Highly dexterous robot hand even works in the dark
Researchers at Columbia Engineering have demonstrated a highly dexterous robot hand, one that combines an advanced sense of touch with motor learning algorithms in order to achieve a high level of dexterity.
As a demonstration of skill, the team chose a difficult manipulation task: executing an arbitrarily large rotation of an unevenly shaped grasped object in hand while always maintaining the object in a stable, secure hold. This is a very difficult task because it requires constant repositioning of a subset of fingers, while the other fingers have to keep the object stable. Not only was the hand able to perform this task, but it also did it without any visual feedback whatsoever, based solely on touch sensing.
In addition to the new levels of dexterity, the hand worked without any external cameras, so it's immune to lighting, occlusion, or similar issues. And the fact that the hand does not rely on vision to manipulate objects means that it can do so in very difficult lighting conditions that would confuse vision-based algorithms--it can even operate in the dark.
“While our demonstration was on a proof-of-concept task, meant to illustrate the capabilities of the hand, we believe that this level of dexterity will open up entirely new applications for robotic manipulation in the real world,” said Matei Ciocarlie, associate professor in the Departments of Mechanical Engineering and Computer Science. “Some of the more immediate uses might be in logistics and material handling, helping ease up supply chain problems like the ones that have plagued our economy in recent years, and in advanced manufacturing and assembly in factories.”
Leveraging optics-based tactile fingers
In earlier work, Ciocarlie’s group collaborated with Ioannis Kymissis, professor of electrical engineering, to develop a new generation of optics-based tactile robot fingers. These were the first robot fingers to achieve contact localization with sub-millimeter precision while providing complete coverage of a complex multi-curved surface. In addition, the compact packaging and low wire count of the fingers allowed for easy integration into complete robot hands.
Teaching the hand to perform complex tasks
For this new work, led by CIocarlie’s doctoral researcher, Gagan Khandate, the researchers designed and built a robot hand with five fingers and 15 independently actuated joints--each finger was equipped with the team’s touch-sensing technology. The next step was to test the ability of the tactile hand to perform complex manipulation tasks. To do this, they used new methods for motor learning, or the ability of a robot to learn new physical tasks via practice. In particular, they used a method called deep reinforcement learning, augmented with new algorithms that they developed for effective exploration of possible motor strategies.
Robot completed approximately one year of practice in only hours of real-time
The input to the motor learning algorithms consisted exclusively of the team’s tactile and proprioceptive data, without any vision. Using simulation as a training ground, the robot completed approximately one year of practice in only hours of real-time, thanks to modern physics simulators and highly parallel processors. The researchers then transferred this manipulation skill trained in simulation to the real robot hand, which was able to achieve the level of dexterity the team was hoping for. Ciocarlie noted that “the directional goal for the field remains assistive robotics in the home, the ultimate proving ground for real dexterity. In this study, we've shown that robot hands can also be highly dexterous based on touch sensing alone. Once we also add visual feedback into the mix along with touch, we hope to be able to achieve even more dexterity, and one day start approaching the replication of the human hand.”
Ultimate goal: joining abstract intelligence with embodied intelligence
Ultimately, Ciocarlie observed, a physical robot being useful in the real world needs both abstract, semantic intelligence (to understand conceptually how the world works), and embodied intelligence (the skill to physically interact with the world). Large language models such as OpenAI’s GPT-4 or Google’s PALM aim to provide the former, while dexterity in manipulation as achieved in this study represents complementary advances in the latter.
For instance, when asked how to make a sandwich, ChatGPT will type out a step-by-step plan in response, but it takes a dexterous robot to take that plan and actually make the sandwich. In the same way, researchers hope that physically skilled robots will be able to take semantic intelligence out of the purely virtual world of the Internet, and put it to good use on real-world physical tasks, perhaps even in our homes.
The paper has been accepted for publication at the Robotics: Science and Systems Conference (Daegu, Korea, July 10-14, 2023), and is currently available as a preprint.
ABOUT THE STUDY
CONFERENCE: Science and Systems Conference (Daegu, Korea, July 10-14, 2023)
STUDY: "Sampling-based Exploration for Reinforcement Learning of Dexterous Manipulation”
AUTHORS: Authors are all from Columbia Engineering: Gagan Khandate and Tristan Luca Saidi (Computer Science), Siqi Shang, Eric Chang, Johnson Adams, and Matei Ciocarlie (Mechanical Engineering). The tactile sensors were developed in collaboration with Ioannis Kymissis (Electrical Engineering).
FUNDING: This work was supported in part by the Office of Naval Research grant N00014-21-1-4010 and the National Science Foundation grant CMMI-2037101.
The authors declare no financial or other conflicts of interest.
The Future of Software-Controlled Cooking
Watch how Columbia mechanical engineers constructed a cheesecake using 3D food printing techniques.
Research produced by Dr. Jonathan Blutinger and his team at the Creative Machine Labs at Columbia University (directed by Prof. Hod Lipson) and Prof. Christen Cooper, Pace University Nutrition and Dietetics.
