腳底壓力辨識系統對於受測者在不同負重支撐點與重量之分析與研究

Abstract

近年來，隨著物聯網應用的興起，網絡通訊不只侷限在手機與電腦間，除了帶來人類生活的便利外，資訊安全的議題也逐漸被重視，因而延伸出具唯一性的生物識別技術，生物辨識的簽名認證有別於傳統的文字或圖像式的帳號與密碼，其不易被偽造的特性也使得安全程度變得更為可靠。在過去的腳底壓力分析的研究中，比較少有提及與探討受測者在身體不同位置處攜帶負重，對於受測者攜帶不同重量的負重的研究也較無著墨。本論文主要在於探討受測者在赤腳情況下於，右側攜帶不同重量的負重與後側攜帶不同重量的負重對於搭配機器學習的腳底壓力感測技術的辨識度和模型訓練時間的影響的分析與研究。實驗結果顯示使用平均腳底壓力資料與攜帶大負重量會提升腳底壓力的辨識率。

In recent years, with the rise of Internet of Things applications, network communication is not limited to mobile phones and computers. In addition to bringing convenience to human life, the issue of information security has gradually been paid attention to. Therefore, a unique biometric technology is extended. Biometric signature authentication is different from traditional text or image-based account numbers and passwords. Its feature that not easy to be forged also makes the degree of security more reliable. In the past research on the analysis of foot pressure, few studies dis-cussed that subjects carried weights at different positions of the body. There are also few studies in research of subjects with carrying different weights. This thesis is mainly to explore the impact of the plantar-pressure recognition systems’ accuracy and the model training time for the subjects who carry different weights on the right side or the back side under barefoot conditions with machine learning methods analysis the subjects which carrying different weights on the right side and the subjects which carrying different weights on the back under barefoot conditions. The experimental results show that using average gait's data and carrying heavy weights will in-crease the recognition rate of gait's pressure.