Loading...

Analysis and Increase Robustness of INS-DVL Navigation Using Rotational Model of AUV in the Presence of Model Uncertainty

Ramezanifard, Amir Hossein | 2021

459 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54338 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Alasti, Aria; Salarieh, Hasaan; Hashemi, Mojtaba
  7. Abstract:
  8. Autonomous underwater vehicle are among the subsurface vehicles used today in applications such as ocean floor mapping, environmental studies, finding the wreckage of drowned vehicles, and military applications such as finding mines. The need for control over the navigation components and the importance of the precise position of the mission have made AUV navigation an important issue in these vehicles. Due to the attenuation of the GPS signal underwater, it is not possible to use this method for subsurface navigation. Therefore, inertial navigation (dead reckoning navigation) is usually used for this purpose. Due to the unlimited growth of inertial sensor error over time, this error must be limited by auxiliary measurements.One of the most common integrated underwater navigation methods is INS-DVL navigation. Unfortunately, INS-DVL navigation has not yet reached sufficient accuracy in the country. One way to increase the INS-DVL navigation accuracy is to use the vehicle dynamic model. In this research, the rotational dynamics model of the vehicle has been used to filter the output of the gyroscope. However, by identifying the rotational model for different tests, sometimes up to 100% parametric uncertainty is seen in this model. This uncertainty is often due to simplifications made to the model. Due to this uncertainty, the use of conventional Kalman filters does not provide sufficient final accuracy in navigation. Studies have shown that this uncertainty can reduce the final accuracy of navigation by up to 40 times if conventional Kalman filters are used. Therefore, in this research, a robust Kalman filter based on the AUV rotational dynamics model has been used to filter the gyro output. Applying this filter to real navigation data shows that using this filter can improve the ultimate navigation accuracy compared to non-robust filters by up to 60%. This review is based on data from several separate tests to ensure reproducibility
  9. Keywords:
  10. Navigation ; Underwater Navigation ; Autonomous Underwater Vehicle ; Rotational Dynamic ; Robust Kalman Filter ; Doppler Velocity Log (DVL) ; Inertial Navigation System ; Inertial Navigation System-Doppler Velocity Log (INS-DVL)

 Digital Object List

 Bookmark

No TOC