基于sEMG的以人为中心的人机协同量化模型研究
Study on Human-centric Human-robot Collaboration Quantization Model Based on sEMG
戚志政 1孙守迁 2郑刚强 3陈力 1纪国伟 1邓志吉4
作者信息
- 1. 浙江大华技术股份有限公司,杭州 310051
- 2. 浙江大学,杭州 310027
- 3. 武汉理工大学,武汉 430070
- 4. 浙江省视觉物联融合应用重点实验室,杭州 311251
- 折叠
摘要
为了对企业不同类别接触类产品进行较为客观的人机工效实验检测和评价,提出了基于表面肌电(sEMG)信号的一种可广泛应用于制造企业产品人机工效评估的通用量化方法.通过中型枪机支架的安装实验,利用时间特征、肌肉疲劳特征、时间特征差值计算方法分别在工作效率,工作负荷,可学习性三个方面实验检测.根据模型的评价指标对实验数据进行了比较与分析,新支架在枪型摄像机壁装过程中工作效率提升43.8％,总负荷降低-70.0％,可学习性提升65.0％,综合评分提升33.5分,人机工效显著提升.结果表明,人机协同量化模型是有效的,通过客观可视化数据弥补了主观经验决策的不足,作为产品研发初期的指导,同时最大限度的提高了工人操作的自主权和人机协同效率.
Abstract
To test and evaluate the ergonomics of different types of contact products,a general quantitative method based on surface electromyography(sEMG)signal is proposed,which can be widely used in the evaluation of ergonomics of products in manufacturing enterprises.Through the installation experiment of medium gun stand,the time feature,muscle fatigue feature and time feature difference calculation method are used to test the work efficiency,workload and learnability,respectively.According to the evaluation index of the model,the experimental data were compared and analyzed.The new bracket increased the working efficiency by 43.8％,the total load was reduced by-70.0％,the learnability was increased by 65.0％,the comprehensive score was increased by 33.5 points,and the ergonomic efficiency was significantly improved during the process of gun camera wall installation.The results show that the human-robot collaboration quantification model is effective.It makes up for the deficiency of subjective empirical decision making through objective visualization data and serves as a guide for the initial stage of product development.At the same time,it maximizes the autonomy of workers in operation and the efficiency of human-robot collaboration.
关键词
人机协同/肌电信号/疲劳评价/人机工效/产品开发Key words
human-centric human-robot collaboration(HHRC)/surface electromyography(sEMG)/fatigue evaluation/ergonomics/product development引用本文复制引用
基金项目
浙江大华技术股份有限公司领军专家课题(DH3.4734)
出版年
2024
NETL
NSTL
万方数据