Recently, the Micro nano Manufacturing and System Integration Research Center of the Chinese Academy of Sciences Chongqing Green Intelligent Technology Research Institute published a research paper entitled Bionic visual audio photodetectors with in sensor performance and preprocessing on Science Advances. This study proposes a biomimetic "audiovisual" photodetector that integrates "audiovisual" signal detection and preprocessing functions at the sensing end by simulating the "excitation" and "inhibition" behaviors between synapses in the human perception system.
The eyes and ears, as the two important organs that dominate vision and hearing, play a crucial role in human perception of external information. Traditional perception technology relies on independent optical and acoustic detection devices, and the data obtained by the detection module is filled with redundant background interference information, which puts pressure on the conversion, transmission, and storage of data and restricts the operational efficiency of the perception system.
Drawing inspiration from human sensory organs, the development of a new generation of neuromorphic sensors has become a key driving force for the development of systems towards intelligence and integration. In human visual and auditory organs, cell and neural circuits not only have the function of detecting external excitation signals, but also can achieve preprocessing of detection signals, which is conducive to eliminating background interference information and accelerating brain computing speed. At present, due to the essential differences in the detection mechanisms of visual and auditory signals, it is difficult to achieve dual-mode detection through a single device. Therefore, there is an urgent need to develop new neural mimetic sensors that can simultaneously detect and preprocess "audio-visual" information.
This study proposes a biomimetic "audiovisual" photodetector based on a graphene germanium heterojunction field-effect transistor structure, which integrates visual and auditory sensing functions on a single device through optical methods to mimic the intelligent perception of sensory cells towards audiovisual signals. By programming gate voltage and utilizing continuously adjustable positive and negative photocurrents to dynamically simulate the "excitation" and "inhibition" behaviors in neural pathways, this device achieves preprocessing operations such as flipping, sharpening, and edge extraction on visual images, and proves its ability to classify target character images. In addition, the device can regulate the strength of the collected sound wave signal, and achieve the restoration of piano tones and Chinese English dialogue audio based on the recorded sound wave "shape". This graphene based device, which integrates "audio-visual" signal detection and preprocessing functions, effectively simplifies the hardware structure and provides new ideas for promoting the development of miniaturization, high integration, multifunctional, and intelligent sensing systems.
The research work has received support from national key research and development plans, etc.
Paper link：
论文链接
Schematic diagram of the working principle of a biomimetic audio-visual photodetector that mimics the human perception system. PPC, positive photocurrent; ZPC, no photocurrent; NPC, negative photocurrent.
