Undergrad Research Project - Embedded SLAM for quadcopters

Fall 2016

Student
Eric Fang
Advisor
Nathan Michael
Project description

The project goal is to achieve SLAM utilizing only the resources of a quadcopter flight controller and independent of any external (from quadcopter) source of mapping or localization. This system is intended to be used as a backup system for more sophisticated but less robust means of SLAM (such as CV based SLAM) for a MAV to roughly localize itself in the event of higher system errors.

The chosen approach is to mount one teraranger tower rangefinder array as well as a downward facing SF02 Lidar sensor on the quadcopter to feed the resulting readings directly into the onboard PX4 firmware. An occupancy grid will then be constructed and populated utilizing the Lidar and laser scan readings for the quad to localized itself within a 2.5D context.

It is expected for this system to produce a robust, high frequency position estimator within the flight controller itself with acceptable localization error that can supplement other SLAM solutions that may be calculated on companion computers.

Return to project list