基于电气参数的电驱动车辆行驶路面附着感知与控制研究

Road perception and adhesion control is critical to implement active safety control for ground vehicles. The traditional solution is based on the slip ratio and adhesion coefficient, which, however, are difficult to obtain, and usually with poor precision and limited bandwidth. Compared with traditional internal combustion engine vehicles (ICEVs), electric-drive vehicles have electrical signals which are easy to measure with high precision and wide bandwidth, inspiring a novel technical route to advance adhesion perception and control. In this research, we will analyze the dynamic behaviors of electrical signals (i.e., current, voltage, and electric torque) under complex road conditions to reveal the relations between electrical signals and complex factors, including road conditions, motor control strategies, and road slope. Then, we will propose a novel identification model of road adhesion states using the on-line sequential extreme learning machine (OS-ELM) methodology. Finally, considering the convex feature of the traction motor’s load current during acceleration, we will propose an optimal adhesion control methodology based on the sliding mode extreme seeking control (SM-ESC) to improve both the stability and the acceleration performance under unknown road conditions. This research only uses the electrical signals from internal sensors, which may achieve high precision, low cost, and strong reliability. It will enhance vehicle’s environmental perception ability significantly, and advance the active safety technologies of electric-drive vehicles.

行驶路面附着条件感知与控制是车辆主动安全系统的关键。而传统方法大多基于滑移率、附着系数等参数来辨识路面条件和实施附着控制，这些参数精度低，带宽有限，且在实际应用中不易获取。与传统车辆相比，电驱动车辆的电气参数具有带宽高、测量简单精确等优点，为这一问题提供了新的研究思路。本项目分析复杂路面附着条件下电气参数（电流、电压、电磁转矩）的动态变化特性，研究变化路面条件、电机控制策略、路面坡度等多因素与电气参数的耦合关系；采用在线序贯极限学习机方法建立路面附着状态的快速训练、全局最优感知模型；针对驱动轮打滑过程中电机负载电流变化的凸函数特性，应用滑模极值搜索算法实现未知路面条件下的最佳附着点控制，保证驱动轮稳定性的同时提升牵引性能。本项目提出的路面附着感知与控制方法只需利用车辆内部电驱动传感器的电气参数，精度高、成本低、可靠性好，有助于提高电动车辆的环境感知能力，推动主动安全控制相关基础研究的发展。

本项目针对电驱动车辆的复杂行驶路面附着条件感知与控制问题，挖掘利用电驱动车辆电气参数带宽高、测量精确等优点，研究了电驱动车辆电气参数在复杂路面附着条件下的动态行为特性，实现了基于数据驱动的附着稳定性的一致感知模型，重点揭示了变化附着路面条件下的附着稳定性表征方法和机理，利用多样化路面附着条件下的附着系数-滑移率存在的一致性凸函数特性，采用滑模极值搜索算法实现了未知路面条件下的自适应的最佳附着点控制。项目的主要成果包括提出并定义了附着稳定参数，实现了对多样性及变化附着路面的附着稳定特性的定量表征，基于数据驱动的极限学习机方法实现了附着稳定参数的辨识，基于力传递模型分析了变化附着路面条件下的表征机理和判别方法，实现了未知路面条件下的最佳附着工作点的自寻优。本项目发展了电驱动车辆的轮-路附着稳定性感知与控制的相关基础方法，推动了快响应、高精度的新型电驱动车辆主动安全控制的方法创新，有望在复杂路面附着条件下的电驱动车辆（包括新能源汽车、移动机器人、轨道电驱动车辆等）的新型主动安全和高性能运动控制等领域得到应用。