“TACTILE OBJECT RECOGNITION ROBOTIC SYSTEM USING SUPERVISED MACHINE LEARNING”
Kamel, M., El-Sayed, A., Awad, M., Abou Elmagd, G. (2023). “TACTILE OBJECT RECOGNITION ROBOTIC SYSTEM USING SUPERVISED MACHINE LEARNING”. EKB Journal Management System, 42(2), 87-99. doi: 10.21608/jaet.2022.100646.1125
Marwa Mohamed Kamel; Abu Hashema Mostafa El-Sayed; Mohammed Ibrahim Awad; Gamal Eldin Ali Abou Elmagd. "“TACTILE OBJECT RECOGNITION ROBOTIC SYSTEM USING SUPERVISED MACHINE LEARNING”". EKB Journal Management System, 42, 2, 2023, 87-99. doi: 10.21608/jaet.2022.100646.1125
Kamel, M., El-Sayed, A., Awad, M., Abou Elmagd, G. (2023). '“TACTILE OBJECT RECOGNITION ROBOTIC SYSTEM USING SUPERVISED MACHINE LEARNING”', EKB Journal Management System, 42(2), pp. 87-99. doi: 10.21608/jaet.2022.100646.1125
Kamel, M., El-Sayed, A., Awad, M., Abou Elmagd, G. “TACTILE OBJECT RECOGNITION ROBOTIC SYSTEM USING SUPERVISED MACHINE LEARNING”. EKB Journal Management System, 2023; 42(2): 87-99. doi: 10.21608/jaet.2022.100646.1125

“TACTILE OBJECT RECOGNITION ROBOTIC SYSTEM USING SUPERVISED MACHINE LEARNING”

Journal of Advanced Engineering Trends
Volume 42, Issue 2, July 2023, Page 87-99  XML PDF (1.46 MB)
Document Type: Original Article
DOI: 10.21608/jaet.2022.100646.1125
View on SCiNiTO View on SCiNiTO
Authors
Marwa Mohamed Kamel email 1; Abu Hashema Mostafa El-Sayed2; Mohammed Ibrahim Awadorcid 3; Gamal Eldin Ali Abou Elmagd4
1Mechatronics and Industrial Robotics Program/ Faculty of Engineering/ Minia University/ Minia
2Electrical Engineering Department Faculty of Engineering, Minia University EI Minia.
3Mechatronic Engineering, Faculty of Engineering, Ain Shams University, Cairo, Egypt
4Production Engineering and Design Department Faculty of Engineering Minia University
Abstract
This paper aims to perform shape estimation actions on 2D uniform objects such as Square, Circle, Triangle and Pentagon to predict the actual shape of the tested objects. A tactile fingertip has been used to construct the shape that encloses some pressure sensor elements (called Taxels). One advantage of this study is the taxel`s number used, which is fewer than any other types of tactile sensor in the previous related studies as well as the new proposed exploratory technique. Moreover, the collected datasets have been used as an input for the three different learning classification algorithms. k-Nearest Neighbors Classifier (KNN), Naïve Bayes and Support Vector Machine (SVM) have been implemented as supervised learning algorithms to recognize the desired object shape from the collected data. As a result, the best performance obtained with SVM is by using the Radial Basis Function (RBF) that gives an average of 96.3% accuracy in shape recognition. Not only that, another performance comparison is made by smalling the scanned area of the same tested objects; the square and circle shapes are explored in the new area because of its lower recognizing performance (94.85 % and 94.71% respectively). Thus, enhancing the accuracy to be 96.03% and 98.8% respectively which is a remarkable performance.
Keywords
Keywords— Tactile sensor; Object recognition; Classification algorithms; shape; classifier
Supplementary Files
Statistics
Article View: 168
PDF Download: 205