Guidance, navigation and control of an autonomous underwater vehicle

Abstract

Autonomous underwater vehicles (AUVs) change the way the ocean is explored and used by letting people move through and explore the deep ocean without any help. These robots can be used in many areas, such as marine science, tracking the environment, underwater archeology, offshore business, the military, and spying. Today's military operations are greatly improved by the unbeatable benefits of autonomous robots, which can work alone, gather information in dangerous situations, and carry out difficult tasks. For autonomous underwater vehicles to work, reliable and powerful guidance, navigation, and control (GNC) systems must be made that are especially made to deal with the challenges of the underwater world. Autonomous underwater vehicles need guidance, navigation, and control systems to stay on the trajectories they're supposed to follow, find their exact locations, and move around easily. This in turn ensures the successful completion of missions in different combat situations. This thesis adds a lot to the area of technology for autonomous underwater vehicles. It works on creating, applying, and testing better methods for Guidance, Navigation, and Control. The goal of these algorithms is to make the AUV more autonomous, better at turning, and better at its general performance. Within the domain of control, a new control architecture is implemented to control dynamics of the underwater vehicle. The controller utilizes an advanced Linear Parameter Varying Model Predictive Control (LPV-MPC) method to manage the vehicle's dynamics, taking into consideration the nonlinearities and uncertainties that are inherent in underwater operations. The controller utilizes model predictive control approaches to provide optimal tracking performance under different operating situations and disturbances, ultimately improving the vehicle's agility and stability. Navigation is an important part of AUV activities because it lets them find their exact location, plan their routes, and carry out their tasks. To get around the problems of underwater guidance, a state estimation method based on the cutting edge Kalman Filter is used. The Kalman Filter lets user to get a real-time estimate of the vehicle's position, speed, and direction, among other things. It does this by putting together data from different types of sensors, such as pressure sensor systems, Doppler velocity logs, and inertial sensors. The navigation algorithm uses optimum estimation and sensor fusion to give accurate and dependable estimates of the state. This lets the AUV move around on its own and adapt to changing conditions in its surroundings. In the thesis, a relatively new algorithm callede Virtual Target Approach is introduced to the field of guidance. This approach makes it possible to accurately hit objects at certain angles. The Virtual Target Approach, which is based on the ideas of optimal control and trajectory optimization, lets the AUV change its path in real time in relation to a virtual target. This improvement makes path planning and engaging targets better. It is possible to do mission-critical tasks like target tracking, spying, and inspection more quickly and accurately underwater by using this method, which considers impact angle and target movement.