| Popis: |
This paper presents integrated silicon-based material that can be configured to sense, perform different classification algorithms through neural computing, and produce an action signal all at the same physical layer. The algorithms will be coded in the mechanical responses of the sensing elements of multiple coupled micro-electro-mechanical systems (MEMS) that also simultaneously capture acceleration measurements to produce an actuated signal. This all-in-one smart material consumes near zero power and runs with zero circuitry. As a demonstration, a material made of three MEMS neurons is designed and fabricated to perform successfully both simple signal classification and activity recognition problems. This smart material will enable emergent technologies such as soft robotics and wearable devices to locally perform complex computations powered by permanent batteries. |
