基于STM32的主动激励式触觉传感器滤波算法对比

Comparison of actively excited haptic sensor filtering algorithms based on STM32

本文通过基于STM32F407ZGT6的单片机作为主控芯片,利用三极板式压电陶瓷作为主动激励式触觉传感器,镶嵌于刚性物体,置于机械手指尖处对不同物体的数据进行采集;通过加法电路和放大电路的处理,将数据存储于SRAM中。纯硬件进行数据采集时,会由于电磁干扰、机械手散热装置等原因产生一些干扰数据,所以在进行主动激励式触觉传感器识别算法的研究之前,要进行数据处理。本课题通过对一阶滤波、中值滤波和五点三次平滑处理等方法进行比较,选择最适合本系统的数据预处理方法,应用于STM32F407ZGT6单片机中。

Based on STM32 F407 ZGT6 single chip microcomputer, the paper uses the three pole plate piezoelectric ceramic as the active excitation tactile sensor embedded in the rigid object, which is placed at the tip of the mechanical finger to collect the data of different objects. Then the paper stores the data in SRAM through the processing of addition circuit and amplification circuit. When data acquisition is carried out with pure hardware, some interference data will be generated due to electromagnetic interference, manipulator heat dissipation device and other reasons. Therefore, data processing should be carried out before the research of active excitation tactile sensor recognition algorithm. Common data preprocessing methods include first-order filtering, median filtering and five-point cubic smoothing. By comparing these three filtering methods, this paper selects the most suitable data preprocessing method for this system and applies it to STM32 F407 ZGT6 single chip microcomputer.