WIND ESTIMATION FOR A QUADROTOR HELICOPTER IN A MODEL-AIDED NAVIGATION SYSTEM

摘要

Unmanned aerial vehicles (UAVs) typically rely on low-cost MEMS inertial measurement units (IMUs) for navigation purposes. However, GNSS outages provide a challenge as the navigation solution quickly deteriorates. A previously presented model-aided navigation system for a quadrotor helicopter shows low positioning errors during purely inertial flights. Since the vehicle dynamics model does not account for wind occurring in outdoor flights, estimating the wind velocity becomes necessary. In this paper, different models for estimating the wind velocity are presented and evaluated. Simulation results show that the wind velocity estimation is accurate, even with a low-grade IMU, GNSS, and no additional sensors such as an air speedometer. During GNSS outages, the new approach significantly outperforms the existing vehicle dynamics model and the pure INS solution.