Toward human-like adaptability in robotics through a retention-engineered synaptic control system
발표자
()
초록
내용
Although advanced robots can mimic human movement and aesthetics, they cannot adapt or evolve based on external experiences. To address this, we propose an innovative approach using parallel-processable retention-engineered synaptic devices in the control system to simulate a human-like learning system without complex computational systems. By modulating the retention properties of these devices through adjusting Ag/AgCl ink, the voltage drop across the gate electrode and electrolyte interface was altered. The free movement of ions in the electrolyte enhanced signal multiplexing in the ion gel, enabling parallel processing. Integrating these synaptic devices with actuators, we emulated a human-like workout process with feedback between acute and chronic responses. This control system reduces complexity and achieves human-like learning in biomimicry.