New Vectorial Propulsion System and Trajectory Control Designs for Improved AUV Mission Autonomy

Ivan Masmitja, Julian Gonzalez, Cesar Galarza, Spartacus Gomaríz, Jacopo Aguzzi, Joaquín del RíoSensors 2018 · DOI

TL;DR

This paper introduces a novel vectorial propulsion system and advanced fuzzy control strategies for the Guanay II Autonomous Underwater Vehicle (AUV). These innovations significantly enhance the AUV's maneuverability and autonomous navigation capabilities, enabling more complex missions in challenging underwater environments.

What Problem Does This Paper Address?

The deep ocean remains largely unexplored due to the limitations of human accessibility and the poor visibility underwater. While AUVs offer a promising solution for marine research, achieving true autonomy in navigation, especially in complex environments with varying depths and currents, remains a significant challenge. Existing AUVs often rely on teleoperation or have limited maneuverability, hindering their ability to perform sophisticated tasks like detailed inspections, mapping, or precise sampling. This paper addresses the need for AUVs with enhanced maneuverability and robust autonomous navigation systems capable of handling dynamic underwater conditions.

Methods

The authors developed a new vectorial propulsion system for the Guanay II AUV, allowing for full maneuverability in both horizontal and vertical planes. This system involves a novel mechanism to adjust the angle of lateral thrusters, enabling precise control over the vehicle's pitch and heave.

For trajectory control, they implemented a two-loop fuzzy control architecture:

  • Inner Loop: This loop controls the vehicle's forward velocity and yaw. It utilizes Type-1 TSK (Takagi-Sugeno-Kang) fuzzy controllers that dynamically adjust parameters of linear controllers (PD for yaw, PI for velocity) based on the vehicle's forward speed. This approach allows for adaptive control across different velocities, overcoming limitations of fixed linear controllers.
  • Outer Loop: This loop handles the kinematic aspects of navigation, employing a pure pursuit strategy with a fuzzy controller. This controller takes into account yaw error, distance to the waypoint, and the angle to the next waypoint to adjust the vehicle's velocity reference, ensuring smoother and more accurate path following.

The control strategies were developed using a simplified 3-DOF mathematical model of the AUV and validated through simulations and field tests.

Datasets

Not discussed in the paper. The paper focuses on the development and testing of the AUV's control system, using simulations and field tests with the Guanay II AUV itself.

Key Insights and Ideas

  • Vectorial Propulsion for Enhanced Maneuverability: The ability to vector the thrust of lateral thrusters is crucial for achieving full 3D maneuverability in AUVs, particularly for depth control and precise attitude adjustments.
  • Fuzzy Control for Adaptive Navigation: Fuzzy logic controllers excel at interpolating between different control strategies designed for specific operating conditions (e.g., different forward velocities). This allows for smoother and more robust navigation compared to traditional gain-scheduling or fixed linear controllers.
  • Decoupled Horizontal and Vertical Control: The paper demonstrates that for many missions, horizontal and vertical navigation can be effectively decoupled, simplifying the control design while maintaining performance.
  • Pure Pursuit with Fuzzy Logic for Path Following: Combining the established pure pursuit algorithm with fuzzy logic for velocity reference generation leads to improved path tracking accuracy, especially when approaching waypoints.
  • Addressing Real-World Constraints: The study highlights the importance of considering physical constraints like thruster torque limits and sensor noise (e.g., from waves) when designing AUV control systems, as these can significantly impact performance and lead to undesirable behaviors.

Key Results

  • Vertical Navigation: The new vectorial propulsion system was shown to provide full maneuverability in the vertical plane. Field tests validated the simulation results, demonstrating effective depth control and trajectory following.
  • Horizontal Navigation (Inner Loop): The fuzzy controller for the inner loop (yaw and velocity) showed improved performance over proportional controllers, particularly in reducing torque switching and achieving faster responses. For instance, the fuzzy controller achieved a settling time of 12.2s for yaw at 0.3 m/s, compared to 29.9s for one of the proportional controllers.
  • Horizontal Navigation (Outer Loop): Field tests comparing the fuzzy outer loop controller with proportional controllers showed that the fuzzy controller achieved faster yaw reference tracking and reduced the vehicle's velocity when approaching waypoints, leading to more accurate path following.
  • Constraint Handling: The fuzzy controllers demonstrated better robustness to torque saturation and noise compared to some linear controllers, although the paper notes that certain noise levels can still cause issues.

Limitations and Future Work

  • Buoyancy Management: The paper emphasizes that careful buoyancy adjustment is mandatory before each mission, as it significantly influences the vehicle's ability to submerge and surface.
  • Model Accuracy: Small deviations between simulation and field test results suggest potential inaccuracies in the AUV's mathematical model coefficients or the final configuration (e.g., buoyancy position, thruster inclination).
  • Future Work: The authors suggest future research directions including:
  • Developing a controller for the vertical plane and integrating it with horizontal plane control for full 3D trajectory control.
  • Implementing path following techniques that allow the vehicle to follow a specific path rather than just waypoints.
  • Exploring other controller techniques.

Summary updated

BibTeX
@article{masmitja2018new,
  author = {Ivan Masmitja and Julian Gonzalez and Cesar Galarza and Spartacus Gomaríz and Jacopo Aguzzi and Joaquín del Río},
  title = {{New Vectorial Propulsion System and Trajectory Control Designs for Improved AUV Mission Autonomy}},
  journal = {Sensors},
  year = {2018},
  doi = {10.3390/s18041241},
}
Figure