Towards efficient and adaptive cyber physical spiking neural integrated systems

Abstract:

This work introduces multi-sensor integration combined with an efficient and adaptive Spiking Neural Network (SNN) emulation architecture for local intelligent processing. For this purpose, we propose CMOS-MEMS with on-chip conditioning electronics together with spike processing by means of a real-time bioinspired and model-programmable SIMD multiprocessor. System integration considerations and results in the MEMS and processor developments are provided.

Año de publicación:

2020

Keywords:

• SNN
• Sensor Neural Computing
• Spiking Neural Networks
• CMOS-MEMS
• Cyber Physical Neural Systems
• Integrated Sensors
• Integration
• MEMs

Fuente:

scopus

google