U.S. Develops Human-Eye-Inspired Memristor to Boost Autonomous Vehicles' Vision in Bright Light

Researchers at Pennsylvania State University in the United States have developed a new optical device called a "photomemristor" that mimics the human eye's adaptation mechanism under drastic changes in light intensity, aiming to address the problem of perception failure in autonomous vehicles operating in environments with intense glare or rapidly alternating light and dark conditions. The device combines titanium dioxide with a PEDOT:PSS hydrogel material; exposure to light triggers the material to either absorb moisture and swell or dehydrate and contract, enabling autonomous adjustment of its light sensitivity. Unlike conventional cameras—which first capture images and then process them—the photomemristor integrates both light-sensing and data-storage capabilities, functioning similarly to a biological neuron. In experiments, a miniature array of 16 such units, coupled with an AI network, achieved 95% accuracy after just seven training rounds in recognizing dark letters against high-contrast backgrounds. The research team has filed a provisional patent application and plans to apply this technology to industrial robots and visual-assistive devices for people with visual impairments.