Vertical Wind Velocity Estimation during UAS Fire Plume Encounters

Wildland fires present significant challenges for aircraft emergency response operations, including firefighting, surveillance, and cargo/crew transportation. The high temperature and strong turbulence near or within fire-generated plumes can be hazardous or catastrophic for manned and unmanned aircraft. In this paper, KHawk Zephyr3, a small fixed-wing unmanned aircraft system (UAS), was sent to fly through fire-generated plumes during a prescribed grass fire to collect the UAS response data in autopilot mode. During the fire plume encounters, the unmanned aircraft experienced significant changes in acceleration, angular rate, attitude, altitude, airspeed, and ground speed. Based on the collected UAS response data, two model-aided wind velocity estimation algorithms, a 2-state extended Kalman filter (EKF) and a 9-state EKF, were developed for the calculation of vertical wind velocity along the plume encounter flight trajectory. Both simulation and flight test results showed the effectiveness of the two vertic l wind estimation algorithms. Based on simulation analysis, the 2-state EKF (inertial angles of ttack and sideslip) performs slightly better than the 9-state EKF for wind velocity estimation. During the two selected UAS plume encounters, the updraft velocity within the fire generated plumes is estimated to be in the range of 6-10 m/s at a height of approximately 115 meters above the ground level.