MIT News breed · Science

Nanoscale Mechanics Enable Brain-Inspired Computing

MIT researchers created a computing platform using nanoscale polymer mechanics for energy-efficient, compact electronics with integrated computing and memory, mimicking brain neurons.

This platform could advance low-power edge computing and smart robotics by enabling devices to perform multiple functions in a single unit.

They used soft polymers' mechanical response to create tiny devices that process and remember information like biological neurons.

Integrating functions into the material reduces component needs for a compact system.

Inspired by biological systems, the device computes through physical transformations at the nanoscale.

They overcame adhesive forces with a PDMS film between electrodes, acting as a 'nano-spring' for reversible reconfiguration.

When voltage is applied, PDMS compression alters current. Its viscoelasticity allows the device to 'remember' forces and voltages, acting as an artificial neuron.

This integrated approach offers high efficiency and a small footprint, with potential applications in prosthetics and wearables. Future work aims to integrate sensing.

AI-samenvatting op basis van de bron.