Robotics & AI Updates vol.99

Paradigm
Paradigm

--

TL;DR

Intelligent tutors powered by AI

Artificial muscles propel jumping robotic leg

Microscale robot folds into 3D shapes and crawls

Hexagonal electrohydraulic modules shape-shift into versatile robots

Researchers discover building blocks that could ‘revolutionize computing’

Robotics market

The global market for robots is expected to grow at a compound annual growth rate (CAGR) of around 26 percent to reach just under 210 billion U.S. dollars by 2025.

Size of the global market for industrial and non-industrial robots between 2018 and 2025 (in billion U.S. dollars):

The global market size for industrial and non-industrial robots between 2018 and 2025 (in billion U.S. dollars). Source: Statista

Latest News & Research

Real-Time multifaceted artificial intelligence vs In-Person instruction in teaching surgical technical skills: a randomized controlled trial

by Recai Yilmaz, Mohamad Bakhaidar, Ahmad Alsayegh, Nour Abou Hamdan, Ali M. Fazlollahi, Trisha Tee, Ian Langleben, Alexander Winkler-Schwartz, Denis Laroche, Carlo Santaguida, Rolando F. Del Maestro in Scientific Reports

Neurosurgery is perhaps one of the most demanding professions in healthcare. Surgeons spend long hours performing operations where expert performance means the difference between a good and bad patient outcome. While operative injuries are rare, when they occur, they can have serious, and lifelong consequences.

Researchers at the Neurosurgical Simulation and Artificial Intelligence Learning Centre at The Neuro (Montreal Neurological Institute-Hospital) of McGill University are striving to improve brain surgery training by designing real-time, intelligent tutors powered by AI. These systems are designed to mimic the role of human surgical instructors in brain surgical training. Intelligent tutors help the learner acquire excellent operative skills by continuously assessing hand movements during simulated brain procedures and providing personalized verbal feedback…

Educating the next generation of neurosurgeons is a long, expensive, and complex process. AI and simulation hold the potential to make the learning process easier while maintaining or enhancing the quality of graduating neurosurgeons’ skills.

Their most recent study was the first randomized controlled trial comparing AI intelligent tutor instruction with human expert human instruction during simulated surgery. They divided 97 medical trainees into three groups, either receiving real-time AI feedback, in-person expert instruction or no real-time feedback.

The trainees who received AI instruction performed significantly better than those who received expert instruction and no real-time instruction. The study found that expert instruction alone led to poorer surgical learning outcomes. By employing their extensive expertise and the new opportunities provided by AI, surgical educators can provide new possibilities for learners to reach their potential as excellent surgeons.

“This study suggests the future of instruction in the operating room may involve human educators utilizing the capacity of AI to further enhance learner surgical skills acquisition,” says Dr. Rolando Del Maestro, Director, Neurosurgical Simulation and Artificial Intelligence Learning Centre.

Electrohydraulic musculoskeletal robotic leg for agile, adaptive, yet energy-efficient locomotion

by Thomas J. K. Buchner, Toshihiko Fukushima, Amirhossein Kazemipour, Stephan-Daniel Gravert, Manon Prairie, Pascal Romanescu, Philip Arm, Yu Zhang, Xingrui Wang, Steven L. Zhang, Johannes Walter, Christoph Keplinger, Robert K. Katzschmann in Nature Communications

Inventors and researchers have been developing robots for almost 70 years. To date, all the machines they have built — whether for factories or elsewhere — have had one thing in common: they are powered by motors, a technology that is already 200 years old. Even walking robots feature arms and legs that are powered by motors, not by muscles as in humans and animals. This in part suggests why they lack the mobility and adaptability of living creatures.

A new muscle-powered robotic leg is not only more energy efficient than a conventional one, it can also perform high jumps and fast movements as well as detect and react to obstacles — all without the need for complex sensors. The new leg has been developed by researchers at ETH Zurich and the Max Planck Institute for Intelligent Systems (MPI-IS) in a research partnership called Max Planck ETH Center for Learning Systems, known as CLS. The CLS team was led by Robert Katzschmann at ETH Zurich and Christoph Keplinger at MPI-IS. Their doctoral students Thomas Buchner and Toshihiko Fukushima are the co-first authors of the team’s publication on an animal-inspired musculoskeletal robotic leg.

As in humans and animals, an extensor and a flexor muscle ensure that the robotic leg can move in both directions. These electro-hydraulic actuators, which the researchers call HASELs, are attached to the skeleton by tendons.

