摘要
路面冰层识别是准确估算路面附着系数的关键技术,路面冰层与路面温度和太阳辐射(映射于地理位置、时刻、空气温度)三者之间存在稳定的非线性关系,以路面温度、地理位置、时刻及空气温度为输入,采用神经网络间接识别路面冰层。实验中,以30 d的1 440个时刻的数据训练BP神经网络,以2 d的96个时刻的数据验证BP神经网络,路面冰层识别结果正确。
Black ice detection is very important to tire-road friction coefficient estimation.There is nonlinear causality between black ice,road temp and solar radiation（mapped to geographical location,time and air temp）,thus,road condition can be detected indirectly by road temp,geographical location,time and air temp with BP neural network.In experiment to detect black ice,BP neural network was trained with 1 440 group data and validated by 96 group data,the method of black ice detection is accurate.
出处
《仪表技术与传感器》
CSCD
北大核心
2010年第11期74-75,78,共3页
Instrument Technique and Sensor
基金
中国博士后科学基金(20090460312)
