A desert locust-inspired wavelength-selective polarization vision system using two-dimensional ferroelectric...

Area of Science:

• Neuromorphic Engineering
• Materials Science
• Artificial Intelligence

Background:

• Traditional von Neumann architectures face limitations in integrating sensing, memory, and computing.
• Advanced bionic vision systems require polarization-sensitivity and wavelength-selectivity with high energy efficiency and fast, non-volatile storage.

Purpose of the Study:

• To develop a bionic polarization vision system (BPVS) capable of simulating polarization-sensitive and wavelength-selective vision.
• To achieve exceptional energy efficiency and high-speed, non-volatile storage in a bionic vision system.

Main Methods:

• Fabrication of a graphene (Gr)/CuInP2S6 (CIPS)/Gr/h-BN/PdSe2 (GCGhP) ferroelectric heterostructure.
• Utilizing optical excitation-induced ferroelectric polarization reversal for wavelength confinement (ultraviolet band).
• Leveraging PdSe2 anisotropy for polarization sensitivity and ferroelectric polarization for low power consumption.

Main Results:

• Demonstrated polarization-sensitive and wavelength-selective vision simulation.
• Achieved ultra-low power consumption: 0.15 pJ/7.44 fJ per spike (optical enhancement/electrical suppression).
• Successfully mimicked polarimetric synaptic plasticity, enabling memory imaging at multiple polarization angles and super-resolution reconstruction via a bionic neural network.

Conclusions:

• The developed GCghP heterostructure offers an effective pathway for advanced, energy-efficient polarization-sensitive bionic neuromorphic computing.
• This system advances the development of sophisticated bionic vision systems with integrated functionalities.