Patent · US Active

AI-driven self adapting microelectronic circuits

US11416664B2 · kind B2 · utility

2Cited by
0References
2Claims
0Family size

Assignee

Inventors

Key dates

Filing dateNov 3, 2020
Grant dateAug 16, 2022
Priority date
Expiry dateNov 3, 2040

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06N3/063
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

The present description relates to a method based on artificial intelligence to implement a wide range of microelectronic circuits that can adapt by themselves to the usage conditions (e.g. loading changes), manufacturing variances or defects (e.g. process variations, device parameter mismatches, device model inaccuracies or changes, etc.), as well as environmental conditions (e.g. voltage, temperature, interference) in order to negate all or part of their effects on the circuit performance characteristics and achieve a very tight set of specifications over the wide range of conditions. Each microelectronic circuit is represented by a neural network model whose behavior is a function of the actual input signals, the usage and environmental conditions. An attached AI engine will infer from the model, the input signals, the usage conditions and the environmental conditions and create the adaptive changes required to modify the microelectronic circuit's behavior to negate all or part of their effects on the circuit performance characteristics and to achieve a very tight set of specifications.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.