Within the Artemis project R3-COP, where I take part as ReLab consultant for Elettric80, I have recently worked on Visual Odometry, i.e. estimating the position of a vehicle exploiting only the data coming from an on-board camera.
Solving odometry through the use of a camera is much cheaper and easier with respect to the use of traditional encoders on wheels and steer, as such solution requires invasive installations over the vehicle, and even more, its accurate kinematic model.
You can have a look to an installation on a forklift and some video results:
Results on such topic are still quite fresh and most of them still belong to research; however I know of some interesting technology trasfers from academia to industry. Contact me if interested.
Acknowledments for this work:
- Elettric80 for providing forklifts and warehouse
- Artemis project R3-COP
- Visual Odometry inspired to the paper: Henning Lategahn, Andreas Geiger and Bernd Kitt, Visual SLAM for Autonomous Ground Vehicles, International Conference on Robotics and Automation (ICRA), 2011.