NRN — Neuromorphic Research Node

Neuromorphic chip architectures for adaptive edge intelligence. Exploring low‑power neuron and synapse circuits, memory and learning mechanisms, and mixed‑signal system integration for practical edge inference.

CollaborateRoadmapTechnical NotesContact

Mission

NRN is an independent research initiative exploring neuromorphic chip architectures for adaptive edge intelligence systems.

The focus is on designing low‑power neuromorphic architectures capable of processing biosignals (e.g., EMG) directly at the sensor edge. Research explores neuron and synapse circuit implementations, programmable synaptic memory, event‑driven computation, and mixed‑signal system integration for efficient edge inference.

Research Roadmap

NRN is being developed as a staged neuromorphic chip research program, progressing from core circuit validation toward scalable system integration and prototype silicon.

Stage 1

Neuron Core Validation

  • LIF neuron design
  • Biasing + operating point validation
  • Power / spike characterization
Stage 2

Synapse + Memory Architecture

  • SRAM-based weight storage
  • Read / write architecture
  • Weight-to-current interfacing
Stage 3

Learning + Adaptation

  • Hybrid training flow
  • Local update mechanisms
  • Spike-driven adaptation logic
Stage 4

Array-Level Integration

  • Multi-neuron scaling
  • Event routing
  • Array-level measurements
Stage 5

Prototype Silicon Path

  • Layout-ready architecture
  • Tape-out planning
  • Publication / translation path

Technical Notes

NRN maintains concise technical notes to document architecture decisions, circuit explorations, and system-level research direction as the program evolves.

NRN-TN-001
Done

LIF Neuron Circuit Direction

Design direction for low-power LIF neuron implementations, including operating regimes, spike generation choices, and circuit-level validation goals.

NRN-TN-002
In Progress

SRAM Synapse + Memory Architecture

Exploration of SRAM-based weight storage, synapse read/write paths, and architectural trade-offs for scalable neuron–synapse systems.

NRN-TN-003
Planned

Hybrid Learning for Edge Neuromorphic Systems

Study of offline training with lightweight on-chip local adaptation for adaptive edge inference and biosignal-driven applications.

Early notes are currently internal working documents and will be expanded into formal technical writeups as results mature.

Collaboration

NRN welcomes technical collaboration in the following areas:

  • Analog IC design and layout
  • Mixed-signal integration
  • FPGA-based neuromorphic inference
  • SNN modeling and training
  • Biosignal processing (EMG/EEG)

Note: NRN is currently a founder-led research initiative. Contributions are structured as collaborative research efforts; future commercialization, if pursued, would involve separate agreements.

Founder

Paul Adu

M.S. Electrical & Computer Engineering • Research focus: Analog VLSI, Neuromorphic IC Design, Edge AI Systems

Email: pbadu816@gmail.com • LinkedIn: Paul Adu