/JME.2020.10.127 The motion control problem of autonomous vehicles is reviewed. From the perspective of model, algorithm, and control structure, the domestic and foreign research progress is reviewed at three levels of longitudinal motion control, path following and trajectory tracking control, and the development prospect of motion control technology for autonomous vehicles is proposed. The current motion control research mainly focuses on normal conditions. In order to realize the potential of autonomous vehicles in handling critical scenarios that human drivers find challenging or lack the ability to navigate, it is necessary to extend the research to extreme working conditions. However, the properties of non-linearity and multi-dimensional coupled dynamics are significantly enhanced in extreme working conditions. The requirements of system modeling and adaptability and robustness of motion control algorithm are further increased. At the same time, in order to deal with the multi-objective coordination in complex scenarios, the integration of motion planning and control considering environmental uncertainty needs to be studied in depth. Adding actuators can increase the lateral response speed and control margin, but the research of control allocation of redundant and heterogeneous actuators is still to be broken through. The realization of motion control depends on road adhesion coefficient, sideslip angle, etc. Therefore, it is urgent to solve the problem of key state and parameter estimation under multi-source sensor information fusion. In addition, the application of machine learning to the field of vehicle motion control is also an important development direction.

高振海 , 朱乃宣 , 高菲 , 等 . 考虑驾驶员特性的自学习换道轨迹规划系统 [J ] . 汽车工程 , 2020 , 42 ( 12 ): 1710 - 1717 . DOI: 10.19562/j.chinasae.qcgc.2020.12.014 http://doi.org/10.19562/j.chinasae.qcgc.2020.12.014 为更好地实现个性化驾驶,本文中提出了一种集成驾驶员特性辨识的自学习换道轨迹规划系统。首先,在高斯分布中引入驾驶员特性系数J_c和驾驶员反应与操作时间t_d,建立了个性化换道轨迹规划模型,并通过DTW算法对实际轨迹和拟合轨迹进行匹配。之后,基于采集的驾驶员换道轨迹进行AP聚类,离线标定J_c和t_d共性化值,同时获得30名驾驶员的标签,将其驾驶特性分为舒适、一般和运动型。然后,将自由驾驶数据进行特征工况的提取,并基于长短期记忆网络(LSTM)搭建驾驶员特性在线辨识模型进行训练。最后,选取15名驾驶员进行实车验证,系统实时提取特征工况,然后基于辨识结果在线调整J_c和t_d,并不断更新拟合轨迹。实验结束后,其中14名驾驶员的实际轨迹与拟合轨迹平方欧氏距离小于1,拟合正确率达93.3%。因此,该系统能够良好地复现真人换道轨迹。

GAO Z H , ZHU N X , GAO F , et al. Self-learning lane-change trajectory planning system with driver characteristics [J ] . Automotive Engineering , 2020 , 42 ( 12 ): 1710 - 1717 . (in Chinese)

LI A , JIANG H B , LI Z J , et al. Human-like trajectory planning on curved road:learning from human drivers [J ] . IEEE Transactions on Intelligent Transportation Systems , 2020 , 21 ( 8 ): 3388 - 3397 .

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 .

BERGMAN K , LJUNGQVIST O , AXEHILL D . Improved path planning by tightly combining lattice-based path planning and optimal control [J ] . IEEE Transactions on Intelligent Vehicles , 2020 , 6 ( 1 ): 57 - 66 .

STANO P , MONTANARO U , TAVERNINI D , et al. Model predictive path tracking control for automated road vehicles:a review [J ] . Annual Reviews in Control , 2023 ,55: 194 - 236 .

杜荣华 , 胡鸿飞 , 高凯 , 等 . 基于变预测时域MPC的自动驾驶汽车轨迹跟踪控制研究 [J ] . 机械工程学报 , 2022 , 58 ( 24 ): 275 - 288 . DOI: 10.3901/JME.2022.24.275 http://doi.org/10.3901/JME.2022.24.275 为了保证自动驾驶汽车轨迹跟踪的精度及行驶过程中的稳定性，提出一种基于车辆横向稳定状态在线识别和模糊算法的变预测时域模型预测控制（MPC）方法。针对车辆稳定状态的在线识别，采用k-means聚类算法对车辆行驶状态参数进行聚类分析，得到聚类质心，通过在线对比当前车辆状态量与不同聚类质心之间的欧氏距离获取车辆的实时安全等级。同时计算出当前车辆的轨迹跟踪横向偏移量，以这二者为输入，通过模糊控制算法在线计算出预测时域的变化量并输出给MPC控制器实现预测时域的自适应调整，最后求解出自动驾驶车辆跟踪轨迹的最优的控制序列，以达到在保持车辆稳定的前提下实现高精度轨迹跟踪控制的目的。CarSim/Simulink联合仿真结果表明，改进后的变预测时域MPC算法在提高自动驾驶汽车轨迹跟踪精度及横向稳定性方面的表现优于传统MPC控制器。

