Memory-electroluminescence for multiple action-potentials combination in bio-inspired afferent nerves.

Area of Science:

• Optoelectronics
• Artificial Intelligence
• Neuroscience

Background:

• Mimicking biological nervous system functions is crucial for advancing artificial intelligence.
• Optoelectronic systems offer potential for creating artificial neural networks.

Purpose of the Study:

• To develop an optoelectronic artificial afferent nerve strategy.
• To enable multiple action-potential combinations via a single optical channel.
• To demonstrate sensor-position recognition using the bio-inspired system.

Main Methods:

• Utilized memory-electroluminescence spikes with history-dependent characteristics to encode sensor signals.
• Proposed a non-carrier injection mode to drive nanoscale light-emitting diodes for generating multi-sub-peak spikes.
• Employed wavelength-multiplexing for spike signals to achieve large signal bandwidth.

Main Results:

• Demonstrated the generation of memory-electroluminescence spikes with diverse morphologies.
• Successfully transmitted multiple action potentials through a single optical channel.
• Achieved a high sensor-position recognition accuracy of 98.88%.

Conclusions:

• The developed strategy effectively mimics biological afferent nerves.
• The memory-electroluminescence spike-based system offers a novel approach for artificial perception.
• This work provides insights for constructing advanced artificial intelligence systems.