Wind Estimation by Unmanned Air Vehicle with Delta Wing

Abstract

In this paper, an algorithm to estimate wind direction by using a small and light Unmanned Air Vehicle(UAV) called KITEPLANE was proposed. KITEPLANE had a big main wing which is a kite-like delta shape and, therefore, it was easy to be disturbed by wind. However, this disadvantage implies that the KITEPLANE has an ability to sense wind and that it is expected to use the KITEPLANE as a sensor for wind estimation. In order to achieve this feature, dynamics of the KITEPLANE under wind disturbance were derived and a numerical estimation method was proposed. Devices equipped on board were also developed and the proposed method was implemented. Results of an experiment showed the effectiveness of the proposed method.

Self-motion and wind velocity estimation for small-scale UAVs

“…Stacking computed epipoles and the corresponding pose error states as vectorsȳ = vec([y (1) · · · y (P ) ]) and δρ = vec([δρ (1) · · · δρ (P ) ]), respectively, yields the compact measurement equation, in terms of the error states,…”

“…When operating in GPS-denied or impaired environments, such as indoor or urban canyons, the UAV has to rely on dead-reckoning systems and/or natural landmarks. Furthermore, changing external conditions may include wind disturbances which affect the performance of autonomous flight control and efficiency of path planning and following [1], [2].…”

Lateral wind force and torque estimation for a driving assistance

“…The problem of wind estimation is often tackled from the point of view of air vehicle [Eld06,Kum05]. But, in the case of air vehicle, the wind has the most impact on the general dynamic.…”

Biologically inspired self-motion estimation using the fusion of airspeed and optical flow

“…This strategy is known as anemotaxis [7]. Engineers use wind speed estimation for pollution tracking, fire prevention, and weather prediction using UAVs [8].…”

“…This strategy has been hypothesized for moths tracking pheromone plumes [9]. Another strategy is to use a knowledge of self-dynamics, control inputs (thrust and control surface deflections), and inertial measurements [8].…”