基于改进BP神经网络的无人驾驶汽车防抱死制动控制系统设计
2025,33(2):103-109
摘要：无人驾驶汽车的状态,如速度、载荷、重心高度等会对制动效果产生影响。车辆状态信息较多且包含噪声信号干扰,导致信息采集精度较差,会导致制动效果不稳定,甚至出现车轮抱死的情况。为此,设计基于改进BP神经网络的无人驾驶汽车防抱死制动控制系统。系统硬件中设计DSP处理器,实现信号的高速处理并生成控制指令,通过CAN实现通讯功能；设计执行模块执行DSP处理器的控制指令；通过采样模块实现无人驾驶汽车防抱死制动信号的采样。在软件设计中,设计引导滤波信号去噪算法,实施防抱死制动信号的去噪处理,获取汽车驾驶信息数据；利用LM算法寻找函数值最小的对应参数向量,获得辨别误差局部最小的权值,改进BP神经网络,设计基于改进BP神经网络的PID控制算法,输出无人驾驶汽车防抱死制动控制指令。实验结果表明,设计系统紧急制动工况下的控制十分稳定,能够很快达到最大制动力,控制反应迅速,主缸压力值上升较快,轻微制动工况下的系统随动性较强。
关键词：DSP处理器;LM算法;改进BP神经网络;无人驾驶汽车;防抱死制动控制;
Design of Autonomous Vehicle Anti lock Braking Control System Based on Improved BP Neural Network
Abstract：The state of an autonomous vehicle, such as speed, load, and center of gravity height, can have an impact on the braking effect. The vehicle has a large amount of status information and contains noise signal interference, resulting in poor information collection accuracy, unstable braking effect, and even wheel lockup. To this end, an anti lock braking control system for autonomous vehicles is designed based on an improved BP neural network. Design a DSP processor in the system hardware to achieve high-speed signal processing and generate control instructions, and achieve communication functions through CAN; Design an execution module to execute the control instructions of the DSP processor; The sampling module is used to sample the anti lock braking signal of autonomous vehicles. In software design, design guidance filtering signal denoising algorithms, implement denoising processing of anti lock braking signals, and obtain driving information data of automobiles; Using the LM algorithm to find the corresponding parameter vector with the smallest function value, obtain the weight value of the local minimum discrimination error, improve the BP neural network, design a PID control algorithm based on the improved BP neural network, and output the anti lock braking control command for unmanned vehicles. The experimental results show that the control of the designed system under emergency braking conditions is very stable, and it can quickly reach the maximum braking force. The control response is rapid, and the pressure value of the master cylinder increases rapidly. The system has strong follow-up under mild braking conditions.
Key words：DSP processor; LM algorithm; Improved BP neural network; Autonomous vehicles; Anti lock braking control;
收稿日期：2023-12-01
基金项目：地方应用型本科高校大数据管理与应用专业多元化人才培养体系与实践研究 编号：SGH22Y1867
