A Tactile Distribution Sensor Based on Inverse Problem Theory

Nagakubo, Akihiko; Alirezaei, Hassan; Kuniyoshi, Yasuo

doi:10.7210/jrsj.25.960

Cited by 2 publications

(1 citation statement)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, we apply the CT technique to inverse-identification of pressure distribution based on the information from around a tactile sensor. In a previous paper, (11) we established the architecture of the CT tactile sensor and surveyed tactile sensors based on a reconstruction algorithm (12) (13) to develop a new architecture for a tactile sensor. After we designed the CT tactile sensor, we evaluated two reconstruction algorithms suited for it: one is the Algebraic Reconstruction Technique (ART) based on iteration; the other is the numerical solution method for linear algebraic equations (hereafter called LUD because LU decomposition is adopted as the numerical linear algebraic solution).…”

Section: Introductionmentioning

confidence: 99%

Sensing Characteristics of an Experimental CT Tactile Sensor

Sawamoto¹,

Ohka

Zhu

2008

JAMDSM

View full text Add to dashboard Cite

The CT (Computed tomography) method is very effective for making thinner optical waveguide-type tactile sensors. In a previous paper, the authors established the architecture of a CT tactile sensor and evaluated its principle using a series of computer simulations. As a result, the LU decomposition method showed better noise robustness compared to the ART method. On the basis of the simulations, we develop an experimental CT tactile sensor comprised of infrared diodes, phototransistors, an acrylic plate and a black rubber sheet. In the present study, the LU decomposition method is employed and we determine the validity of the algorithm by examining use of a loading machine produced for the CT tactile sensor. In a series of experiments, it is found that absorption of light intensity increases monotonically with increase of applied force and that the present reconstruction method can be applied to not only simply but also multiply connected contact domains. The estimated position of the pressure centroid possesses a rather large error of 4.9 mm, which is 10.8% of a segment of the sensing area.

show abstract

Section: Introductionmentioning

confidence: 99%