摘要
为了降低超声测距系统因温、湿度差异带来的误差,在Matlab2010平台下,运用BP神经网络对标定的温、湿度样本及对应的超声波速度进行融合训练仿真实验,并将设计的补偿算法移植到STM32测距系统,测试其补偿能力.经多次测距实验表明,与现有方法相比,能有效补偿温、湿度变化引起的误差,精度提高了1个数量级.
In order to reduce the ultrasonic ranging system error due to temperature and / or humidity difference in Matlab2010 platform,and to calibrate the temperature and humidity samples and hence the corresponding ultrasonic velocity with Fusion training simulation experiment through BP neural network,the design of the compensation algorithm is transplanted into STM32 ranging system to test its compensation ability. The current experiment showed that compared with existing methods,STM32 ranging system effectively compensates the error due to temperature and / or humidity difference,resulting in the improvement of precision to one order of magnitude.
出处
《琼州学院学报》
2016年第2期18-22,共5页
Journal of Qiongzhou University
基金
安徽省教育厅医电子仪器与维护省级特色专业质量工程项目(20101459)
安徽省教育厅医用电子仪器省级示范实习实训中心项目(2011131)
安徽省教育厅重点教研项目(2015jyxm535)
关键词
数据融合
神经网络
超声波
data fusion
neural network
ultrasound