A CORDIC based real-time implementation and analysis of a respiratory central pattern generator

Xinyu Hao, Shuangming Yang, Bin Deng, Jiang Wang, Xile Wei, Yanqiu Che

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

Central pattern generators (CPGs) are dedicated neural circuits which can generate rhythmic motor patterns even in absence of sensory input with extraordinary robustness and flexibility. In this paper, a biologically realistic model of a respiratory CPG with four neurons is implemented on a reconfigurable Field-Programmable Gate Array (FPGA) system. Considering the limitations of hardware resources, we first propose a modified respiratory CPG model with Coordinate Rotation Digital Computer (CORDIC) algorithm to save limited resources and reduce complexity. And then, all the multipliers are replaced with a method which is appropriate and effective for hardware implementation to avoid the use of the area-intensive multipliers. The implementation results show that rhythmic oscillations are successfully generated by the respiratory CPG network and the resource utilization is greatly reduced, which shows the potential for building large-scale spiking neural networks. The proposed high-performance and real-time implementation of the respiratory CPG network on the FPGA system can speed up the process to gain new insights into the respiratory network and can also be developed into applications for respiratory rhythm generation and modulation.

Original languageEnglish (US)
Pages (from-to)373-388
Number of pages16
JournalNeurocomputing
Volume423
DOIs
StatePublished - Jan 29 2021

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Cognitive Neuroscience
  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'A CORDIC based real-time implementation and analysis of a respiratory central pattern generator'. Together they form a unique fingerprint.

Cite this