Neural networks integrated circuit with switchable gait pattern for insect-type microrobot

摘要

Insects have millimeter-sized bodies but excellent autonomous control. Therefore, numerous researchers have been bio-inspired by their attractive features to create highly functional robots. In this research, we investigated a neural network integrated circuit (NNIC) that generated the driving waveform of an insect-type microrobot. Previously, we discussed an NNIC that had a switching mechanism for the gait pattern. It could switch between the tripod gait pattern and ripple gait pattern typical of insects. The simulation results revealed that the gait was unstable when changed by varying the coupling coefficient voltage of the synaptic model. In addition, the order of the ripple gait pattern had an error under certain conditions. In this paper, we compare the measurement results with the simulation results. As unstable gaits do not appear in the measurement results, the fabricated NNIC generated a more stable gait pattern than the simulation.