基于线控转向系统的路感参数化重构和控制机理研究

Steer-by-wire vehicle is different from the conventional mechanical steering system, the dynamic coupling is realized by the distributed electricity transmission systems. Due to the lack of mechanical constraints, the issue of steer-by-wire system is how to ensure accurate delivery of steering feel. Previous study of steering feel is limited on the design of power-assisted system, the dynamic of steering torque and its reconfigurable mechanism is lack of systematic analysis currently. Our research proposed parametric expression method for steering feel through steer-by-wire system. Through the study of relationship between steering feel and driving performance, the optimization mechanism for steering characteristic is analyzed to satisfy different requirements. Firstly by using the previous research in dynamic model, we will investigate the solution to realize the position/torque coupling control with multidimensional state parameters. Then by taking analysis of the dynamic factors of steering torque, we proposed the online reconfigurable method for the parameterized steering feel. Since the nonlinear perturbations of the vehicle, the online state observer and compensation algorithm are designed to improve control robust. Finally the influence of virtual road feeling on the handling characteristics is investigated. The objective measures for virtual steering feel is investigated by the analysis of evaluation index for vehicle handling. The results of research can support the relevant scientific theory for the reconfiguration and optimization of steering feel.

线控转向车辆有别于传统刚性连接的机械转向，是通过分布式的电传系统实现动力学的耦合。由于没有机械约束，线控系统面临如何确保转向路感特性精准传递的问题。目前针对转向路感的研究集中在助力特性设计上，对于路感的动力学产生机理以及再重构方法还缺乏系统的理论分析和实验研究。本课题创新性地提出基于线控转向系统的路感参数化重构模型，研究路感重构和操纵性能之间的映射关系，分析满足不同驾驶需求的转向特性优化策略。首先基于前期的转向动力学模型研究，拟解决线控系统多维状态参量下的复杂位置/力矩耦合问题；然后通过转向路感动力学构成因子分析，提出多维的参数化路感在线重构理论；同时针对车辆非线性扰动，提出状态观测器和动态补偿方法以改善控制鲁棒性；最后分析虚拟路感对操纵性能的影响，结合车辆操控性评价指标提出虚拟路感特性的客观度量方法，为转向路感的重构和优化提供科学理论依据。

线传操控技术是未来智能驾驶车辆的核心，线控转向系统通过解耦的电传系统实现动力学的传输，没有了物理上的传动约束。线控系统提供了转向动力学参数化重构的特点，可以优化提升自动驾驶中的底盘稳定性和操控性。项目主要研究成果围绕两方面完成，一方面通过对于路感的动力学产生机理进行理论分析，提出了基于扩展观测器的非线性转向齿条力评估算法。通过相关算法可以实现更加有效的主动转向跟随控制，并为路感的参数化重构提供观测模型。另一方面通过转向路感动力学构成因子分析，提出多维的参数化路感在线重构理论，然后结合车辆操控性评价指标提出虚拟路感特性的客观度量方法，分析了虚拟路感对操纵性能的影响。最后研究成果通过搭建的硬件在环平台完成了验证，测试工况包括直角转弯、蜿蜒道路、比利时道路等典型的极端场景。这些研究成果为智能驾驶车辆上的线控转向系统设计提供方法和数据依据。