US 12,291,446 B2
Artificial intelligence-based analog sensors and wearable devices incorporating the same
Fadi Alsaleem, Omaha, NE (US); and Mohammad H. Hasan, Omaha, NE (US)
Assigned to NUtech Ventures, Lincoln, NE (US)
Filed by NUtech Ventures, Lincoln, NE (US)
Filed on Aug. 4, 2021, as Appl. No. 17/394,169.
Claims priority of provisional application 63/060,903, filed on Aug. 4, 2020.
Prior Publication US 2022/0041433 A1, Feb. 10, 2022
Int. Cl. B81B 7/00 (2006.01); G05B 13/04 (2006.01); G06N 3/044 (2023.01); G06N 3/063 (2023.01); G06N 3/08 (2023.01)
CPC B81B 7/008 (2013.01) [G05B 13/04 (2013.01); G06N 3/044 (2023.01); G06N 3/063 (2013.01); G06N 3/08 (2013.01); B81B 2201/0235 (2013.01); B81B 2203/0118 (2013.01); B81B 2203/04 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method of operating a micro-electro-mechanical-systems (MEMS) device as a reservoir computer, the method comprising:
sensing an input signal;
generating a modulated input signal based on the input signal, wherein the input signal is a continuous signal and the modulated input signal is a discretized signal;
generating a MEMS deflection signal based on the modulated input signal and a time-delayed MEMS deflection signal;
sampling the MEMS deflection signal N times during a time internal T to generate a MEMS deflection matrix, wherein MEMS deflection matrix has a size M×N, wherein N corresponds to a number of virtual nodes of the reservoir computer and M is a number of time steps of time interval T;
receiving a trained weight matrix, wherein the trained weight matrix is trained by linear regression;
multiplying the MEMS deflection matrix by the trained weight matrix to generate an output signal, wherein the output signal comprises an output of the reservoir computer; and
classifying the input signal based on the output signal, wherein the sensing and the classifying of the input signal are performed simultaneously based on interactions of the virtual nodes of the reservoir computer, and wherein the interactions occur utilizing delayed feedback.