The ILB-25 chooses itself of its decisional area according to the length of the vectors (vector = signature of the controlled object)
Learning of the neuronal network
The neuronal system is based on expanding space, new vectors learning fine-tune the whole neuronal network. Complex limit cases can be learned to the system to improve its specificity and selectivity.
Use of a neuronal learning
When a new product vector is presented to neuronal network, 4 cases may be found:
- product vector matches with fail region,
- product vector matches with pass region,
- product vector is in a region without learning, the result is in this case “unknown”,
- product vector is in uncertain region, but analysis of environment help to find a solution.
|Example of sorting on neuronal basis, sorting of pills :|