Environment mapping using hybrid octree knowledge for UAV trajectory planning
Cedric Cocaud, Amor Jnifene, Bumsoo Kim · Canadian Journal of Remote Sensing · 2008
In addressing the need for enhanced autonomous unmanned aerial vehicle (UAV) controllers, this work presents a knowledge base providing a synthetic representation of the UAV’s environment, enabling autonomous trajectory planning. The knowledge base uses a memory-efficient hybrid structure combining octree and standard three-dimensional maps. Essential elements including terrain, obstacles, and atmospheric conditions are mapped in 3D through layers grafted onto the octree structure. The knowledge base also integrates the capability of adding user-defined layers to represent mission-dependent features such as radar exposure or communication link intensity.