基于GRNN的糖尿病足足底压力信号检测系统设计  

Design of diabetes foot’s planter pressure signal detection system based on GRNN

作　　者：吴玉广 张文栋 桑胜波 杨琨 阳佳 李晟嘉 WU Yuguang;ZHANG Wendong;SANG Shengbo;YANG Kun;YANG Jia;LI Shengjia(Key Laboratory of Micro Nano Sensors&Artificial Intelligence Perception Shanxi Province,College of Information and Computers,Taiyuan University of Technology,Jinzhong 030024,China;R&D Department,China Academy of Launch Vehicle Technology,Beijing 100076,China)

机构地区：[1]太原理工大学信息与计算机学院、微纳传感与人工智能感知山西省重点实验室,山西晋中030024 [2]中国运载火箭技术研究院研究发展部,北京100076

出　　处：《电子设计工程》2023年第9期84-89,共6页Electronic Design Engineering

摘　　要：信号采集系统在实际应用中,会受到传感器本身及外界温湿度等内外因素的影响,导致系统采集到的数据存在较大误差。该文基于STM32单片机对传统压力采集电路结构进行了改进设计,并采用了广义回归神经网络(GRNN)拟合算法,研制出一套针对多传感采集(1 024个)及采样频率为100 Hz的足底信号检测系统。文中对传感阵列施压测得电导,基于最小二乘法和GRNN两种算法在Matlab中对压-导值进行拟合对比,结果显示,应用了GRNN力-导映射关系的足压检测系统的相对误差为2.89%,相较于最小二乘法,降低了9.11%。由此可得,采用GRNN拟合关系的压力信号检测系统生成的压力云图,可以更好地反应检测者的足底压力分布,能够为糖尿病足病况分析提供可靠的测试平台和高精度的环境。In practical application,the signal acquisition system will be affected by the sensor itself and external factors such as temperature and humidity,in turn,lead to large errors in the data.This study,which based on STM32 microcontroller to improve the design of the pressure acquisition system structure and adopts the Generalized Regression Neural Network(GRNN)to curve fitting,develops a planar signal acquisition system for 1024 sensors units and the sampling frequency of 100 Hz.This paper uses cond⁃uctance value from the sensor array to fit and compare in Matlab based on the above two algorithms.The results show that the relative error of foot pressure measurement system applied GRNN force⁃conductance mapping is 2.89%,which is 9.11%lower than the least square method.The pressure cloud map,generated by the pressure signal detection system based on GRNN fitting relationship,can better reflect the plantar pressure distribution of the tester and provide a reliable testing platform and high⁃precision environment for the analysis of diabetic foot conditions.

关 键 词：压阻式传感阵列 足底压力检测 传感器标定 GRNN 曲线拟合 

分 类 号：TN06[电子电信—物理电子学]