Addressing food printing challenges
Food printing technology has existed since Lipson’s lab first introduced it in 2005, but to date the technology has been limited to a small number of uncooked ingredients, resulting in what many perceive as less than appetizing dishes. Blutinger’s team broke away from this limitation by printing a dish comprising seven ingredients, cooked in situ using a laser. For the paper, the researchers designed a 3D-printing system that constructs cheesecake from edible food inks — including peanut butter, Nutella, and strawberry jam. The authors note that precision printing of multi-layered food items could produce more customizable foods, improve food safety, and enable users to control the nutrient content of meals more easily.
“Because 3D food printing is still a nascent technology, it needs an ecosystem of supporting industries such as food cartridge manufacturers, downloadable recipe files, and an environment in which to create and share these recipes. Its customizability makes it particularly practical for the plant-based meat market, where texture and flavor need to be carefully formulated to mimic real meats,” Blutinger said.
To demonstrate the potential of 3D food printing, the team tested various cheesecake designs, consisting of seven key ingredients: graham cracker, peanut butter, Nutella, banana puree, strawberry jam, cherry drizzle, and frosting. They found that the most successful design used a graham cracker as the foundational ingredient for each layer of the cake. Peanut butter and Nutella proved to be best used as supporting layers that formed “pools” to hold the softer ingredients: banana and jam. Multi-ingredient designs evolved into multi-tiered structures that followed similar principles to building architectures; more structural elements were needed to support softer substrates for a successful multi-ingredient layered print.
Is 3D food printing healthy?
“We have an enormous problem with the low-nutrient value of processed foods,” Cooper said. “3D food printing will still turn out processed foods, but perhaps the silver lining will be, for some people, better control and tailoring of nutrition--personalized nutrition. It may also be useful in making food more appealing to those with swallowing disorders by mimicking the shapes of real foods with the pureed texture foods that these patients--millions in the U.S. alone--require.”
The potential of 3D food-printing
Laser cooking and 3D food printing could allow chefs to localize flavors and textures on a millimeter scale to create new food experiences. People with dietary restrictions, parents of young children, nursing home dieticians, and athletes alike could find these personalized techniques very useful and convenient in planning meals. And, because the system uses high-energy targeted light for high-resolution tailored heating, cooking could become more cost-effective and more sustainable.
“The study also highlights that printed food dishes will likely require novel ingredient compositions and structures, due to the different way by which the food is ‘assembled,’ ” said Lipson. “Much work is still needed to collect data, model, and optimize these processes.”
Blutinger added, “And, with more emphasis on food safety following the COVID-19 pandemic, food prepared with less human handling could lower the risk of foodborne illness and disease transmission. This seems like a win-win concept for all of us.”
About the Study
JOURNAL: npj Science of Food
STUDY: “The Future of Software-Controlled Cooking”
AUTHORS: Jonathan David Blutinger (1, Christen Cupples Cooper (2), Shravan Karthik(1), Alissa Tsai (1), Noa Samarelli (1), Erika Storvick (1), Gabriel Seymour (1), Elise Liu (1), Yoran Meijers (1,3) and Hod Lipson (1)
- Department of Mechanical Engineering, Columbia Engineering
- Department of Nutrition and Dietetics, Pace University
- Department of Food Technology, Wageningen University, Netherlands.
FUNDING: The study was supported by NSF AI Institute for Dynamical Systems, grant 2112085, and by a grant from the Redefine Meat Ltd.
The authors declare no financial or other conflicts of interest.
Cyber Security Breefing
ABOUT THE STUDY
JOURNAL: Science Advances
TITLE: "Liquid solution centrifugation for safe, scalable, and efficient isotope separation"
AUTHORS: Joseph F. Wild, Heng Chen, Keyue Liang, Jiayu Liu, Stephen E. Cox, Alex N. Halliday, Yuan Yang
FUNDING: This work was supported by the U.S. Department of Energy, grant number DE-SC0022256, and the seed funding support from Columbia University’s Research Initiatives in Science and Engineering (RISE) competition, started in 2004 to trigger high-risk, high-reward, and innovative collaborations in the basic sciences, engineering, and medicine.
A provisional patent (U.S. 63/425,181) has been filed related to this work. The authors declare that they have no other competing interests.
ABOUT THE STUDY
JOURNAL: Nature Materials
AUTHORS: Claudia Cea(†1), Zifang Zhao(†1), Duncan J. Wisniewski(1), George D. Spyropoulos(1§), Anastasios Polyravas(1), Jennifer N. Gelinas(*2,3), Dion Khodagholy(*1)
- Department of Electrical Engineering, Columbia University, New York, NY 10027, USA
- Department of Neurology, Columbia University Medical Center, New York, NY 10032, USA.
- Institute for Genomic Medicine, Columbia University Medical Center, 630 W 168th St. New York, NY 10032, USA
§ Current address: Department Information Technology, Waves, UGhent Technology Campus, iGhent, Technologiepark 126, 9052 Zwijnaarde, Belgium
FUNDING: This work was supported by the National Institute of Health grants R01NS118091, R21 EY 32381-01, and RF1NS128669, National Science Foundation 1944415 and 2219891, and the Odysseus program from the Research Foundation – Flanders (FWO) G0F9421N.
The authors declare no financial or other conflicts of interest.