DU R H , HU H W , GAO K , et al. Research on trajectory tracking control of autonomous vehicle based on mpc with variable predictive horizon [J ] . Journal of Mechanical Engineering 58 ( 24 ): 275 - 288 . (in Chinese)

HE H W , SHI M , LI J W , et al. Design and experiential test of a model predictive path following control with adaptive preview for autonomous buses [J/OL ] . Mechanical Systems and Signal Processing , 2021 ,157:107701.

李韶华 , 杨泽坤 , 王雪玮 . 基于T-S模糊变权重MPC的智能车轨迹跟踪控制 [J ] . 机械工程学报 , 2023 , 59 ( 4 ): 199 - 212 . DOI: 10.3901/JME.2023.04.199 http://doi.org/10.3901/JME.2023.04.199 为了协调智能驾驶车辆的轨迹跟踪精确性和稳定性，提高控制算法对不同工况的自适应能力，提出基于Takagi-Sugeno模糊变权重模型预测控制(Takagi-Sugeno fuzzy model predictive control，T-S FMPC)的轨迹跟踪控制策略。以前轮转角为控制变量建立MPC控制，并以实时横向位移误差和横摆角误差为模糊输入，通过T-S模糊控制在线优化MPC目标函数权重，协调权重矩阵对轨迹跟踪精确性和稳定性的影响。基于Carsim建立分布式驱动电动汽车的整车动力学模型，基于Simulink建立控制策略，通过双移线工况仿真及实车试验，验证了所提控制策略的有效性。仿真结果表明，相比于传统MPC控制，所提出的T-S模糊变权重MPC控制可降低横向位移误差达62.24%，有效提高轨迹跟踪精度；并且可使前轮转角波动减小37.46%、横摆角误差减小84.19%，显著增强轨迹跟踪稳定性；试验结果表明，在20 km/h、沥青路面双移线工况下，横向位移误差在0.12 m以内，横摆角误差在1°以内，且前轮转角控制曲线平滑，说明所提算法具有良好的控制效果和实用性。

LI S H , YANG Z K , WANG X W . Trajectory tracking control of an intelligent vehicle based on T-S fuzzy variable weight MPC [J ] . Journal of Mechanical Engineering 2023 , 59 ( 4 ): 199 - 212 . (in Chinese)

ROKONUZZAMAN M , MOHAJER N , NAHAVANDI S . Human-tailored data-driven control system of autonomous vehicles [J ] . IEEE Transactions on Vehicular Technology , 2022 , 71 ( 3 ): 2485 - 2500 .

马跃 , 郭烈 , 秦增科 , 等 . 考虑驾驶风格的路径跟踪控制方法 [J ] . 重庆理工大学学报(自然科学) , 2022 , 36 ( 11 ): 20 - 30 .

MA Y , GUO L , QIN Z K , et al. Path tracking control method considering driving styles [J ] . Journal of Chongqing University of Technology( Natural Science) , 2022 , 36 ( 11 ) : 20 - 30 . (in Chinese)

HONDA K , OKUDA H , SUZUKI T . Multi-task model predictive control based on continuation with intermediate mode[C ] //Proceedings of the 2020 IEEE 23rd International Conference on Intelligent Transportation Systems ( ITSC).Rhodes,Greece:IEEE ,2020: 1 - 7 .

HONDA K , OKUDA H , SUZUKI T , et al. MPC builder for autonomous drive:automatic generation of MPCs for motion planning and control[C ] //Proceedings of the 2023 IEEE Intelligent Vehicles Symposium ( IV). Anchorage,AK,US:IEEE ,2023: 1 - 8 .

