基于单层感知结构的五指机械手机器学习研究  

Research on Learning of Five-Finger Robot Based on Single-Layer Perceptual Structure

作　　者：王湘慧[1] 兰浩[2] WANG Xiang-hui;LAN Hao(School of Economics,Trade and Information,Changsha Vocational and Technical College,Changsha 410217,China;College of Engineering and Design,Hunan Normal University,Changsha 410081,China)

机构地区：[1]长沙职业技术学院经济贸易与信息学院,长沙410217 [2]湖南师范大学工程与设计学院,长沙410081

出　　处：《湖南师范大学自然科学学报》2022年第6期130-136,共7页Journal of Natural Science of Hunan Normal University

基　　金：湖南省教育厅重点项目(20A325)。

摘　　要：为了有效减少企业劳动力成本,提高生产产品的效率,有必要设计出具有智能感知性能的五指机械手。本设计采用柔性薄膜压力传感器作为感知器件,结合五指机械手的结构特点实现了传感器布置,选择了单层Perception感知器,设计了一个具有感知、学习、推理和判断能力的机器学习算法,调用TensorFlow库对算法进行优化,使得五指机械手识别率得到明显改善,实验结果显示:优化前算法判断物体形状的准确率为70%,优化后算法的准确率提升了20%。In order to effectively reduce the dependence of enterprises on the equipment and labor,and reduce the price and cost of production products,it is necessary to design a five-finger manipulator with intelligent sensing performance.In this design,the flexible membrane pressure sensor was used as the sensing device,and the sensor layout was realized in combination with the structural characteristics of the five finger manipulator.A single-layer perception sensor was selected,and a machine learning algorithm with the ability of sensing,learning,reasoning and judgment was designed.The Tensor Flow library was called to optimize the algorithm,so that the recognition rate of the five finger manipulator is significantly improved.The experimental results show that the accuracy before optimization is 70%,which is improved by 20% after optimization.

关 键 词：STM32 机械手抓取 PWM 信号调理电路 机器学习 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]