Additive Manufacturing of Neuromorphic Systems.

Journal: Advanced materials (Deerfield Beach, Fla.)
Published Date:

Abstract

Neuromorphic engineering aims to create brain-inspired computing systems based on synaptic electronic hardware and neural network software. It combines intelligent materials, advanced processing technology, and computation programs. Additive manufacturing (AM), despite being one of the advanced manufacturing technologies capable of multimaterial processing at the microscale, is not widely applied in neuromorphic hardware fabrication. This gap suggests not only process incompatibility and limited resolution of AM but opportunities to create novel intelligent systems. Here, the state-of-the-art in AM-printed neuromorphic hardware (synaptic electronics and mechanical systems) is reviewed and discussed the integration of AM techniques with neuromorphic engineering. An outlook of printed neuromorphic systems is provided with low cost and environmental impact but high customizability and design flexibility. With ongoing innovation in AM technologies and materials, AM is envisioned with high throughput and resolution for affordable, scalable, and customizable neuromorphic hardware production. The crossover of AM and neuromorphic engineering facilitates prototyping of brain-inspired computing architectures for efficient and analog computation. This approach may facilitate various applications including neuromorphic robotics, bionics, and real-time sensing.

Authors

  • Jiongyi Yan
    School of Cellular and Molecular Medicine, University of Bristol, Bristol, BS8 1TD, UK.
  • Yutai Su
    School of Mechanics, Civil Engineering and Architecture, Northwestern Polytechnical University, Xi'An, 710129, China.
  • James P K Armstrong
    Department of Translational Health Sciences, Bristol Medical School, University of Bristol, Bristol, BS1 3NY, UK.
  • Andrew Gleadall
    Wolfson School of Mechanical Electrical and Manufacturing Engineering, Loughborough University, Loughborough, LE11 3TU, UK.

Keywords

No keywords available for this article.