Ultrasensitive Mesh-Structured Position-Sensitive Detectors Enabling Eye Tracking for Human-Machine Interaction.

Eye tracking provides highly efficient, intuitive, and seamless human-machine interaction by measuring eye movements and revealing user attention and intention. Currently, video-based eye tracking systems, while achieving sub-degree accuracy, suffer from high power consumption due to data acquisition and processing of redundant pixels. Here, we present an energy-efficient eye-tracking system employing a position-sensitive detector (PSD) with a graphene photoconductive mesh, which operates with only four readout signals, reducing data volume by over three orders of magnitude compared to video-based systems, while achieving superior accuracy. With the synergistic effects of lateral photoelectric and interfacial gating, the mesh-structured PSD operates under eye-safe ultra-weak light (as low as 10 pW) without a detection dead zone. Leveraging the real-time, precise tracking capability of the PSD, the eye-tracking system determines gaze orientation by tracking the reflected trajectories of light beams from the cornea, achieving a 1 kHz tracking frequency and < 0.1° angular accuracy. Human-machine interaction applications such as eye-controlled typing and a Gluttonous Snake game are demonstrated using the PSD-enabled eye tracker. The PSD-based eye tracking holds significant promise for the application of human-machine interaction in miniaturized systems.