基于PNN的汽车ABS系统中压力调节器和轮速传感器的故障诊断

Fault?diagnosis?of pressure regulator?and??wheel?speed sensor for the?anti-lock?braking?system?of?automobiles?based?on?probabilistic?neural?network

摘要:

在汽车防抱死制动系统(ABS)中，压力调节器和轮速传感器起着非常重要的作用，为了进一步完善汽车防抱死制动系统的制动性能，文中提出一种基于概率神经网络(PNN)的压力调节器和轮速传感器的故障诊断方法。基于高附着均一路面，起车时制动及单一的压力调节器或者轮速传感器故障的试验数据，分别建立了基于概率神经网络的压力调节器故障诊断模型和轮速传感器故障诊断模型，并与BP神经网络进行了比较。仿真结果表明，利用相同的训练样本集对概率神经网络和BP神经网络进行训练时，基于概率神经网络的压力调节器故障诊断模型和轮速传感器故障诊断模型在训练时间和诊断精度上明显优于BP神经网络，并且利用测试样本对建好的压力调节器故障模型和轮速传感器故障模型进行检测时，无论测试样本的顺序发生什么变化，基于概率神经网络的故障模型都能够准确的进行故障识别。

Abstract:

Pressure?regulator?and?wheel?speed?sensor?play?an?important?role?in?the?anti-lock?braking?system?of?automobiles.?In?order?to?further?improve?the?braking?performance?of?the?anti-lock?braking?system?of?automobiles,?a?fault?diagnosis?method?of?pressure?regulator?and?wheel?speed?sensor?based?on?probabilistic?neural?network?is?proposed.?Based?on?the?test?data?of?braking?and?single?pressure?regulator?or?wheel?speed?sensor?faults?when?start-up?on?high?adhesion?uniform?road?surface,?the?fault?diagnosis?models?of?pressure?regulator?and?wheel?speed?sensor?based?on?probabilistic?neural?network?are?established?respectively,?and?compared?with?BP?neural?network.?The?simulation?results?show?that?when?the?probabilistic?neural?network?and?BP?neural?network?are?trained?with?the?same?training?sample?set,?the?fault?diagnosis?model?of?pressure?regulator?and?wheel?speed?sensor?based?on?probabilistic?neural?network?is?obviously?superior?to?BP?neural?network?in?training?time?and?diagnostic?accuracy,?and?When?the?fault?models?of?pressure?regulator?and?wheel?speed?sensor?are?detected?by?the?test?sample,?the?fault?models?based?on?probabilistic?neural?network?can?accurately?identify?the?faults?no?matter?what?the?order?of?test?samples?changes.