Layered Network of Representatives for Control of Robot
Shinji Karasawa · 2014
A decoder translates to one output from many inputs. Concurrent excitations of neurons, such as eyes and ears, are decoded by an annex neuron in a brain [1]. The pattern of data to decode is memorized in newly added circuit. Such reaction can be understood as a representative of intelligence. Subroutine in program will correspond to a series of such reactions. The intelligent system is created by bottom-up. But it is operated by top-down. Output of representative is able to assign a series of reaction in lower layer. The reaction in bottom layer is operated along real time. As for upper layered activation, the next operation is started by a signal of the next segmentation. The system of layered representatives can economically achieve many of reliable signal-processing [2].