Pengembangan Autopilot AUV Berbasis Unscented Kalman Filter Dan Artificial Potential Field Untuk Navigasi Adaptif Dan Penghindaran Halangan Multi-Arah

Ardian, Rafi Maulana (2026) Pengembangan Autopilot AUV Berbasis Unscented Kalman Filter Dan Artificial Potential Field Untuk Navigasi Adaptif Dan Penghindaran Halangan Multi-Arah. Other thesis, Institiut Teknologi Sepuluh Nopember.

Text 5022221234-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until August 2026. Download (14MB) | Request a copy

Abstract

Autonomous Underwater Vehicle (AUV) membutuhkan sistem autopilot yang mampu mengestimasi state dan melakukan penghindaran halangan pada lingkungan bawah laut dengan keterbatasan sensor. Penelitian ini mengembangkan autopilot AUV berbasis integrasi Unscented Kalman Filter (UKF) dan Artificial Potential Field (APF) pada model AUV 5-DOF menggunakan MATLAB/Simulink. UKF digunakan untuk mengestimasi posisi, kecepatan, dan orientasi, sedangkan Forward Looking Sonar (FLS) digunakan sebagai masukan APF untuk menghasilkan koreksi heading, pitch, dan kecepatan. Improved APF dengan modifikasi gaya repulsif dan Regular Hexagon-Guided (RHG) diterapkan untuk menangani local minima dan GNRON. Pengujian dilakukan pada lintasan tanpa halangan, lintasan 1-waypoint, dan lintasan 2-waypoint dengan variasi K_rep=100, 200, dan 350 pada kondisi sensor normal serta sensor gangguan. Hasil pengujian menunjukkan bahwa UKF menghasilkan RMSE posisi horizontal kurang dari 1 m pada kondisi sensor normal dan RMSE kedalaman sekitar 0,010 m. Pada kondisi sensor gangguan, RMSE posisi x meningkat hingga 3,707 m. FLS dan APF mampu menghasilkan respons penghindaran halangan multi-arah, tetapi pada lintasan 2-waypoint kondisi sensor gangguan khususnya variasi 3 terjadi collision. Secara umum, integrasi UKF dan APF mampu mendukung navigasi adaptif dan penghindaran halangan multi-arah, meskipun performanya dipengaruhi oleh gangguan sensor seperti GPS tidak tersedia dibawah laut, DVL dropout, IMU bias serta konfigurasi halangan, efek hidrodinamika, dan keterbatasan aktuator.
=================================================================================================================================
An Autonomous Underwater Vehicle (AUV) requires an autopilot system capable of estimating its state and performing obstacle avoidance in the underwater environment despite sensor limitations. This research develops an AUV autopilot based on the integration of the Unscented Kalman Filter (UKF) and the Artificial Potential Field (APF) on a 5-DOF AUV model using MATLAB/Simulink. The UKF is used to estimate position, velocity, and orientation, while the Forward-Looking Sonar (FLS) is used as input to the APF to generate corrections for heading, pitch, and velocity. An improved APF with a modified repulsive force and Regular Hexagon Guided is applied to address local minima and GNRON. Testing was conducted on obstacle-free trajectories, 1-waypoint trajectories, and 2-waypoint trajectories with variations of 200 and 350 under both normal sensor conditions and sensor interference conditions. Test results show that the UKF produces a horizontal position RMSE of less than 1 m under normal sensor conditions and a depth RMSE of approximately 0.010 m. Under sensor disturbance conditions, the position RMSE increases to 3.707 m. The FLS and APF are capable of generating multi-directional obstacle avoidance responses, but on the two-waypoint route under sensor disturbance conditions particularly with a variation of 3 a collision occurred. In general, the integration of UKF and APF supports adaptive navigation and multi-directional obstacle avoidance, although performance is affected by sensor disturbances such as GPS underwater unavailability, DVL dropouts, and IMU bias as well as obstacle configuration, hydrodynamic effects, and actuator limitations.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Autonomous Underwater Vehicle, Unscented Kalman Filter, Artificial Potential Field, Forward Looking Sonar, Penghindaran Halangan. ======================================================================================================================== Autonomous Underwater Vehicles, Unscented Kalman Filter, Artificial Potential Field, Forward-Looking Sonar, Obstacle Avoidance.
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1007 Electric power systems control
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Rafi Maulana Ardian
Date Deposited:	27 Jul 2026 01:37
Last Modified:	27 Jul 2026 01:37
URI:	http://repository.its.ac.id/id/eprint/137791

Actions (login required)

View Item