A robust pattern of neuronal response to outer phenomena in "Vitroid", the hybrid neuro-robot

概要

論文の詳細を見る
Rat hippocampal neurons reorganized a complexnetworks on microelectrodes array dish. The living neuronalnetwork can distinguish patterns of action potentials evoked bydifferent inputs, suggesting that a cultured neuronal networkcan represent particular states as symbols. A neuro-robot-hybridsystem with living neuronal network and miniature moving robotwas developed. We use a Khepera II robot for interfacing witha living neuronal network and the outer world and succeeded inperforming collision avoidance behavior with premised controlrule sets. Using self-tuning fuzzy reasoning, we associated adistinct spatial pattern of electrical activity with a particularphenomenon in the outside of the culture dish. The particularrelationship between network activity and outer phenomenonwas performed by control rules of electrical stimulation to theneuronal network, responding to outer phenomenon, while aspatio -temporal patterns of neuronal activity were linked tooutput devices by premised control rules. We succeeded inperforming collision avoidance, and found that fluctuation ofneuronal responses evoked by sensor output of robot body wascontrolled by the interaction between neurons via synapses.
日本知能情報ファジィ学会の論文