The actuators are oil-filled plastic bags, similar to those used to make ice cubes. About half of each bag is coated on either side with a black electrode made of a conductive material. Buchner explains that “as soon as we apply a voltage to the electrodes, they are attracted to each other due to static electricity. Similarly, when I rub a balloon against my head, my hair sticks to the balloon due to the same static electricity.” As one increases the voltage, the electrodes come closer and push the oil in the bag to one side, making the bag overall shorter.

Pairs of these actuators attached to a skeleton result in the same paired muscle movements as in living creatures: as one muscle shortens, its counterpart lengthens. The researchers use a computer code that communicates with high-voltage amplifiers to control which actuators contract, and which extend.

A robotic leg built with electrohydraulic artificial muscles enables versatile robots.

The researchers compared the energy efficiency of their robotic leg with that of a conventional robotic leg powered by an electric motor. Among other things, they analysed how much energy is unnecessarily converted into heat. “On the infrared image, it’s easy to see that the motorised leg consumes much more energy if, say, it has to hold a bent position,” Buchner says. The temperature in the electro-hydraulic leg, in contrast, remains the same. This is because the artificial muscle is electrostatic.

“It’s like the example with the balloon and the hair, where the hair stays stuck to the balloon for quite a long time,” Buchner adds. “Typically, electric motor driven robots need heat management which requires additional heat sinks or fans for diffusing the heat to the air. Our system doesn’t require them,” Fukushima says.

The robotic leg’s ability to jump is based on its ability to lift its own weight explosively. The researchers also showed that the robotic leg has a high degree of adaptability, which is particularly important for soft robotics. Only if the musculoskeletal system has sufficient elasticity can it adapt flexibly to the terrain in question.

“It’s no different with living creatures. If we can’t bend our knees, for example, walking on an uneven surface becomes much more difficult,” Katzschmann says. “Just think of taking a step down from the pavement onto the road.”

In contrast to electric motors requiring sensors to constantly tell what angle the robotic leg is at, the artificial muscle adapts to suitable position through the interaction with the environment. This is driven just by two input signals: one to bend the joint and one to extend it. Fukushima explains: “Adapting to the terrain is a key aspect. When a person lands after jumping into the air, they don’t have to think in advance about whether they should bend their knees at a 90-degree or a 70-degree angle.” The same principle applies to the robotic leg’s musculoskeletal system: upon landing, the leg joint adaptively moves into a suitable angle depending on whether the surface is hard or soft.

The research field of electrohydraulic actuators is still young, having emerged only around six years ago. “The field of robotics is making rapid progress with advanced controls and machine learning; in contrast, there has been much less progress with robotic hardware, which is equally important. This publication is a powerful reminder of how much potential for disruptive innovation comes from introducing new hardware concepts, like the use of artificial muscles,” Keplinger says. Katzschmann adds that electro-hydraulic actuators are unlikely to be used in heavy machinery on construction sites, but they do offer specific advantages over standard electric motors. This is particularly evident in applications such as grippers, where the movements have to be highly customised depending on whether the object being gripped is, for example, a ball, an egg or a tomato.

Katzschmann does have one reservation: “Compared to walking robots with electric motors, our system is still limited. The leg is currently attached to a rod, jumps in circles and can’t yet move freely.” Future work should overcome these limitations, opening the door to developing real walking robots with artificial muscles. He further elaborates: “If we combine the robotic leg in a quadruped robot or a humanoid robot with two legs, maybe one day, when it is battery-powered, we can deploy it as a rescue robot.”

Electronically configurable microscopic metasheet robots

by Qingkun Liu, Wei Wang, Himani Sinhmar, Itay Griniasty, Jason Z. Kim, Jacob T. Pelster, Paragkumar Chaudhari, Michael F. Reynolds, Michael C. Cao, David A. Muller, Alyssa B. Apsel, Nicholas L. Abbott, Hadas Kress-Gazit, Paul L. McEuen, Itai Cohen in Nature Materials

Cornell University researchers have created microscale robots less than 1 millimeter in size that are printed as a 2D hexagonal “metasheet” but, with a jolt of electricity, morph into preprogrammed 3D shapes and crawl.

The robot’s versatility is due to a novel design based on kirigami, a cousin of origami, in which slices in the material enable it to fold, expand and locomote.

The paper’s co-lead authors are postdoctoral researchers Qingkun Liu and Wei Wang. The project was led by Itai Cohen, professor of physics. His lab has previously produced microrobotic systems that can actuate their limbs, pump water via artificial cilia and walk autonomously.

Kirigami structure of metabots.

