EKF-Based Parameter Identification of Multi-Rotor Unmanned Aerial VehiclesModels

Area of Science:

• Robotics and Control Systems
• Aerospace Engineering
• System Identification

Background:

• Accurate model parameter estimation is crucial for multi-rotor unmanned aerial vehicle (UAV) control and performance.
• Traditional test-bed based identification methods can be time-consuming and may not reflect real-world flight conditions.
• Onboard sensors offer a potential source for real-time parameter estimation during operation.

Purpose of the Study:

• To develop and validate a novel online method for estimating all model parameters of multi-rotor UAVs.
• To investigate the system's observability for parameter identification using onboard sensor measurements.
• To demonstrate the practical applicability of the proposed method through simulations and experimental flight data.

Main Methods:

• Development of dynamic models for three classes of multi-rotor UAVs.
• Nonlinear observability analysis by augmenting the state vector with parameters to be identified.
• Implementation of an extended Kalman filter (EKF) for online parameter estimation using onboard sensor data.
• Validation through extensive computer simulations and real flight log data from a custom quadrotor.

Main Results:

• The nonlinear observability analysis identified specific measurement sets and conditions for successful parameter estimation.
• Simulation results confirmed the feasibility of estimating all model parameters in a single online process using the EKF.
• Experimental validation using flight data demonstrated the practical suitability of the proposed estimation method.

Conclusions:

• The proposed EKF-based online method effectively estimates all model parameters for multi-rotor UAVs using readily available onboard sensor data.
• The method overcomes limitations of traditional test-bed identification, offering a practical solution for real-time parameter estimation.
• This approach enhances the potential for adaptive control and improved performance of UAVs in diverse operational scenarios.