PPAC-Pilot: Prescribed-performance augmented control for fixed-wing autopilots.

Area of Science:

• Aerospace Engineering
• Control Systems
• Robotics

Background:

• Traditional autopilots for fixed-wing Unmanned Aerial Vehicles (UAVs) often require extensive tuning for optimal performance.
• Existing control systems struggle to guarantee specific performance bounds, particularly for altitude tracking.
• Historical flight data is an underutilized resource for control system development.

Purpose of the Study:

• To introduce a novel Prescribed-Performance Augmented Control (PPAC) framework for UAV autopilots.
• To enhance existing Proportional-Integral-Derivative (PID) control loops without complete redesign.
• To leverage historical flight data for deriving dynamic models and control laws.

Main Methods:

• Developed a PPAC framework to augment existing open-source autopilot PID controllers.
• Utilized historical flight data to derive dynamic linearization models and control laws, avoiding explicit UAV modeling.
• Integrated the PPAC framework with the Total Energy Control System (TECS) for practical application.
• Conducted numerical simulations and Hardware-in-the-Loop (HIL) tests for validation.

Main Results:

• PPAC successfully augmented baseline autopilot performance in takeoff and cruising scenarios.
• The framework ensured prescribed performance bounds for altitude tracking errors.
• Comparative analysis demonstrated improved performance of PPAC-augmented systems over baseline autopilots.
• Validation through simulations and HIL tests confirmed the effectiveness of the PPAC strategy.

Conclusions:

• The Prescribed-Performance Augmented Control (PPAC) framework effectively enhances fixed-wing UAV autopilots.
• PPAC guarantees altitude tracking performance while minimizing redesign efforts for existing control systems.
• Leveraging historical flight data provides a robust method for control law derivation without explicit UAV models.