The objective of this research is to discover new fundamental principles, design methods, and technologies for realizing distributed networks of sub-cm3, ant-sized mobile micro-robots that self-organize into cooperative configurations. The approach is intrinsically interdisciplinary and organized along four main thrusts: (1) Algorithms for distributed coordination and control under severe power, communication, and mobility constraints. (2) Electronics for robot control using event-based communication and computation, ultra-low-power radio, and adaptive analog-digital integrated circuits. (3) Locomotion devices and efficient actuators using rapid-prototyping and MEMS technologies that can operate robustly under real-world conditions. (4) Integration of the algorithms, electronics, and actuators into a fleet of ant-size micro-robots.
No robots at the sub-cm3 scale exist because their development faces a number of open challenges. This research will identify and determine means for solving these challenges. In addition, it will provide new solutions to outstanding questions about resource-constrained algorithms, architectures, and actuators that can be widely leveraged in other applications. The PIs will adopt a co-design philosophy that promotes cross-disciplinary research and tight collaboration.
Networks of ant-sized robots are expected to be useful in disaster relief, manufacturing, warehouse management, and ecological monitoring, as well as in new unforeseen applications. In addition, the new methods and principles proposed here can be transitioned to other highly-distributed and resource-constrained engineering problems, such as air-traffic control systems. This research program will train Ph.D. students with unique skills in the design of hybrid distributed networks and it will involve undergraduate students, particularly underrepresented minorities and women.
Micro-robots will revolutionize the use of engineered systems in tasks that require intervention at small scales in constrained spaces that are not reachable by a human. Possible uses are endless and include assisted micro-surgery, micro-fabrication, equipment monitoring and repair, and infrastructure inspection. In spite of the notable progress in miniaturization technology, the development of functional tiny robots presents challenges that require the investigation of new fundamental principles and methods. These robots are powered by minuscule batteries that require the development of small and low-power locomotion methods, software and algorithms that have to be extra-efficient to run on tiny low-power micro-processors to perform coordination tasks subject to very limited short range communication. Equally important is the development of methods to manufacture and package tiny parts and system integration techniques to make them work together effectively. This project supported an interdisciplinary group of graduate and undergraduate students to work on the aforementioned challenges under the supervision of four faculty affiliated with the departments of Electrical and Computer Engineering and Mechanical Engineering of the University of Maryland at College Park. The multi-disciplinary background gained through this project allowed students, who already graduated, to pursue successful careers in industry and academia. The main results were reported in top journals and peer-reviewed conference papers. The following is a summary of some of the highlights of our research. Previous small scale robots have been limited to slow, tethered treks if they were able to move at all. Many of the failures were due to lack of mechanical robustness in their environment – a failure that we have begun to address through the introduction of new, softer materials in microfabrication. These critical contributions were acknowledged with a cover photo and "IOP Select Article" in two recent journal papers. We are also excited about use of these same processes to create the first batch-microfabricated, entirely soft dielectric elastomer actuators with composite electrodes. We also developed an open resource hardware and software platform (TinyTeRP), which is a resource contribution to research and education. These robots have been featured in an NSF Science Nation video and have continually attracted bright undergraduate students to research. We have developed methods for creating chip-scale robots: chips that can walk because they have moving legs and know how to operate them. The combination of integrated circuits (ICs) with actuators required the development of novel hardware and microfabrication techniques. Since the chip surface is covered with circuits for programmable walking gaits, energy-efficient actuation, and computation, we needed to use exclusively surface micromachining, and it had to be compatible with commercially-available complementary metal-oxide-semiconductor (CMOS) technology. To enable the chip-scale robot to support its own weight and to walk forward, we optimized the actuator structure. Furthermore, to provide power to these tiny chips (3 mm) we needed flexible wire connections that did not interfere with their motion. We were pleased to see actuation in prototype devices and hope to report on the first legged chip with onboard control in the very near future. Various micro-scale actuators were investigated, including surface micromachined dielectric elastomer actuators, thermal actuators, and nastic actuators. The latter are hydraulic actuators within an elastomer operated by electro-osmotic pumping. The most important improvement to the nastic technology was the discovery of a pumping fluid that does not generate gas bubbles when high voltage is applied, which will allow large forces and stable performance to be achieved. Other technology advances include the demonstration of carbon/elastomer composite electrodes for pumping and the development of meso-scale fabrication processes. Our work on algorithms unveiled new methods to design controllers for small robots operating individually and in groups or swarms. In contrast with previous approaches, ours is guaranteed to optimize energy and performance subject to limits on inter-robot information exchange, such as when the network is fragmented due to the limitations imposed by low power wireless communication that is viable only among robots within close range. We participated in various outreach activities throughout the duration of the project, including University of Marylandâ€™s Robotic Day, National Robotics Week at the Air and Space Museum, Robotics Week at Koshland Science Museum, 2011 ISCAS in Rio de Janeiro Brazil, and University of Marylandâ€™s 4-H Youth Program. The PIs were invited to give seminars, talks, and tutorials at venues around the country and abroad. The main ideas and outcomes have also received media attention, such as the recent TED talk by Bergbreiter and the article at the Pacific Standard entitled "Antbots to the rescue" - June 7th, 2012. Motivated by the new research directions that resulted from this project, the Institute for Systems Research (University of Maryland) provided space and resources for the creation of the CPS and Cooperative Autonomy laboratory, where two other NSF CPS projects are being developed and student research activities continue to be supported at the graduate and undergraduate levels.