Due to the shortcomings of traditional pattern recognition, conventional robots are almost blind. We make your robots see - like a human being.
A new generation of low cost, highly autonomous robots results.
Please contact us for finding a solution for your specific application.
As an example we have built a little robot, called Egon, which orders little objects just by itself:
Egon recognizes objects in a reasonably natural environment: the objects are placed on a wooden table, the surface is not painted, illumination is quite natural, too, with strong shadows. One and the same program does not only recognize quite different objects with totally different features, but it also can easily be extended to recognize additional new objects – this can be done in a “linear process” : each new object will need one more subroutine, which means that the length of the program will increase in a linear, not in an exponential way.
In case you have personally seen any other device in this world performing like Egon, please do not hesitate to contact us.
Egon uses one web cam (30 Euro). But with isomorph computing, stereo vision is trivial – an additional webcam and about one man year of work would extend Egon to a 3-D system.
Note, that Egon is somewhat slow. That is, because it is still running on a conventional computer (Von Neumann). On the CELL microprocessor (IBM) it would be 20 times faster.
Our work on robotics is an application example of linear pattern recognition, derived from our research on the physics theory of information