Feature Extraction and Analysis from a Biomimetic Tactile Sensor

仿生觸覺感測器之特徵擷取與分析

古宏麒(Hong-Chi Ku) ； 龔宗鈞(Yan-Chen Liu) ； 劉彥辰(Jui-Yiao Su) ； 蘇瑞堯(Chung-Chun Kung)

《中國機械工程學刊》 35卷4期 (2014/08) Pp. 313-319

摘要

對於人類之靈巧性而言，觸覺感測與訊號處理是絕對必要的。如果要實現像人一樣靈巧的機器手或義肢，機電系統亦不能缺少觸覺感測能力。在本文中，我們探討一手指型仿生觸覺陣列感測器（BioTac）的感測訊號作特徵擷取及辨識處理。此感測器具備模仿人的手指皮膚感覺能力，可同時提供接觸壓力、微振動以及熱傳通量之資訊。當感測器與物體接觸時，所有觸覺資訊是協同作用的且非線性。當觸覺感測器沿著待測物表面滑動，我們藉由所擷取到的觸覺資訊進行演算法和機器學習技術的開發及驗證。

並列摘要

Tactile sensing and signal processing is necessary for human dexterity and is likely to be required in mechatronic systems such as robotic and prosthetic limbs if they are designed to achieve similar dexterity. In this paper, we discuss feature extraction and interpretation of the modality data from a finger-shaped biomimetic tactile sensor (BioTac) which provides simultaneous information about contact forces, microvibrations and thermal fluxes, mimicking the full cutaneous sensory capabilities of the human finger. The data produced by these sensory modalities are synergistic and non-linear during sensor-object interaction. Algorithms and machine learning techniques were developed and validated for extracting the cutaneous information during sliding motion against the surface of object.

並列關鍵字

biomimetic tactile sensor ； machine learning ； tactile feature extraction