Tactile tomographic imaging using robotic whiskers

Robotic Whiskers;Bio-inspired Signal Processing;Tactile Sensing;Tomography;Shape Recognition;Flow\rimaging;Inverse problem;Image Reconstruction

Tuna, Cagdas

关键词: Robotic Whiskers;    Bio-inspired Signal Processing;    Tactile Sensing;    Tomography;    Shape Recognition;    Flow\rimaging;    Inverse problem;    Image Reconstruction;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/49680/Cagdas_Tuna.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Animals such as rats and seals sense through movement by oscillating their whiskers back and forth to extract environmental information including nearby object features and fluid-flow characteristics. However, current array sensory systems do not fully utilize tactile sensing strategies extensively used by these animals. This dissertation focuses on developing advanced bio-inspired signal processing algorithms for the reconstruction of surroundings with arrays of vibrissal sensors. Inspired by the oscillatory motions of the rat’s whiskers, we introduce a new tactile tomographic imaging model using robotic metal whiskers to extract object features including size, shape and location, and map out the cross-sectional fluid-flow characteristics via tomographic imaging. Comparing the whisker position at the very initial object contact to a ray path in X-ray tomography, we show that the problem of object shape recognition with robotic whiskers can be expressed as a 2-D tactile tomographic imaging procedure by using only the whisker base position and the angle at the whisker base during the very initial contact recorded at different locations around the object.At high Reynolds numbers, the drag force on a whisker segment is proportional to the relative velocity squared and hence, whether the flow is laminar or turbulent, we propose that it is possible to map out the 2-D cross-sectional mean fluid-flow velocity field using the moment measurements collected by a robotic whisker array from different directions for tomographic reconstruction. We also present a linear state-space formulation for the tactile dynamictomographic fluid-flow imaging for the sequential estimation of the fluid-flow characteristics in a dynamically changing environment.The experimental results strongly demonstrate that this new tactile sensing technology developed in this dissertation may find a potential future use in variousrobotic applications including object feature extraction, object tracking, underwater navigation, high-resolution fluid-flow imaging, source localization and environmental mapping.