Model-Based Autonomous Navigation with Moment of Inertia Estimation for Unmanned Aerial Vehicles.

Area of Science:

• Aerospace Engineering
• Robotics
• Navigation Systems

Background:

• Unmanned Aerial Vehicles (UAVs) commonly rely on Inertial Navigation Systems (INS) integrated with Global Navigation Satellite Systems (GNSS).
• Navigation accuracy rapidly degrades during GNSS signal outages, limiting UAV operational capabilities.
• Low-cost Micro-Electro-Mechanical Systems (MEMS) inertial sensors are susceptible to drift and noise, exacerbating navigation errors.

Purpose of the Study:

• To develop and evaluate a model-based navigation system for fixed-wing UAVs resilient to GNSS outages.
• To enhance navigation accuracy by integrating a Vehicle Dynamics Model (VDM) with MEMS INS and GNSS data.
• To calibrate system parameters, including moment of inertia, using an Unscented Kalman Filter (UKF).

Main Methods:

• Implementation of a model-based integration approach for fixed-wing UAV navigation.
• Utilizing a Vehicle Dynamics Model (VDM) as the primary process model.
• Employing low-cost MEMS inertial sensors and GNSS measurements.
• Calibrating moment of inertia terms via an Unscented Kalman Filter (UKF).

Main Results:

• Position error remained under 14.5 meters in all directions after 140 seconds of GNSS outage.
• Roll and pitch errors were bounded to within 0.06 degrees.
• Yaw error increased gradually to 0.65 degrees after 140 seconds of GNSS outage.
• The UKF successfully estimated model parameters, including coupled moment of inertia terms.

Conclusions:

• The proposed model-based integration approach significantly improves UAV navigation robustness during GNSS outages.
• Accurate estimation of dynamic parameters, like moment of inertia, is achievable even with sensor coupling.
• Further dynamic maneuvers could enhance the observability of estimated parameters for even greater accuracy.