Optical Localization to 0.1mm? No problemo!

Quick recap, “the mission”: we want to build an affordable open source industry quality robot arm. We think we can make it out of low quality components but stick an accurate sensor on the end effector (the ‘hand’) and sidestep the huge costs of precision mechanical components. Web cams are cheap, and people already own computers capable of the required visual processing, so we think that optical localization is a better strategy for many applications that researchers, entrepreneurs and engineers might like to do – if we can get the optical localization accurate enough.

Last time we calibrated our £18 Microsoft LiveCam 3000, and experimented using Vision Visp’s moving edge tracker. We found that the moving edge required a highly accurate object and we could not build one to the 0.1mm accuracy target we set ourselves. Furthermore, the actual moving edge tracker was not particularly CPU friendly, so we decided to pursue a different strategy this time.


May Music Night this Friday

This Friday, 18th May, is our monthly music night, starting from around 8pm in the lab. Come along to talk about music, hacking, and make some noise! The month before last I wasn’t there, and I missed out on Andrew and Tom’s resistive fabric synth, and James and Matt of Madlab making instruments out of[…]