应用于列车完整性检测系统的改进型滤波器

Application of Modified Filter in Train Integrality Detection System

列车完整性检测系统是新型车载设备,对防止列车抛车有重要意义。其硬件构架上以MSP430微处理器为主控制器,由ADXL202传感器实现加速度信息采集。作为判据的加速度信息在复杂的行车条件下被干扰。本文以加速度信息为研究对象,进行理论分析、数学建模和结果仿真。文中设计了一个二维的改进型滤波器。主要采用最小二乘滤波思想,选取适当的加权矩阵和遗忘因子,它不需要被估计量和观测量的统计规律,不易发散,完全符合观测数据被随机干扰,难以建立精确模型的特点。为了进一步提高滤波精度在滤波器中设计了一个野值剔除环节,在前期去掉粗大误差。对系统采集的数据进行上述的滤波,显著提高了数据的可靠性。

The train integrality detection system is a new type of on-hoard equipment. Its application greatly alleviates the trouhle of train separateness. The hardware design consists of Microprocessor MSP430 as the main controller and Sensor ADXL202 as the device to complete acceleration information collection. The acceleration information serves as the criteria for detection. Complicated transportation conditions badly interfere with the acceleration information. This paper takes the acceleration information as the research object and deals with the related theoretical analysis, mathematical modeling and result simulation. An improved two-dimension filter is designed. The least square filtering theory is applied. The weighting matrix and restricted memory are suitably selected. The filter does not require statistical rules of estimated and observed quantities and is not easy to filter divergence. The filter is featured by random interference with observation data and by difficult establishment of an accurate model. In order to upgrade filtering precision, the outlier eliminating function is designed to get rid of gross errors ahead of the least square filtering. It is illustrated that filtering the data collection results in better data reliability.