A research team from the School of Engineering at the Hong Kong University of Science and Technology (HKUST) has introduced an artificial compound eye system. This new system is not only more cost-effective but also exhibits sensitivity levels at least twice as high as existing market products within small areas. The development holds significant promise for transforming robotic vision, potentially enhancing robots’ capabilities in navigation, perception, and decision-making, and encouraging further development in human-robot collaboration as well as commercial applications.
The artificial compound eye system mimics the visual abilities of natural compound eyes and is applicable in a wide range of scenarios. For instance, it can be installed on drones to improve their accuracy and efficiency in tasks such as irrigation or emergency rescue operations at disaster sites. With its heightened sensitivity, the system is expected to facilitate closer collaboration between robots and other connected devices. Over the long term, this technology could enhance safety in autonomous driving and accelerate the adoption of intelligent transportation systems, which would, in turn, foster the development of smart cities.
This innovative technology was developed under the leadership of Prof. Zhiyong Fan, Chair Professor at HKUST’s Department of Electronic & Computer Engineering and Department of Chemical & Biological Engineering. It represents a significant advance in the field of biomimetic vision systems.
Historically, roboticists have focused on replicating the visual systems of insects, which offer a broad field of view and advanced motion-tracking capabilities. However, the integration of compound eye systems into autonomous platforms such as robots or drones has presented numerous challenges. These challenges include the complexity and instability of the systems during deformation, geometric constraints, and potential mismatches between optical components and detectors.
To overcome these obstacles, Prof. Fan’s team developed a pinhole compound vision system by employing new materials and structures. This system boasts several key features, including an inherent hemispherical perovskite nanowire array imager with high pixel density, which serves to expand the imaging field. Additionally, the system incorporates a 3D-printed, lens-free pinhole array with a customisable layout designed to regulate incoming light and eliminate blind spots between neighbouring ommatidia—the individual units that make up an insect’s compound eye.
Thanks to its angular selectivity, wide field of view, broad-spectrum response in both monocular and binocular configurations, and dynamic motion-tracking capability, the pinhole compound eye can not only accurately locate targets but also track moving objects, such as a quadruped robot when incorporated onto a drone.
Prof. Fan highlighted the simplicity, lightweight nature, and affordability of this compound eye design. While it may not completely replace traditional cameras, it is expected to provide significant benefits in specific robotic applications, such as drone swarms flying in close formation. Furthermore, by miniaturising the device and increasing the number of ommatidia, imaging resolution, and response speed, this type of system could find widespread applications in both optoelectronics and robotics.
Prof. Fan is a well-regarded researcher in biomimetic optoelectronics, known for combining practical approaches with bold ideas to drive innovative research. This compound eye system marks another milestone in the field of vision and robotic systems, following his previous development of the world’s first spherical artificial eye with a 3D retina in 2020.
The team’s research has been recognised at an international level, with the work being published and featured as a cover article in the journal Science Robotics. The research team includes Dr Yu Zhou and Dr Zhibo Sun, both postdoctoral researchers, as well as Yucheng Ding, a PhD student, who are credited as co-first authors, with Prof. Fan serving as the corresponding author.