Prakoso, Muhamad Rizal Ridlo (2019) Perancangan Algoritma Integrasi Sistem Navigasi Inersia dengan GPS Berbasis Unscented Kalman Filter. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
07111540000016-Undergraduate_Theses.pdf Download (4MB) | Preview |
Abstract
— Pada tugas akhir ini, dikembangkan algoritma integrasi antara sistem navigasi inersia (INS) dengan GPS menggunakan algoritma Unscented Kalman Filter. Inertial Navigation Sysem atau sistem navigasi inersia adalah suatu sistem yang terdiri dari sensor, dan algortima mekanisasi yang dapat digunakan untuk mengetahui 9 state, yaitu orientasi (roll, pitch, yaw), kecepatan, dan posisi wahana berdasarkan inisialisasi dan penyelarasan, namun sensor pada INS ini sangat rentan terhadap gangguan dari luar maupun dari dalam sensor tersebut, sehingga ketidak akuratan terhadap solusi navigasi sangat besar. GPS sangat umum digunakan untuk mengoreksi kesembilan state tersebut, sehingga dibutuhkan state estimator berupa Kalman Filter nonlinier dikarenakan model kesalahan 15 state nonlinier, 9 state pada mekanisasi ditambah dengan 6 state bias akselerometer dan giro. Peran estimasi bias akselerometer dan giroskop sangat penting karena nilainya berubah setiap waktu. Integrasi INS-GPS berhasil meningkatkan performansi sebesar 40,30% , 6,73%, dan 48,85% untuk state roll, pitch, yaw. 82,44%, 94,66%, dan 90,85% untuk state kecepatan north, east, down, dan 91,52%, 97,11%, 98,18% untuk state latitude, longitude, dan ketinggian dibandingkan dengan tanpa integrasi GPS. Hasil menunjukkan perbedaan yang tidak signifikan antara performansi UKF dan KF. Peningkatan performansi yang terlihat adalah pada state yaw yaitu sebesar 0,1055%, kecepatan sumbu down sebesar 0,1403%, posisi lintang sebsar 0,006%, dan ketinggian sebesar 0,1%, dan waktu komputasi yang diperlukan UKF lebih lama jika dibandingkan dengan KF.======================================================In this final project, the integration algorithm between INS (Inertial Navigation System) and GPS is developed using Unscented Kalman Filter. INS or Inertial Navigation System is a system which consists of sensors, and a mechanization algorithm used to know 9 states, which are attitude (roll, pitch, yaw), velocity, and position that is based on initialization and alignment processes, but the sensor of INS is prone to disturbance and noise both from external sources and internal sources, so that the inaccuracies of navigation solution is large. GPS is commonly used to correct nine states using the information of velocity, and position, so that the Kalman Filter is used as a non-linear state estimator because of the nonlinearity of the 15 states of error model, which are 9 states from mechanization, and 6 states from accelerometer and gyroscope biases. Bias role estimation affects the INS performance since the value changing with time. The INS-GPS integration sucessfully enhace the performance about 40,30% ,6,73%, and 48,85% for roll, pitch, yaw. 82,44%, 94,66%, dan 90,85% for north, east, down velocities, and 91,52%, 97,11%, 98,18% for latitude, longitude, and height compared to no GPS integration.But the results show that there are no significant difference of performance improvement between using UKF compared to KF. The improvements using UKF are the yaw state which is 0.1055%, velocity at down axis which is 0,1403%, latitutde which is 0,006%, and the altitude which is 0,1%, and with more computational time consumption.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSE 629.89 Pra p-1 2019 |
| Uncontrolled Keywords: | Unscented Kalman Filter, Nonlinier, Navigasi, INS GPS, State estimator |
| Subjects: | H Social Sciences > HA Statistics > HA31.7 Estimation Q Science > QA Mathematics > QA402.3 Kalman filtering. U Military Science > UG1242 Drone aircraft--Control systems. (unmanned vehicle) |
| Divisions: | Faculty of Electrical Technology > Electrical Engineering > 20201-(S1) Undergraduate Thesis |
| Depositing User: | Muhamad Rizal Ridlo Prakoso |
| Date Deposited: | 16 Mar 2023 03:36 |
| Last Modified: | 16 Mar 2023 03:36 |
| URI: | http://repository.its.ac.id/id/eprint/63710 |
Actions (login required)
 |
View Item |