考虑NMS的新型驾驶员模型在转向控制中的应用研究

发布时间：2018-02-23 20:24

本文关键词： 驾驶员模型多点预瞄线性二次型神经肌肉系统车辆稳定性仿真　出处：《南京航空航天大学》2014年硕士论文　论文类型：学位论文

【摘要】：驾驶员驾驶机动车时，因个人自身因素引起的道路交通安全事故发生频繁，驾驶员的驾驶行为，是诱发交通事故的主要原因，如何对驾驶员转向行为特性进行分析与研究进而营造一个安全、舒适的驾驶环境也就成为了目前研究的热点。鉴于目前采用车辆测试成本较高、时间长、同时还会受到空间等环境条件的限制等因素的影响，在这种情况下建立精确、合理、符合驾驶员特性的驾驶员转向控制模型，对于车辆实车试验验证与分析、车辆辅助设计及确保行车安全具有重要意义。本文针对驾驶员转向行为所涉及的问题，，对基于线性二次型（LQR）的驾驶员转向控制模型、神经肌肉动态性模型及车辆稳定性控制模型等内容进行了较深入的研究，其具体研究内容可以归纳如下：首先，设计了基于线性二次型（LQR）并考虑驾驶员延时特性的驾驶员转向控制模型。针对驾驶员转向控制方法进行研究，根据车辆动力学及线性规划最优控制理论，采用多点预瞄机制建立了“最优-预瞄”驾驶员转向控制模型。针对不同控制权值进行控制轨迹跟随性能的分析。根据反馈矩阵系数来探究最优驾驶员的内在驾驶特性。采用状态移位寄存器的形式将驾驶员延时特性考虑到驾驶员转向控制模型中，使设计的模型更贴近实际驾驶员转向行为特性。采用双移线的仿真方法验证该驾驶员转向控制模型的有效性。其次，将神经肌肉动态性模型融合到驾驶员转向控制模型中。根据神经肌肉具体作用过程及肌肉的反应反射的内在机理，设计了手臂转向动态性模型与神经肌肉控制环节构成的神经肌肉动态性模型。其中神经肌肉控制环节包含反射控制、刚度因子与参考模型三部分，并得到各部分模型或传递函数的具体形式：根据手臂及车辆转向系统的动态耦合性，得到手臂转向动态性模型的运动方程；通过肌肉协同收缩得到刚度因子；通过系统辨识的方法得到了耦合的驾驶员、车辆及转向系统之间肌肉力矩及转向角之间的传递函数，从而得到内部参考模型。考虑了转向力矩反馈对驾驶员神经肌肉动态性的影响，将神经肌肉动态性模型融合到驾驶员转向控制模型中，并进行仿真分析。通过仿真验证神经肌肉动态性参数对车辆转向及轨迹跟踪性能的影响。最后，在包含神经肌肉动态性的驾驶员转向控制模型的基础上，提出建立车辆稳定性控制模型。考虑到轮胎的非线性特性，基于“魔术公式”及车辆动力学建立了整车动力学模型。根据车辆的线性动力学模型及线性二次型最优控制理论，采用横摆角速度和质心侧偏角作为控制变量，建立一种增加附加横摆力矩的稳定性控制模型将其应用到整车动力学模型中，并进行仿真验证。验证该稳定性控制模型能够保证车辆在不同附着系数路面上的稳定性。综上所述，本文一方面设计了驾驶员转向控制模型，通过双移线仿真实验，验证不同控制权值下的轨迹跟踪性能，控制效果较好，以选取最优的驾驶员转向控制模型。另一方面将神经肌肉动态性模型融合到驾驶员转向控制模型中，研究结果表明驾驶员神经肌肉动态性参数（反射增益、刚度因子）及内部参考模型的精确性，对于车辆轨迹跟踪性能具有重要影响。在此基础上，建立了车辆稳定性控制模型，不仅提高了车辆在不同附着系数路面上的稳定性，而且其研究结果在相关领域中有较重要的理论意义和工程应用价值。
[Abstract]:The driver when driving the motor vehicle, road traffic safety accidents caused by the personal factors of the frequent occurrence of the driving behavior, is mainly caused by traffic accidents, how to driver steering behavior analysis and research and to create a safe, comfortable driving environment has become the hotspot of the research. In view of the vehicle test of high cost, long time, but also by the influence of space environment conditions and other factors, in this case to establish accurate, reasonable, accord with the characteristics of drivers driver for vehicle steering control model, vehicle test and analysis, has the important meaning of vehicle design and ensure traffic safety.
Aiming at the problems involved in driver's steering behavior, this paper makes a deep research on driver's steering control model based on linear two times (LQR), neuromuscular dynamic model and vehicle stability control model, etc. the specific research contents can be summarized as follows:
First of all, the design based on the linear two type (LQR) and considering the delay characteristics of the driver steering control model. According to the driver's steering control method is studied, according to the vehicle dynamics and linear programming optimal control theory, established the "best - Preview" driver steering control model using multipoint Preview for different control mechanisms. Weight control analysis of trajectory following performance. According to the internal driving characteristics to explore the optimal feedback matrix coefficients of the driver. The state of the shift register form driver delay characteristics considering the driver's steering control model, the design of the model closer to the actual driver steering behavior. The simulation method using double lane change to verify that the driver's steering effectiveness control model.
Secondly, the fusion of nerve muscle dynamic model to the driver's steering control model. According to the internal mechanism of reaction process and the specific role of reflex muscle muscle, the steering arm design of nerve muscle model dynamic model and a dynamic neuromuscular control link. The neuromuscular control link includes reflection control, stiffness factor with the three parts of the reference model, and get the specific form of each part of the model or the transfer function: according to the dynamic coupling system of vehicles and arms, arm steering motion equation obtained dynamic model; CO contraction stiffness factor by muscle; through the method of system identification has been coupled to the driver, vehicle and steering system. Muscle torque and steering angle between the transfer function, so as to obtain the internal reference model. Considering the steering torque feedback to the driver by God The influence of muscle dynamics is integrated into the driver's steering control model, and the simulation analysis is carried out. The influence of neuro muscular dynamic parameters on vehicle steering and trajectory tracking performance is verified by simulation.
Finally, the dynamic nature of the nerve muscle contains the driver's steering control based on the model, proposed the establishment of vehicle stability control model. Considering the nonlinear characteristics of tires, based on the "magic formula" and vehicle dynamics vehicle dynamic model has been established. According to the linear dynamics model and linear vehicle two optimum control theory, the yaw angle speed and sideslip angle as control variables, establish a stability of additional yaw moment control model is applied to the vehicle dynamics model, and simulation. Verify the stability control model can ensure the vehicle stability with different adhesion coefficient road.
In summary, this paper designed a driver steering control model, through the double lane change simulation, verify the tracking performance under different weight control trajectory, good control effect, to select the best driver steering control model. On the other hand the fusion neuromuscular dynamic model to the driver's steering control model, the results show that the nerve muscle the dynamic parameters (reflection gain, stiffness factor) accuracy and internal reference model, has an important effect on the tracking performance of the vehicle trajectory. On this basis, establish a vehicle stability control model not only improves the stability of the vehicle in different attachment coefficient road, and the research results have important theoretical significance and engineering the application value in related fields.

【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.25

【参考文献】