Abstract
The underwater realm, characterized by its complexity and vastness, necessitates the use of autonomous underwater vehicles (AUVs) for exploration and operational tasks. Traditional propeller-driven AUVs have been instrumental in this field, but their limitations in agility are becoming increasingly evident. This thesis rigorously investigates the capabilities of fin-actuated AUVs, which draw inspiration from the locomotion principles of marine life.A significant challenge in underwater operations is the robust and accurate tracking of divers, particularly in conditions of limited visibility. Existing methodologies often fall short in such demanding scenarios. To address this, the thesis introduces a diver tracking system that is both sophisticated and cost-effective. By synergistically integrating visual and acoustic data, enhanced tracking accuracy is achieved, demonstrating the adaptability of fin-actuated AUVs in dynamic marine environments.Transitioning from diver tracking, the thesis addresses a gap in current literature concerning the locomotion dynamics of fin-actuated AUVs. Specifically, models that relate fin actuation parameters to generated thrust, and that can be easily inverted for control purposes, are lacking. A nonlinear dynamic model is presented, capturing the relationship between fin thrust and its oscillation parameters. Empirically validated, this model serves as the foundation for the development of an inverse model, offering a novel approach to vehicle control.Control precision is a critical aspect of AUV operations. Traditional AUVs face challenges in achieving 6-DOF (Degrees of Freedom) control, particularly when constrained by a limited number of actuators. A novel 6-DOF control method tailored for fin-actuated AUVs is presented. This method enables intricate underwater motions, achieving 6-DOF control with only four actuators, a significant advancement in the field.The necessity for fault-tolerant control in underwater robotics extends beyond academic interest; it is a critical requirement for real-world missions. Given the harsh and unpredictable underwater conditions, actuator failures are not uncommon. For fin-actuated AUVs, a single fin failure can severely impede the vehicle's operational capabilities. The thesis introduces an active fault-tolerant control scheme specifically designed for four fin-actuated AUVs. The scheme features a control allocation switching mechanism that adaptively redistributes control forces among the fins in the event of a failure.In conclusion, the thesis offers a comprehensive study on fin-actuated AUVs, covering aspects of perception, modeling, and control. Each contribution not only addresses a specific challenge but also lays the groundwork for future advancements in underwater robotics. The experimentally validated results further underscore the potential utility of fin-actuated AUVs in complex underwater exploration and operations.