📝 SummarXiv

Abstract

Bio-inspired whisker sensors are employed in diverse applications such as fluid-flow sensing, texture analysis, and environmental exploration. However, existing designs often face challenges related to durability, fabrication complexity, and response consistency. To address these issues, we propose a modular architecture that decomposes whisker sensors into five functional components: the whisker element (WE), compliant element (CE), sensing element (SE), support structure (SS), and data acquisition module (DAQ). We develop and compare four in-house sensor designs built using this architecture, each differing in material choice, sensing modality, and mechanical structure. To unify heterogeneous sensor outputs, we introduce a calibration strategy that maps raw sensor readings-whether from pressure, magnetic flux, or visual features-into a common representation: the bending moment at the whisker base. This representation supports consistent interpretation and comparison across sensing techniques. We adopt texture roughness analysis as a representative sensing task to evaluate design trade-offs. Each whisker sensor's frequency-domain response is benchmarked against a high-resolution laser sensor using standardized roughness specimens. Empirical results show that rigid whiskers improve accuracy in texture classification, while flexible whiskers provide robustness for exploratory robotics tasks. Among the evaluated designs, the Hall-effect sensor with a rubber CE demonstrates the most favorable balance of durability, reconfigurability, and fabrication simplicity.