王博洋 , 龚建伟 , 张瑞增 , 等 . 基于真实驾驶数据的运动基元提取与再生成 [J ] . 机械工程学报 , 2020 , 56 ( 16 ): 155 - 165 . DOI: 10.3901/JME.2020.16.155 http://doi.org/10.3901/JME.2020.16.155 类人驾驶系统是通过学习人类驾驶员知识与经验来提升无人驾驶系统适用性与接受度的重要技术途径。为解决驾驶员轨迹和操控层面经验的表述问题，以采集得到的大量真实驾驶数据为依托，提出一种基于轨迹基元与操控基元的分层式驾驶员经验表述模型。轨迹基元以动态运动基元算法进行表征，并由概率提取算法完成基元从无标签连续轨迹数据中的分割提取。操控基元在轨迹基元的提取分类结果上，利用高斯混合模型完成基元的训练，并利用高斯回归算法完成转向操控序列的预测。结果表明，概率提取算法既利用到了表征与提取之间的相互关联关系，又借助于初始分割点的合理设置，提升了算法的效率并使得提取得到的运动基元符合特定的驾驶假设。此外，所提出的运动基元既能以较高精度完成对驾驶员轨迹和操控层面数据的表征，又具备良好的泛化能力以应对运动基元再生成时在期望位置和时间尺度上的变化需求。最终构建了描述全工况驾驶行为的运动基元库，并大幅提升了运动基元应对不同行车环境时的适用性。

WANG B Y , GONG J W , ZHANG R Z , et al. Motion primitives extraction and regeneration based on real driving data [J ] . Journal of Mechanical Engineering , 2020 , 56 ( 16 ): 155 - 165 . (in Chinese) DOI: 10.3901/JME.2020.16.155 http://doi.org/10.3901/JME.2020.16.155 The human-like driving system is an essential technical way to improve the applicability and acceptance of an unmanned driving system by learning the knowledge and experience of human drivers. In order to solve the driving skills representation problem at trajectory and control level, by utilizing a large amount of collected real driving data, a hierarchical driver model based on trajectory primitives and operation primitives is proposed. The trajectory primitives are represented by the dynamic movement primitive, and the probabilistic extraction algorithm is applied to extract primitives from the unlabeled continuous trajectory data. The operation primitives use the Gaussian mixture model to complete the training process based on the extraction and classification results of the trajectory primitives. The Gaussian mixture regression(GMR) algorithm is applied to predict the steering angle. The results show that the probabilistic extraction algorithm not only utilizes the correlation between representation and extraction but also uses the reasonable setting of the initial segmentation point, which improves the efficiency of the algorithm and makes the extracted motion primitives conform to specific driving assumptions. The proposed motion primitives can not only represent the driver's driving data with high precision but also have good generalization ability to deal with the desired position and time duration change when the motion primitives are regenerated. Finally, the motion primitive library describing the driving behavior under all conditions is established, and the applicability of the motion primitives to different driving situations is significantly improved.

WANG B Y , GONG J W , LIANG W , et al. Regeneration and joining of the learned motion primitives for automated vehicle motion planning applications[C ] //Proceedings of 2019 IEEE Intelligent Vehicles Symposium ( IV).Paris,France:IEEE ,2019: 790 - 797 .

RAJAMANI R . Vehicle dynamics and control [M ] . Berlin , Germany:Springer , 2011 .

卢佳兴 , 刘海鸥 , 关海杰 , 等 . 基于双参数自适应优化的无人履带车辆轨迹跟踪控制 [J ] . 兵工学报 , 2023 , 44 ( 4 ): 960 - 971 .

LU J X , LIU H O GUAN H J , et al. Trajectory tracking control of unmanned tracked vehicles based on adaptive dual-parameter optimization [J ] . Acta Armamentarii , 2023 , 44 ( 4 ): 960 - 971 . (in Chinese) DOI: 10.12382/bgxb.2022.0009 http://doi.org/10.12382/bgxb.2022.0009 To improve the poor adaptability of trajectory tracking controllers with fixed parameters, an optimized adaptive dual-parameter trajectory tracking algorithm for unmanned tracked vehicles based on the improved Particle Swarm Optimization (IPSO) and Multi-Layer Perceptron (MLP) algorithms is proposed. In the offline state, based on the collected actual vehicle data, the IPSO algorithm is used to construct the optimal parameter data set under different motion primitives, aiming for high accuracy, high stability, and low time cost of trajectory tracking. With the motion primitive type and vehicle speed as feature vectors, control time domain length and control time step length as labels, adaptive learning rate optimization algorithm is used to complete the training of the MLP neural network model. In the online state, according to the trajectory information and vehicle state feedback information provided by the planning layer, the MLP neural network outputs the predicted optimal control time domain length and control time step. These parameters are then input to the model predictive controller as dual parameters, enabling the adaptive trajectory tracking control. ROS-VREP co-simulation test and actual vehicle test based on a bilateral electric drive platform are carried out. Vehicle test results show that under various working conditions including large curvature steering, the proposed controller achieves a 30.5% reduction in average lateral error, a 17.2% decrease in average heading error, and a 7.8% reduction in average change rate of rotation angle, compared with the fixed-parameter trajectory tracking control method with the same calculation time cost. The results verify the feasibility and effectiveness of the new algorithm.

