Perancangan Sistem Kontrol Altitude Pid-Neural Network Pada Drone Quadcopter.

Abiesa, Addo Adantya (2022) Perancangan Sistem Kontrol Altitude Pid-Neural Network Pada Drone Quadcopter. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 02311740000125-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (5MB)

Abstract

Seiring berjalannya waktu, manusia dapat menciptakan pesawat terbang dan helikopter. Lalu munculah drone karena kecanggihan teknologi kecerdasan buatan dan teknologi sistem kendali pada pesawat terbang. Oleh sebab kedua macam kecanggihan teknologi itu pula, belakangan ini banyak sekali perguruan tinggi maupun lembaga penelitian di dunia yang telah melakukan penelitian menggunakan quadcopter. Salah satunya penelitian kendali PID yang dijadikan sistem tertanam di dalam quadcopter. Namun, kendali PID sendiri memiliki kekurangan yang disebabkan konstanta yang bersifat tetap, sedangkan sistem quadcopter yang non linear membutuhkan kendali adaptif saat terkena gangguan tertentu. Jaringan syaraf tiruan menjadi salah satu dari beberapa metode yang digunakan untuk meningkatkan performansi sistem kontrol. PID-NN yang diaplikasikan berarsitektur multilayer-backpropagation. Pengujian performansi dilakukan dengan set point tetap dan set point tracking. Alhasil, hasil pengujian kedua kontrol dengan set point tracking: rerata maximum overshoot dan undershoot pada PID-NN sebesar 23.86% dan 1.607%, sedangkan pada PID saja sebesar 40.667% dan 0.211%. Settling time pada PID-NN sebesar 0.211 detik, sedangkan pada PID saja tidak diperoleh nilainya karena sinyal respons keluaran yang terus berosilasi hingga tak mencapai keadaan tunak. Hasil tersebut mengindikasikan bahwa PID-NN dapat membuat sistem kontrol altitude lebih baik.
===================================================================================================================================
Over time, humans can create airplanes and helicopters. Then came the drones because of the sophistication of artificial intelligence technology and control system technology on airplanes. Because of the two kinds of technological sophistication, lately there are not once universities and research institutions in the world that have conducted research using quadcopters. One of them is PID control research which is used as an embedded system in the quadcopter. However, PID control itself has shortcomings caused by constants that are fixed, while quadcopter systems that are non-linear require adaptive control when exposed to certain disturbances. Artificial neural networks are one of several methods used to improve the performance of the control system. The PID-NN applied is multilayerbackpropagation architecture. Performance testing is carried out with fixed set points and set point tracking. As a result, the test results of the two controls with set point tracking: the average maximum overshoot and undershoot in PID-NN was 23.86% and 1,607%, while in PID alone it was 40.667% and 0.211%. The settling time on the PID-NN is 0.211 seconds, while in the PID alone the value is not obtained because the output response signal continues to oscillate until it does not reach the tunak state. These results indicate that the PID-NN can make the altitude control system better.

Item Type:	Thesis (Other)
Additional Information:	RSF 629.89 Abi p-1 2022
Uncontrolled Keywords:	PID, Neural network, Quadcopter, Altitude, Performansi. PID, Neural network, Quadcopter, Altitude, Performance.
Subjects:	Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	11 May 2026 01:50
Last Modified:	11 May 2026 01:50
URI:	http://repository.its.ac.id/id/eprint/133098

Actions (login required)

View Item