Robot Navigation using a Biomimetic Vision System Test 4
Using a biomimetic vision system for mobile robot navigation. Vision system developed in The University of Adelaide’s Physiology Dept. This is Phase 2 of the Autonomous Control System for a Mobile Robot final year project for BEng (Electrical & Electronics) under the supervision of Dr Russell Brinkworth and Mr Charlie Green. Test 4 - Using ‘natural world’ images instead of black & white stripes and old newspapers. The vision system was tuned to work best with these types of images. The ‘tunnel’ was a bit too short for the robot. We’ll make longer ones soon. This is a work in progress, lots more to do. Control system is definitely not optimally tuned yet. Filmed on 28th April 2009.
5 Comments »
RSS feed for comments on this post. TrackBack URL
Leave a comment
You must be logged in to post a comment.
Depth is determined using ‘optic flow’. A good explanation is on Centeye’s website, centeye dot com (I cant post links here).
This company, centeye does research on biomimetic vision systems. Have a look.
@davepamn In theory, biomimetic systems are supposed to be highly efficient. Just look at how small a fly brain is, but that tiny thing can do extremely impressive flight maneuvers. But you’ll need to do some googling if you want specific performance comparison.
How does the biomimetic system translate in terms of performance? How is depth determined? Is depth a perception feedback input?
@davepamn A biomimetic system is basically a system that mimics / copies biology. So in this case, the vision system that is used has been developed by really studying how insect eyes work and then translate it into a software algorithm. Does that answer your question?
What is biomimetic vision?