Self-driving cars have encountered issues with their visual systems due to difficulties in processing static or slow-moving objects in 3D space. This obstacle is akin to the monocular vision found in many insects, which exhibit great motion-tracking abilities and a wide field of view but struggle with depth perception.
Contrary to common insects, the praying mantis possesses a unique field of view that overlaps between its left and right eyes, granting it binocular vision with depth perception in 3D space. Drawing inspiration from these natural capabilities, researchers at the University of Virginia School of Engineering and Applied Science embarked on a mission to develop artificial compound eyes that could potentially revolutionize the way machines collect and process visual data in real-time.
Through a combination of meticulous design, optoelectrical engineering, and innovative “edge” computing techniques, the team successfully replicated the biological capabilities of praying mantis eyes. By integrating microlenses, multiple photodiodes, and flexible semiconductor materials, the artificial compound eyes were able to emulate the convex shapes and complex positions within mantis eyes.
The newly developed system showcases a wide field of view, superior depth perception, and precise spatial awareness, making it ideal for various applications such as low-power vehicles, drones, self-driving vehicles, robotic assembly, surveillance systems, security systems, and smart home devices. The groundbreaking technology has the potential to reduce power consumption by more than 400 times compared to traditional visual systems, paving the way for real-time processing without the need for external computation or excessive energy usage.
The key to the success of the sensor array lies in its ability to continuously monitor changes in the scene, differentiate between pixels, and encode this information into smaller data sets for processing. This approach mirrors how insects perceive the world through visual cues, allowing for rapid and accurate understanding of motion and spatial data.
The integration of advanced materials, conformal devices, in-sensor memory components, and post-processing algorithms has enabled the team to achieve real-time, efficient, and accurate 3D spatiotemporal perception. This significant scientific breakthrough serves as a testament to the power of biomimicry in addressing complex visual processing challenges, inspiring engineers and scientists to explore innovative solutions rooted in nature’s wisdom.
Leave a Reply