Laser Based Altimetry for Unmanned Aerial Vehicle Hovering Over a Snow Surface

Abstract

A microwave radar for non-invasive snow stratigraphy measurements has been developed. Results were promising, but it failed to detect light powder snow in the air-snowpack interface. The aim of this thesis is to find and verify a system for estimating altitude on centimeter scale over a snow surface, independent of snow conditions. Also, relative pitch and roll angle estimation between the UAV and local surface should be resolved, to help directing the radar beam perpendicularly to the surface. After a variety of technical solutions were examined, we propose a system of three time-of-flight near-infrared altimeters pointing at different directions towards the surface. Experimental results showed RMS error of 1.39 cm for range estimation averaged over the most common snow types, and 2.81 cm for wet snow, which was the least reflective medium. An experiment conducted for an array of two altimeters scanning over a snow surface, showed that the local, relative surface tilt was found to be accurate within ±2° given that it was sufficiently planar. Further, the altitude RMS error was estimated to 1.57 cm. We conclude that the chosen altimeter was within the requirements, and that an array of three altimeters would give acceptable relative tilt estimation in to planes on the snow surface. The system should be subject to flight testing and implemented on UAV platform such that it can aid the microwave radar system during snow scanning.