DESIGN AND IMPLEMENTATION OF HYBRID UNDERWATER GLIDER (HUG) BERBASIS ROBOT OPERATING SYSTEM

• Main Author: Ihsan Hadi, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/48572
• Institution: Institut Teknologi Bandung

The underwater mapping process in Indonesia is currently still generally using surveyor ships, air vehicles and satellites. The use of ships and air vehicles requires a long time so it is costly for the crew and also fossil fuels, while the use of satellites has a low level of accuracy.So we need a tool that can retrieve data accurately but does not require human resources and fossil fuels in running it. The solution we are working on in this thesis to solve the problem is a vehicle in the form of Hybrid Underwater Glider (HUG) that uses a combination of Autonomous Underwater Glider (AUG) and Autonomous Underwater Vehicle (AUV) technology so that the vehicle is able to move with great energy efficiency but still able to maneuver according to the desired mapping path. In designing a HUG vehicle, one of the important systems so that the vehicle is able to move on its own without an operator is a navigation system. The navigation system functions to determine and update the position, attitude and speed data of the vehicle during the operation which will then be used by the control and guide system on the vehicle. The focus of this final project is the design and implementation of a navigation system from HUG. The navigation system is designed to use the Inertial Navigation System (INS) sensor system as the main sensor and the Doppler Velocity Log (DVL) sensor as an auxiliary sensor. Data readings from the two sensors are processed using the Kalman Filter integration algorithm so that the error growth caused by bias in the estimation of the navigation data of the vehicle can be eliminated and the noise that arises can be muted. The implementation of the vehicle navigation system is implemented with the Robot Operating System (ROS) framework on UDOO's single board computer. The test was conducted using the Hardware in Loop System (HILS) method using vehicle dynamics data in matlab. The test results showed that the navigation system that was designed was able to reduce noise and bias in the speed and angle data of the vehicle for gliding for 1 hour. However, for the estimation of position data, differences in performance are found according to the initial initialization of the system covariance matrix, which concluded that there is an optimal initial covariance matrix system value to guarantee the estimated performance of the system.