Jannaty, Baity (2020) Desain Kendali Haluan Kapal dengan Unscented Model Predictive Control (UMPC). Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06111850010008-Master_Thesis.pdf Download (2MB) | Preview |
Abstract
Pada Tesis ini dibahas kendali haluan kapal dengan metode Unscented Model Predictive Control (UMPC). Sudut kemudi dari kapal dikendalikan agar sudut haluan kapal sesuai dengan yang diharapkan. Model yang digunakan mempertimbangkan tiga derajat kebebasan, yaitu sway, yaw, dan roll. UMPC merupakan pengembangan dari Model Predictive Control (MPC) untuk sistem non linier dengan gangguan yang bersifat stokastik. Terdapat noise yang melekat pada model sistem, sehingga sistem berbentuk dinamik stokastik dengan kendala probabilistik. Noise pada model proses menambah ketidakpercayaan terhadap state secara langsung, sedangkan noise pada model pengukuran menyebabkan estimasi state menjadi tidak pasti. Keberadaan noise tersebut menyebabkan perubahan pada variabel keadaan dari memiliki nilai yang pasti menjadi variabel acak berdistribusi. Fungsi tujuan berubah dari deterministik menjadi bentuk ekspektasi peubah acak. Agar metode MPC dapat diterapkan, kendala variabel keadaan diubah dari probabilistik menjadi deterministik. Fungsi tujuan diubah ke dalam bentuk deterministik. Sistem dilinierkan dengan pelinearan statistik untuk mengaproksimasi transisi dari state. Untuk menggantikan prediksi yang dilakukan MPC, digunakan Unscented Kalman Filter (UKF). Hasil prediksi yang diperoleh kemudian digunakan pada algoritma MPC dengan meminimumkan fungsi tujuan. Hasil simulasi menunjukkan bahwa UMPC dapat mengatasi permasalahan dengan gangguan yang bersifat stokastik.
================================================================================================================================
This Thesis discussed a ship heading control problem with Unscented Model Predictive Control (UMPC). Rudder angle is controlled such that the heading angle follows the desired angle. This paper used model with three degree of freedom, that is sway, yaw, and roll. UMPC is based on Model Predictive Control (MPC) for non linear system with stochastic disturbance. There are noises in the system model, therefore the system become dynamic stochastic with probabilistic constraint. Process noise which adds uncertainty to the state directly and measurement noise which makes the state estimate uncertain. This noises cause a changing in the state variable from a definite value to a distributed random variable. The objective function change from deterministic into the form of expectations of random variable. To handle this problem, state variable constraints are changed from probabilistic to deterministic. The objective function are changed into deterministic forms. System is linearized using stochastic linearization to approximate state transition. The Unscented Kalman Filter (UKF) is used as prediction process for MPC. The prediction process result is used by MPC algorithm by minimizing the objective function. The simulation results showed that UMPC can handle problem with stochastic disturbance.
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTMa 515.642 Jan d-1 2020 |
| Uncontrolled Keywords: | Model Predictive Control (MPC), Unscented Kalman Filter (UKF), kendali haluan kapal, gangguan stokastik |
| Subjects: | Q Science > QA Mathematics > QA401 Mathematical models. Q Science > QA Mathematics > QA402 System analysis. V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM161 Ships--Hydrodynamics V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM293 Shipping--Indonesia--Safety measures |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44101-(S2) Master Thesis |
| Depositing User: | Baity Jannaty |
| Date Deposited: | 05 Jun 2023 03:32 |
| Last Modified: | 05 Jun 2023 03:32 |
| URI: | http://repository.its.ac.id/id/eprint/73181 |
Actions (login required)
 |
View Item |