Skip to main content

Zhizun Xu

Artificial Intelligent control for autonomous underwater vehicles.

Email: z.xu21@ncl.ac.uk

Project supervisors

Project description

I am developing control and navigation methods based on artificial intelligent technologies.

Methodology and objectives

Due to high nonlinear dynamics, control is a challenge for underwater vehicles. A fuzzy controller would be able to guide the vehicle to approach the desired heading angle and depth. Underwater navigation is also crucial because an electronic wave can’t go through water. I will use visual odometry to estimate the motion trajectory. Previously, drones have used this technology.

Result

For the control, there are five different controllers. The AI controllers such as fuzzy control algorithms have a big overshoot compared with others. The modified sliding mode control has better performance over others. In navigation, I will create an innovative visual odometry. It will integrate camera, IMU and high accurate compass. Hence, such a method can work in a sparse feature environment. Using these methods, the error ratio is less than 5% after about 30m of travel.

Interests

Automatic control, computer vision, deep learning

Qualifications

  • MSc in marine engineering from Mokpo Maritime University
  • B.E in marine engineering from Shanghai Maritime University