"A novel adaptive gesture recognition framework for bionic hands using Stacked Autoencoder (SAE), Adaptive Bayesian Feature Selection (ABFS), MODWT, and Hybrid sEMG-MMG Sensor Modality"

“一种基于堆叠式自编码器(SAE)、自适应贝叶斯特征选择(ABFS)、MODWT和混合sEMG-MMG传感器模态的仿生手自适应手势识别框架”

阅读:1

Abstract

For amputees and those who rely on prosthetic limbs, the dream of natural, intuitive movement is now within reach. This work present an intelligent bionic hand that doesn't just follow commands-it learns, adapts, and responds in real time, almost like a natural extension of the body. By combining muscle magnetic sensing (MMG) and electrical signal detection (sEMG), this system captures even the subtlest muscle movements with incredible precision [1]. But what truly sets it apart is its brain-like adaptability:•Self-Learning AI - Using advanced neural networks, the hand continuously improves its recognition of your unique muscle patterns, whether you're an amputee or not.•Instant Response - With accuracy reaching up to 99.9 % under subject-specific cross-validation in controlled trials and a processing delay as low as 12 ms, movements feel fluid and natural.•Real-World Testing - Tested on 15 participants and evaluated through both ADAMS-MATLAB co-simulation and preliminary hardware experiments with a 3D-printed prototype, the system demonstrates strong feasibility [2]. It is a personalized, evolving solution that bridges the gap between human and machine. While larger-scale and long-term studies are still required, this work paves the way for next-generation prosthetic control systems that restore not only function but also confidence [3].

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。