Prof. CHAI Yang, Core Member of the Research Institute for Intelligent Wearable Systems (RI-IWEAR), collaborated with researchers at Yonsei University in Korea to publish an article titled “Optoelectronic Graded Neurons for Bioinspired In-sensor Motion Perception” in Nature Nanotechnology.
Motion perception by conventional machine vision usually entails significant computational resources, which greatly restricts its application at edge terminals. For perception of dynamic motion at sensory terminals, efficient visual processing hardware is required. Flying insects have a tiny visual system (~800 photoreceptors and 105 neurons in the brain) that allows them to detect motion with great agility. Prof. Chai and his team were inspired to emulate these characteristics with hardware devices for in-sensor motion perception.
The agile visual system of a flying insect is facilitated by its graded neural structure, which exhibits a much higher information transmission rate (>1000 bit/s) than that of a spiking neuron (~300 bit/s) and allows spatiotemporal information to be fused at sensory terminals. Prof. Chai’s team used a MoS2 phototransistor to emulate the non-spiking graded neurons of insect vision systems. The charge dynamics of the shallow trapping centres in MoS2 phototransistors mimic the characteristics of graded neurons, exhibiting multilevel response and high volatility.
The optoelectronic graded neuron array can directly perceive different types of motion. The bioinspired sensor array detects trajectories in the visual field with very economical hardware devices, allowing the efficient perception of the direction of moving objects. By modulating the charge dynamics of the shallow trapping centres in the MoS2 phototransistor, the bioinspired sensor array can recognise the motion with a temporal resolution ranging from 101 to 106 ms.
The bioinspired sensors have potential applications in robotics and artificial intelligence. For example, the sensors could be used to develop more efficient robots with better abilities to detect and respond to moving objects in the surrounding environment.
