By now everyone knows about Phillipp Robbel’s weekend hacking project where he used the Microsoft Kinect as a robotic sensor. Kinect is a fantastic 3D sensor that could have a ton of robotic applications – and the scale (price) at which it is produced, makes it very ideal.
An obvious place for Kinect is using it as a sensor for robust SLAM (Simultaneous Localization and Mapping).
SLAM is one of those interesting problems that predates computers. Scientists and mathematicians have been working on SLAM algorithms for a long time and though there are now a few different viable approaches, a low-cost robust SLAM system still eludes us.
Recently I had a brief conversation with William Cox of RobotBox.net who is offering an $800 prize to the first person who demonstrates live data from the Neato Robotics’s Neato vacuum cleaner, called the Open Lidar Project.
Update: the Open Lidar Project has a winner!
I’m therefore offering an additional $200 to the first person who can demonstrate any SLAM algorithm on a mobile robot using either the Microsoft Kinect or the Neato Vacuum Cleaner’s Lidar.
To win this prize, you’re free to choose any SLAM algorithm found on OpenSlam.org or any other algorithm of your choice. The robot must be able to map (with or without human assistance) but then be able to take command to navigate autonomously from Point A to Point B within the created map. The first person to post a video showing the result and sharing the code will win $200 from RoboDynamics.
Special thanks to William Cox of RobotBox.net, Phillipp Robbel of MIT, and of course Adafruit who put the original bounty on the Microsoft Kinect hack.