基于神经网络的ESC线控制动压力估计与控制  

Estimation and Control of ESC Brake-by-wire Pressure Based on Neural Network

作　　者：邵东 尹思维 李亮[1] 王翔宇 魏凌涛 周道林 SHAO Dong;YIN Siwei;LI Liang;WANG Xiangyu;WEI Lingtao;ZHOU Daolin(State Key Laboratory of Automotive Safety and Energy,Tsinghua University,Beijing 100084;DFAC Commodity Research and Development Institute,Wuhan 430056)

机构地区：[1]清华大学汽车安全与节能国家重点实验室,北京100084 [2]东风汽车股份有限公司商品研发院,武汉430056

出　　处：《机械工程学报》2025年第4期219-228,共10页Journal of Mechanical Engineering

基　　金：山东省重大科技创新工程(2019TSLH0701);博士后创新人才支持计划(BX20200184)资助项目。

摘　　要：汽车电子稳定性控制(Electronic stability controller,ESC)作为车辆主动安全领域的关键技术之一,不仅能在车辆制动抱死、驱动滑转、转向侧滑等工况下保障车辆处于安全状态,而且支持高级驾驶辅助系统(Advanced driving assistance system,ADAS)的线控制动需求。传统ESC的线控制动功能往往需要加装轮缸压力传感器从而实现压力闭环控制,为了降低基于ESC的线控制动产业化成本,采用基于BP神经网络的ESC线控制动压力估计方法,通过分析ESC线控制动过程中的压力控制模式,利用BP神经网络在多维度数据处理上具有拟合能力强的优势,在不同控制模式下对ESC液压特性进行建模,拟合后的模型可在给定的ESC电磁阀与电机控制指令下进行轮缸压力估计。同时提出一种基于逻辑门限选择的PI反馈压力控制策略,利用压力估计结果进行反馈控制,提高线控制动的压力控制精度。在无轮缸压力传感器的基础上,对搭载ESC控制器与泵体的硬件在环台架进行测试。最终试验结果表明,ESC线控制动压力估计方法具有较高的精度,满足反馈控制需求,同时压力控制的精度在0.3 MPa以内,满足ADAS和高级自动驾驶的线控制动需求,为低成本线控制动方案提供一种新思路。Automotive electronic stability controller(ESC)is one of the key technologies in the field of vehicle active safety.It can not only ensure that the vehicle is in a safe state under conditions such as vehicle braking lock,driving slip,and steering sideslip,but also support control-by-wire demand of advanced driving assistance system(ADAS).Traditional ESC’s control by wire function often requires the installation of wheel cylinder pressure sensors to achieve pressure closed-loop control.In order to reduce the industrialization cost of brake-by-wire based ESC,the pressure estimation method of ESC based on BP neural network is adopted.By analyzing the pressure control mode in the process of ESC brake-by-wire,the hydraulic characteristics of ESC are modeled under different control modes by taking advantage of the strong fitting ability of BP neural network in multi-dimensional data processing,the fitted model can estimate the wheel cylinder pressure under the given ESC solenoid valve and motor control command.At the same time,a PI feedback pressure control strategy based on logical threshold selection is proposed,which uses the pressure estimation results for feedback control to improve the pressure control accuracy.Based on without wheel cylinder pressure sensor,the hardware-in-the-loop bench equipped with an ESC controller and a pump body is tested.The final experimental results show that the ESC brake-by-wire pressure estimation method has high accuracy and satisfies feedback control.At the same time,the accuracy of pressure control is within 0.3 MPa,which meets the demand for brake-by-wire of ADAS and advanced automatic driving,and provides a new idea for low-cost brake-by-wire solutions.

关 键 词：电子稳定性控制 BP神经网络 线控制动 压力估计与控制 

分 类 号：U461[机械工程—车辆工程]