Cornell researchers and the ERC developed microrobots that synchronize via electronic pulses or AI “dreaming,” enabling new medical and environmental applications.
The studies, published in Science Robotics and Nature Machine Intelligence, detail decentralized control systems for microscale machines. In 2024, Cornell University researchers engineered micromachines that autonomously synchronize their movements using electronic pulses, a breakthrough published in Science Robotics on November 27. The system, led by Alyssa Apsel, the IBM Professor of Engineering and director of the School of Electrical and Computer Engineering, and Itai Cohen, professor of physics in the College of Arts and Sciences and faculty member in the Department of Design Tech, uses complementary metal-oxide-semiconductor oscillators to coordinate up to 16 devices in linear and two-dimensional arrays. Each machine features a 7-nanometer-thick bending paddle actuator, mimicking stadium waves through periodic electronic signals. We’re essentially designing local timing systems that communicate with each other to produce global behaviors. This approach is ideal for microscale machines that lack the power, capability or space to be wired over long distances,
Apsel said. The decentralized approach allows sub-groups to maintain synchronization even when separated, with potential applications in drug delivery, chemical mixing, and environmental remediation.
Micromachines autonomously coordinate using electronic pulses
The Cornell team’s pulsed-coupling technique aligns micromachines without centralized control, drawing inspiration from natural systems like firefly synchronization. Milad Taghavi, Ph.D. ’21, and co-lead researcher Wei Wang, Ph.D. ’23, noted that the system self-corrects under changing conditions. If a group becomes severed, each sub-group can synchronize independently, Taghavi explained. The oscillators, operating at sub-nanowatt power, enable complex swarms without wired connections. This scalability could lead to elastronic materials where electronics integrate into materials to create emergent behaviors, as theorized by Cohen. The research, supported by the National Science Foundation, the U.S. Army Research Office, the Cornell Center for Materials Research, and the Kavli Institute at Cornell for Nanoscale Science, marks the first demonstration of synchronization in micromachines with such oscillators. The team plans to continue work on micromachines, including coordinated microrobots that mimic inchworms or split into autonomous pieces.
A study introduced a new method for microrobots to “dream” in simulated environments, enabling real-time adaptation. Led by Daniel Ahmed, the project, supported by the European Research Council (ERC)’s Starting Grant SONOBOTS, used model-based reinforcement learning (RL) to train ultrasound-propelled microrobots in virtual microfluidic channels. We created virtual worlds that mimic the physics of real-world conditions, enabling the microrobots to learn how to move, avoid obstacles, and adapt to new environments,
Ahmed said. Experimental trials showed 90% success rates after one hour of fine-tuning, with performance improving from 50% to over 90% in unfamiliar settings. Mahmoud Medany, co-lead author, highlighted the potential for autonomous navigation in living systems. This shows that ultrasound-driven microrobots can learn to adapt in real time,
Medany said. The research, published in Nature Machine Intelligence, aims to revolutionize non-invasive medical procedures like drug delivery and microsurgery.
Teaching microrobots to dream
The Cornell and ERC approaches represent distinct strategies for microrobot coordination. While Cornell’s system relies on physical synchronization via electronic pulses, the ERC’s method leverages AI to simulate and adapt to dynamic environments. Both advancements address challenges in microrobot control, such as the impracticality of traditional sensors at microscopic scales. The Cornell team’s work emphasizes scalability for large swarms, while the ERC’s focus on ultrasound propulsion and AI opens pathways for real-time medical applications. Together, these developments underscore the growing potential of microrobots to transform fields from healthcare to environmental science.
Implications for medicine and technology include precise, non-invasive interventions. Synchronized arrays of micromachines might target hard-to-reach tumors or clear vascular blockages, while AI-trained devices could navigate complex bodily environments. The Cornell team’s elastronic materials concept suggests future applications where electronics embedded in materials exhibit novel behaviors. Meanwhile, the ERC’s research highlights the feasibility of real-time adaptation, a critical factor for microrobots operating in unpredictable biological systems. Both studies reflect a broader trend in microtechnology toward autonomy and adaptability, driven by interdisciplinary insights from physics, biology, and artificial intelligence. Syl Kacapyr is associate director of marketing and communications for Cornell Engineering.