摘要
为提升用户对人形机器人的接受度与信任度,以生活和商业服务场景中的人形机器人为设计对象,通过探索亲和力外观的评价维度和设计特征,研究该类型机器人外观设计的人工智能(artificial intelligence, AI)模型训练方法。采用感性工学法将亲和力与人形机器人外观设计的各要素相对应;通过问卷调研和量化分析用户对人形机器人的造型、材质及色彩等设计要素的偏好,得出各设计要素特征亲和力分值表,形成精确的设计要求与评价机制;依据亲和力分值表指导训练样本的筛选和制作,开展模型训练,使稳定扩散模型(stable diffusion, SD)能在短时间内生成大量具有亲和力的外观设计方案,结果表明:该方法在提升设计效率、优化设计方案方面有显著效果,为后续人形机器人外观设计提供方法指导。
To enhance user acceptance and trust in humanoid robots,the humanoid robots in life and business service scenarios were taken as the design object to explore AI-driven training approach for their exterior design through exploring the evaluation dimensions and design features of the affinity appearance.The kansei engineering method was used to correlated affinity with each aspect of the robot exterior design.Questionnaires and quantitative analysis were performed on the user preferences for design elements,such as modeling,materials,and color.The affinity-scoring table for each design feature was obtained to establish precise design requirements and evaluation metrics.The model training was carried out according to screening and production of training samples based on affinity score table.The model was constructed and trained to enable stable diffusion(SD)to efficiently produce a vast array of affinity-focused design concepts.These results revealed that this method could significantly promote design efficiency,optimize design scheme and provide methodological guidance for the subsequent design of humanoid robot appearances.
作者
邓俊
李晓煜
袁佳俊
DENG Jun;LI Xiaoyu;YUAN Jiajun(School of Urban Design,Wuhan University,Wuhan 430072)
出处
《机械设计》
北大核心
2025年第10期218-224,共7页
Journal of Machine Design
基金
武汉大学“351人才计划”教学岗位建设项目(WDBZ202234)。
