Optical-based sensing of shear strain using reflective color patterns

Abstract

There is an increasing need to measure shear stress and strain for biomedical and robotics applications. Many shear sensors exist, but are often impractical for these applications as they can be bulky, require wired power sources, or are sensitive to electromagnetic interference from the human body or mechatronic systems. This paper presents an optical-based sensor, which measures shear strain based on a red, green, and blue (RGB) light-emitting diode (LED) projecting a cyclical pattern of RGB light onto a color pattern. As shear strain causes displacement between the LED and color pattern, the relative intensities of the reflected RGB lights change due to the shift of the color pattern position. A photoresistor captures the reflected light intensity during each color illumination, allowing determination of shear strain along two axes. The purpose of this work was to determine the optimal sensor configuration by studying the efficacy of different color patterns and data classification algorithms to measure two-axis shear strain. The optimal sensor configuration was a randomized color pixel pattern with a weighted k-nearest-neighbor classifier. Accuracy was 97.5% and 98.0%, and misclassification cost was 0.11 mm and 0.07 mm in the horizontal and vertical directions, respectively. Importantly, this sensing technique allows determination of directionality of shear measurement, overcoming a major limitation of other optical-based shear sensors. These properties, along with its small, low-cost, and scalable design support the use of this sensor and sensing paradigm for a variety of biomedical and robotics applications