Planning for Autonomous Aerial Vehicles


The goal of this project is to enable unmanned aerial vehicle to operate fully autonomously in near-earth spaces such as urban environments and sparse forests. Robust planning is challenging because it needs to occur in a relatively high-dimensional space in order to generate dynamically feasible trajectories, it needs to be fast in order to allow for fast flying, and it needs to account for massive uncertainty due to noisy perception and inaccurate localization.

This project is a collaboration with Dragonfly Pictures, Inc company that designs and builds unmanned helicopters.

Quadrotor-based system we built for indoor testing. It is equipped with on-board processing and sensors. Movie of autonomous landing site selection using PPCP-based planner [IROS'11] Fully autonomous flight by the quadrotor. Everything (motion planner, landing planner, SLAM, 3D map building, and terrain analysis) is run on-board the quadrotor. Here is a movie.
Hexarotor MAV with on-board sensing and processing that my students built for testing autonomous flight and landing. Movie of autonomous flight amongst people using ASIPP planner [IROS'13] Ground-air robotic system built in our lab. Movie showing this system in action can be found here