A research team led by the School of Engineering of the Hong Kong University of Science and Technology (HKUST) has developed a liquid metal-based electronic logic device that mimics the intelligent prey-capture mechanism of Venus flytraps. The study is published in the journal Nature Communications.

Exhibiting memory and counting properties, the device can intelligently respond to various stimulus sequences without the need for additional electronic components. The intelligent strategies and logic mechanisms in the device provide a fresh perspective on understanding “intelligence” in nature and offer inspiration for the development of “embodied intelligence.”

The unique prey-capture mechanism of Venus flytraps has always been an intriguing research focus in the realm of biological intelligence. This mechanism allows them to effectively distinguish between various external stimuli such as single and double touches, thereby distinguishing between environmental disturbances such as raindrops (single touch) and insects (double touches), ensuring successful prey capture.

This functionality is primarily attributed to the sensory hairs on the carnivorous plants, which exhibit features akin to memory and counting, enabling them to perceive stimuli, generate action potentials (a change of electrical signals in cells in response to a stimulus), and remember the stimuli for a short duration.

Inspired by the internal electrical signal accumulation/decay model of Venus flytraps, Prof. Shen Yajing, Associate Professor of the Department of Electronic and Computer Engineering (ECE) at HKUST, who led the research, collaborated with his former Ph.D. student at City University of Hong Kong, Dr. Yang Yuanyuan, now Associate Professor at Xiamen University, to propose a liquid metal-based logic module (LLM) based on the extension/contraction deformation of liquid metal wires.

The device employs liquid metal wires in sodium hydroxide solution as the conductive medium, controlling the length of the liquid metal wires based on electrochemical effects, thereby regulating cathode output according to the stimuli applied to the anode and gate. Research results demonstrate that the LLM itself can memorize the duration and interval of electrical stimuli, calculate the accumulation of signals from multiple stimuli, and exhibit significant logical functions similar to those of Venus flytraps.

To demonstrate, Prof. Shen and Dr. Yang constructed an artificial Venus flytrap system comprising the LLM intelligent decision-making device, switch-based sensory hair, and soft electric actuator-based petal, replicating the predation process of Venus flytraps. Furthermore, they showcased the potential applications of LLM in functional circuit integration, filtering, artificial neural networks, and more.

Their work not only provides insights into simulating intelligent behaviors in plants, but also serves as a reliable reference for the development of subsequent biological signal simulator devices and biologically inspired intelligent systems.

“When people mention ‘artificial intelligence,’ they generally think of intelligence that mimics animal nervous systems. However, in nature, many plants can also demonstrate intelligence through specific material and structural combinations. Research in this direction provides a new perspective and approach for us to understand ‘intelligence’ in nature and construct ‘life-like intelligence,'” said Prof. Shen.

“Several years ago, when Dr. Yang was still pursuing her Ph.D. in my research group, we discussed the idea of constructing intelligent entities inspired by plants together. It is gratifying that after several years of effort, we have achieved the conceptual verification and simulation of Venus flytrap intelligence.

“However, it is worth noting that this work is still relatively preliminary, and there is much work to be done in the future, such as designing more efficient structures, reducing the size of devices, and improving system responsiveness,” added Prof. Shen.

A compact, lightweight sensor system with infrared imaging capabilities developed by an international team of engineers could be easily fitted to a drone for remote crop monitoring.

This flat-optics technology has the potential to replace traditional optical lens applications for environmental sensing in a range of industries.

This innovation could result in cheaper groceries as farmers would be able to pinpoint which crops require irrigation, fertilization and pest control, instead of taking a one-size-fits-all approach, thereby potentially boosting their harvests.

The sensor system can rapidly switch between edge detection—imaging the outline of an object, such as a fruit—and extracting detailed infrared information, without the need for creating large volumes of data and using bulky external processors.

The capability to switch to a detailed infrared image is a new development in the field and could allow farmers to collect more information when the remote sensor identifies areas of potential pest infestations.

This research by engineers at the City University of New York (CUNY), the University of Melbourne, RMIT University and the ARC Center of Excellence for Transformative Meta-Optical Systems (TMOS) is published in Nature Communications in a paper titled “Reconfigurable image processing metasurfaces with phase-change materials.”

How does the sensor system work?

The prototype sensor system, which comprises a filter made with a thin layer of a material called vanadium dioxide that can switch between edge detection and detailed infrared imaging, was engineered by TMOS Chief Investigator Professor Madhu Bhaskaran and her team at RMIT in Melbourne.

“Materials such as vanadium dioxide add a fantastic tuning capability to render devices ‘smart'”, she said.

“When the temperature of the filter is changed, the vanadium dioxide transforms from an insulating state to a metallic one, which is how the processed image shifts from a filtered outline to an unfiltered infrared image.”

“These materials could go a long way in futuristic flat-optics devices that can replace technologies with traditional lenses for environmental sensing applications—making them ideal for use in drones and satellites, which require low size, weight and power capacity.

RMIT holds a granted US patent and has a pending Australian patent application for its method of producing vanadium dioxide films, which may be suitable for a broad range of applications.

Lead author Dr. Michele Cotrufo said the system’s ability to switch between processing operations, from edge detection to capturing detailed infrared images, was significant.

“While a few recent demonstrations have achieved analog edge detection using metasurfaces, most of the devices demonstrated so far are static. Their functionality is fixed in time and cannot be dynamically altered or controlled,” said Corufo, who conducted his research at CUNY.

“Yet, the ability to dynamically reconfigure processing operations is key for metasurfaces to be able to compete with digital image processing systems. This is what we have developed.”

Next steps

Co-author Shaban Sulejman from the University of Melbourne said the design and materials used make the filter amenable to mass-manufacturing.

“It also operates at temperatures compatible with standard manufacturing techniques, making it well-placed to integrate with commercially available systems and therefore move from research to real-world usage rapidly.”

TMOS Chief Investigator Ann Roberts, also from the University of Melbourne, said flat optics technologies had the potential to transform countless industries.

“Traditional optical elements have long been the bottleneck preventing the further miniaturization of devices. The ability to replace or complement traditional optical elements with thin-film optics breaks through that bottleneck.”