龚建伟 . 无人驾驶车辆模型预测控制 [M ] . 北京 : 北京理工大学出版社 , 2020 .

GONG J W . Model predictive control for self-driving vehicles [M ] . Beijing : Beijing Institute of Technology Press , 2020 . (in Chinese)

臧勇 , 蔡英凤 , 孙晓强 , 等 . 基于可拓博弈的智能汽车轨迹跟踪协调控制方法研究 [J ] . 机械工程学报 , 2022 , 58 ( 8 ): 181 - 194 . DOI: 10.3901/JME.2022.08.181 http://doi.org/10.3901/JME.2022.08.181 针对智能汽车在高速过弯工况下，轨迹跟踪误差大、横向稳定性无法保障的问题，提出一种可拓博弈轨迹跟踪协调控制方法，通过可拓划区域切换控制和博弈协调相结合，突破单一控制策略的工况适应性和多策略切换的抖动问题。所提方法基于分层控制体系，将轨迹跟踪控制分解为上层测度模式识别层和下层博弈协调层。上层基于可拓理论，提出并联可拓测度模式识别策略，将车-路系统实时状态映射至对应的可拓控制架构中经典域、可拓域和非域三种测度模式。下层针对不同测度模式对应设计三种控制策略，根据上层测度模式识别结果进行实时策略切换，引入博弈协调方法对并联可拓权重进行协调控制，有效避免了模式切换带来的抖动问题。通过Simulink/Carsim建立联合仿真模型，在双移线和“8字”形时变曲率高速工况开展算法对比验证，所提方法相较于比例-积分-微分(Proportion-integral-derivative，PID)控制方法，平均跟踪误差精度提升45.08%，尤其在大曲率突变的恶劣工况下，车辆稳定性提升44%。最后利用智能汽车试验平台进行了对比验证，对设计智能汽车高速轨迹跟踪控制策略具有极强的指导意义和参考价值。

ZANG Y , CAI Y F , SUN X Q , et al. Research on intelligent vehicle trajectory tracking coordination control method based on extension game [J ] . Journal of Mechanical Engineering , 2022 , 58 ( 8 ): 181 - 194 . (in Chinese) DOI: 10.3901/JME.2022.08.181 http://doi.org/10.3901/JME.2022.08.181 For the problems of intelligent vehicle under high-speed cornering conditions, large trajectory tracking errors and lateral stability cannot be guaranteed, an extension game trajectory tracking coordination control method is proposed, which combines extension zone switching control and game coordination. It breaks the working condition adaptability of a single control strategy and the jitter problem of multiple strategy switching control. The proposed method is based on a hierarchical control system, which decomposes the trajectory tracking control into an upper measurement pattern recognition layer and a lower game coordination layer. Based on the extension theory, the upper layer proposes a parallel extension measurement pattern recognition strategy, and maps the real-time state of the vehicle-road system to the corresponding three measurement modes:classic domain, extension domain, and non-domain in extension control architecture. The lower layer designs three control strategies corresponding to different measurement modes. The real-time policy switching is performed based on the recognition results of upper measurement modes. The game coordination method is introduced to coordinate the parallel extension weights, which effectively avoids the jitter problem caused by mode switching control. The joint simulation model is established by Simulink/Carsim, and the algorithm is compared and verified in the double-shift and "8-shaped" with time-varying curvature and high-speed conditions. Compared with the Proportion- Integral-Derivative(PID) control method, the proposed method improves the average tracking error accuracy by 45.08%, especially under bad working conditions with sudden changes in curvature, vehicle stability is improved by 44%. Finally, the intelligent vehicle experiment platform is used for comparison and verification, which has strong guiding significance and reference value for designing high-speed trajectory tracking control strategies for intelligent vehicles.

NA X X , COLE D J . Linear quadratic game and non-cooperative predictive methods for potential application to modelling driver-AFS interactive steering control [J ] . Vehicle System Dynamics , 2013 , 51 ( 2 ): 165 - 198 .