Perencanaan Gerakan Tendangan Robot Humanoid Sepak Bola Berdasarkan Kurva Bezier Kubik Dengan Optimasi Neural Network

Rizqifadiilah, Muhammad Azriel (2022) Perencanaan Gerakan Tendangan Robot Humanoid Sepak Bola Berdasarkan Kurva Bezier Kubik Dengan Optimasi Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 07111840000221-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2024. Download (5MB) | Request a copy

Abstract

Robot sepak bola humanoid merupakan robot yang memiliki kemampuan bermain sepak bola menirukan keterampilan sepak bola manusia seperti berdiri, berjalan, berlari, menggiring bola, mengoper, menendang dan bermain tim. Kualitas akurasi tendangan robot dapat di lihat dari perencanaan menendang pada saat menendang. Penggunaan metode kurva Bezier kubik dapat memudahkan untuk membangun perencanaan menendang dengan baik maka dirancang agar dapat melakukan pergerakan menendang dengan pergerakan lebih halus. Konstruksi perencanaan berupa kurva dengan dimensi X, Y, Z yang telah diinterpolasi, setelah itu inverse kinematika dilakukan sehingga diperoleh nilai perubahan sudut setiap joint kaki kanan robot. Perencanaan tendangan juga dioptimalkan untuk membuat bola menyentuh garis gawang dengan waktu tercepat. Dengan perkembangan ini menghasilkan perencanaan yang efektif menggunakan neural network optimization. Dari hasil tes tersebut, ditemukan bahwa waktu pergerakan 25 mili detik dan 10 mili detik mencatat waktu ketika bola menyentuh garis gawang 1 detik 836 mili detik, sedangkan waktu 50 mili detik mengangkat kakinya dan 15 mili detik dengan waktu 2 detik 208 mili detik untuk mencapai garis gawang. Setelah optimasi kedua variasi waktu dilakukan, akurasi 0,7% diperoleh dengan MSE 16,8657 dan MAE 3,5966 untuk nilai akurasi 25-10 sebesar 0,4% dengan MSE 16,9844 dan MAE 3,6157 untuk 50-15.
========================================================================================================================
A humanoid soccer robot is a robot that has ability to play soccer just like humans, such as walking, kicking, and standing, and demonstrates soccer skills such as dribbling, passing, and team play. The quality of the robot's kick accuracy can be seen from the planning of kicking at the time of kicking. The use of the cubic Bezier curve method can make it easier to build a good kicking plan, so it is designed to be able to carry out kicking movements with dynamic movements. The construction of the planning is in the form of a curve with dimensions X, Y, Z which has been interpolated, after which inverse kinematics are carried out so that the value of changes in the angle of each joint of the robot's right leg is obtained. Kicking planning is also optimized to get the ball to touch the goal line with the fastest time. With this development resulted in effective planning using neural network optimization. From the test results, it was found that the movement time of 25 milli seconds and 10 milli seconds recorded a time when the ball touched the goal line of 1 seconds 836 milli seconds, while a time of 50 milli seconds raised his legs and 15 milli seconds with a time of 2 seconds 208 milli seconds to reach the goal line. After the second optimization of time variations was carried out, an accuracy of 0.7% was obtained with MSE 16.8657 and MAE 3.5966 for an accuracy value of 25-10 of 0.4% with MSE 16.9844 and MAE 3.6157 for 50-15.

Item Type:	Thesis (Other)
Additional Information:	RSE 629.892 Riz p-1 2022
Uncontrolled Keywords:	Robot Humanoid, Kurva Bezier Kubik, Motion Planning, Neural Network, Inverse Kinematics, Bezier Cubic Curve
Subjects:	Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Anis Wulandari
Date Deposited:	01 Nov 2022 02:43
Last Modified:	01 Nov 2022 02:43
URI:	http://repository.its.ac.id/id/eprint/95038

Actions (login required)

View Item