Praying mantises have advanced vision with exceptional depth perception. (Patricia Chumillas/Shutterstock)

CHARLOTTESVILLE, Va. — Autonomous vehicles can navigate complex traffic patterns and recognize road signs, but they sometimes fail at a seemingly simple task: determining how far away a parked car is. This limitation shares a surprising connection with the insect world, where most compound eyes excel at detecting motion but struggle with depth perception. Enter an unlikely hero in solving this technological challenge: the praying mantis.

A team from the University of Virginia’s School of Engineering and Applied Science has developed high-tech artificial eyes that mimic those of insects. These eyes mark a big step forward in visual sensing technology. According to a study published in Science Robotics, the researchers built them using two curved surfaces, each packed with 256 tiny light sensors arranged in a 16-by-16 grid. To recreate the unique shape and structure of mantis eyes, they used flexible semiconductor materials, allowing the artificial eyes to capture images in a way similar to real insect vision.

“Making the sensor in hemispherical geometry while maintaining its functionality is a state-of-the-art achievement, providing a wide field of view and superior depth perception,” explains Byungjoon Bae, a Ph.D. candidate in the Charles L. Brown Department of Electrical and Computer Engineering at UVA, in a statement.

The way this artificial vision system works is remarkably elegant. Each eye contains light sensors similar to those in digital cameras, but with a crucial difference; they can process and store information right where they detect it. Traditional cameras must send all their visual data to a separate computer for processing, which takes time and energy. This new system handles much of the visual processing right at the source, similar to how insects process visual information in their eyes rather than sending raw data to their brains.

The artificial eyes are constructed using flexible materials that can be shaped into hemispheres, mimicking the curved surface of natural compound eyes. Each eye contains an array of tiny lenses that focus light onto the sensors beneath them. When an object moves through its field of view, both eyes track it independently and then share their information through a sophisticated computer network that combines their observations to determine the object’s position and movement in three dimensions.

This binocular vision system proves particularly effective because it combines two different ways of perceiving depth. First, it uses stereopsis, which is the same principle that gives humans depth perception by comparing slightly different views from two eyes. Second, it employs motion parallax, where closer objects appear to move faster across our field of view than distant ones. While many animals use one method or the other, praying mantises are unique among insects in using both, making them particularly adept at judging distances.

It can track objects with an error margin of just 0.3 centimeters, about the width of a pencil eraser. Even more remarkable is its processing speed: it can analyze visual information in just 1.8 milliseconds, fast enough to track rapidly moving objects in real time. All this while using minimal power—about 400 times less energy than conventional camera systems.

The researchers achieved this efficiency through an innovative approach called “edge computing” where data is processed as close as possible to where it’s collected. The system continuously monitors scenes for changes, creating compact data packages that require minimal processing power. This mirrors how biological vision systems work, focusing on changes in the visual scene rather than constantly processing everything in view.

One of the most promising aspects of this technology is its potential to improve the safety and reliability of autonomous vehicles. Current self-driving cars sometimes struggle to accurately judge the distance between stationary or slow-moving objects, leading to safety concerns. By incorporating this mantis-inspired vision system, future vehicles could better perceive their environment, potentially reducing accidents and improving navigation in complex environments.

Beyond autonomous vehicles, this technology could revolutionize other fields where accurate visual perception is crucial. Robotic systems could become more precise in assembly lines or more adept at navigating dynamic environments. Surveillance systems could track objects more accurately while using less power. Even smart home devices could benefit from improved depth perception and motion-tracking capabilities.

“The seamless fusion of these advanced materials and algorithms enables real-time, efficient, and accurate 3D spatiotemporal perception,” explains Kyusang Lee, an associate professor from UVA.

This research could mark the beginning of a new era in machine vision, where efficient, accurate depth perception becomes standard. By following nature’s blueprint, researchers have shown that sometimes the best innovations come from understanding and adapting existing biological solutions.