Related Experiment Video
Updated: Jul 6, 2026
10:18
Photodiode-Based Optical Imaging for Recording Network Dynamics with Single-Neuron Resolution in Non-Transgenic Invertebrates
Published on: July 9, 2020
Standard cell-based implementation of a digital optoelectronic neural-network hardware
Applied Optics
|March 22, 2008
Summary
A novel digital optoelectronic neural network architecture was designed using standard cell-based implementation. This microelectronic circuit design offers a significant performance advantage over purely electronic neural networks.
Area of Science:
- Digital electronics
- Optoelectronics
- Neural network architectures
Background:
- Multilayer perceptron networks are computationally intensive.
- Existing electronic neural networks face performance limitations.
- Optoelectronic components offer potential for faster computation.
Purpose of the Study:
- To present a standard cell-based implementation of a digital optoelectronic neural network.
- To define the structure and components of the optoelectronic multilayer perceptron.
- To describe the design process and identify tool limitations.
Main Methods:
- VHDL-based modeling and synthesis for circuit design.
- Partially automatic placement and routing for microelectronic layout.
- Standard cell-based design approach for optoelectronic systems.
Main Results:
- The overall structure of the multilayer perceptron network was defined.
- The optoelectronic interconnection system between layers was detailed.
- The microelectronic circuit layout for one layer was successfully designed.
Conclusions:
- A viable standard cell-based design approach for optoelectronic systems was established.
- The designed optoelectronic neural network layer shows potential for performance exceeding purely electronic networks.
- Shortcomings in current design tools for optoelectronic systems were identified.
