Soft 3 Axis Force Sensor for Surface Quality Control of Non

Soft 3 Axis Force Sensor for Surface Quality Control of
Non-Ferromagnetic Products
Damith Suresh Chathuranga, Zhongkui Wang, and Shinichi Hirai
Department of Robotics, Ritsumeikan University, Kusatsu, Shiga 525-8577, Japan
Abstract— Quality controlling the surface finish in wood
products, painted surfaces, and metal castings ect. is an important step in manufacturing process. Presently most of the
surface quality is assessed by trained factory personal. They
feel the soft or roughness of the product surface by sliding
their hands over the product surface. Tactile awareness of
hands is used in deciding if the surface has acceptable surface
finish to pass the quality control check. This process is highly
individualized and worker decide the quality by his or her own
judgement. Automating such tasks is difficult as tactile systems
that are reliable and cheap are rare whilst standards for surface
quality in terms of easily measurable values are difficult to
define. We believe that if a system had the ability to classify
and distinguish materials and textures, it has the potential to be
used in assessing surface finish and be used in quality control
tasks. We propose a new three axis tactile sensor, based on
magnetic flux measurements, to obtain three dimensional tactile
data. This data is then used in a texture classification algorithm
utilizing support vector machine (SVM) classifier. Frobenius
norm calculated from the covariance matrix of the above tactile
data and the three means of the three dimensional data were
used as features. Palpation velocity and small vertical load
variances had minimum influence on the proposed algorithm,
making it robust and utilizable to industrial quality control
tasks.
The proposed soft 3D force sensor, calculate force and
sense vibrations. It uses three Hall Effect sensors orthogonally placed near a cylindrical soft beam made of silicon
rubber. A niobium permanent magnet is embedded in the
silicon. When a force is applied to the free end of the
cylinder, it is compressed and bent displacing the magnet.
This displacement causes change in the magnetic field near
the ratiomatric linear (Hall effect) sensors. By detecting
these magnetic field changes, the position of the magnet is
calculated. Next, using the spring theory and bending theory,
forces in three directions are calculated. Read [1] to further
understand the sensor operation and the characteristics.
The proposed algorithm classifies textures, based on three
dimensional tactile data obtained by 3D soft tactile sensors.
The covariation between the tactile data of the three dimensions is used as a feature in the classification. Support
Vector Machine (SVM) was used as the classifier. This study
describes experiments showing that this classification method
is robust and applicable to any system that could detect tactile
data in three dimensions.
The sensor and the classification algorithm are combined
to be used in the surface finish assessing system. Initially, the
system is trained using the samples that have the acceptable
surface finish. Next the system examines the products that
need quality check and if the surface finish is not similar
to the standard quality it is rejected. The products can be
classified to different grades of quality depending on the
method the system was initially trained.
(a)
(b)
(c)
Fig. 1. (a) Tactile sensor, (b) Materials used for classification, (c) Gaussian
ellipsoids of the data obtained for the eight materials and textures. The
centers of the ellipsoids are the mean tactile sensor values in the x, y and
z directions and the length of the semi-principal axis is the covariance
values of the sensor signal. It is observed that the ellipsoids are spread
apart depending on the texture.
R EFERENCES
[1] D. S. Chathuranga, Z. Wang, Y. Noh, T. Nanayakkara and S. Hirai,
Disposable Soft 3 Axis Force Sensor for Biomedical Applications,
The 37th Annual International Conference of the IEEE Engineering
in Medicine and Biology Society (EMBS), 2015.