Power Consumption Estimation of Digital Predistortion based on Spiking Neural Networks
Résumé
Different neuromorphic circuits have been proposed in recent years for low-power implementation of spiking neurons. However, there are very few clear quantitative studies on the power consumed by these spiking neurons as well as their networks. In this paper, we present a study on power consumed by some recently designed analog electronic neurons. An application of different structures of spiking neural network (SNN) to digital predistortion for MIMO antenna array linearization through over-the-air (OTA) feedback is taken as an example for numerical analysis, which demonstrates the effectiveness of SNN.
Origine | Fichiers produits par l'(les) auteur(s) |
---|