In a sense, the origins of the kirigami robot were inspired by “living organisms that can change their shape.” Liu said. “But when people make a robot, once it’s fabricated, it might be able to move some limbs but its overall shape is usually static. So we’ve made a metasheet robot. The ‘meta’ stands for metamaterial, meaning that they’re composed of a lot of building blocks that work together to give the material its mechanical behaviors.”

The robot is a hexagonal tiling composed of approximately 100 silicon dioxide panels that are connected through more than 200 actuating hinges, each about 10 nanometers thin. When electrochemically activated via external wires, the hinges form mountain and valley folds and act to splay open and rotate the panels, allowing the robot to change its coverage area and locally expand and contract by up to 40%. Depending which hinges are activated, the robot can adopt various shapes and potentially wrap itself around other objects, and then unfold itself back into a flat sheet.

Cohen’s team is already thinking of the next phase of metasheet technology. They anticipate combining their flexible mechanical structures with electronic controllers to create ultra-responsive “elastronic” materials with properties that would never be possible in nature. Applications could range from reconfigurable micromachines to miniaturized biomedical devices and materials that can respond to impact at nearly the speed of light, rather than the speed of sound.

“Because the electronics on each individual building block can harvest energy from light, you can design a material to respond in programmed ways to various stimuli. When prodded, such materials, instead of deforming, could ‘run’ away, or push back with greater force than they experienced,” Cohen said. “We think that these active metamaterials — these elastronic materials — could form the basis for a new type of intelligent matter governed by physical principles that transcend what is possible in the natural world.”

Hexagonal electrohydraulic modules for rapidly reconfigurable high-speed robots

by Zachary Yoder, Ellen H. Rumley, Ingemar Schmidt, Philipp Rothemund, Christoph Keplinger in Science Robotics

Scientists at the Max-Planck-Institute for Intelligent Systems (MPI-IS) have developed hexagon-shaped robotic components, called modules, that can be snapped together LEGO-style into high-speed robots that can be rearranged for different capabilities. The team of researchers from the Robotic Materials Department at MPI-IS, led by Christoph Keplinger, integrated artificial muscles into hexagonal exoskeletons that are embedded with magnets, allowing for quick mechanical and electrical connections.

Six lightweight rigid plates made from glass fiber serve as the exoskeleton of each HEXEL module. The inner joints of the hexagons are driven by hydraulically amplified self-healing electrostatic (HASEL) artificial muscles. Applying a high voltage to the module causes the muscle to activate, rotating the joints of the hexagon and changing its shape from long and narrow to wide and flat.

“Combining soft and rigid components in this way enables high strokes and high speeds. By connecting several modules, we can create new robot geometries and repurpose them for changing needs,” says Ellen Rumley, a visiting researcher from the University of Colorado Boulder. She and Zachary Yoder, who are both Ph.D. students working in the Robotic Materials Department, are co-first authors of the publication.

In a video, the team shows the many behaviors that can be created with HEXEL modules. A group of modules crawls through a narrow gap, while a single module actuates so fast that it can leap into the air. Multiple modules are connected into larger structures that produce different motions depending on how the modules are attached. For instance, the team combined several modules into a robot which rapidly rolls.

“In general, it makes a lot of sense to develop robots with reconfigurable capabilities. It’s a sustainable design option — instead of buying five different robots for five different purposes, we can build many different robots by using the same components. Robots made from reconfigurable modules could be rearranged on demand to provide more versatility than specialized systems, which could be beneficial in resource-limited environments’’, Yoder concludes.

Linear symmetric self-selecting 14-bit kinetic molecular memristors

by Deepak Sharma, Santi Prasad Rath, Bidyabhusan Kundu, Anil Korkmaz, Harivignesh S, Damien Thompson, Navakanta Bhat, Sreebrata Goswami, R. Stanley Williams, Sreetosh Goswami in Nature

A research team at University of Limerick has made a major discovery by designing molecules that could revolutionise computing.

The researchers at UL’s Bernal Institute have discovered new ways of probing, controlling and tailoring materials at the most fundamental molecular scale.

The results have been used in an international project involving experts worldwide to help create a brand-new type of hardware platform for artificial intelligence that achieves unprecedented improvements in computational speed and energy efficiency.

The UL team, led by Damien Thompson, Professor of Molecular Modelling at UL and director of SSPC, the Research Ireland Centre for Pharmaceuticals, in an international collaboration with scientists at the Indian Institute of Science (IISc) and Texas A&M University, believe that this new discovery will lead to innovative solutions to societal grand challenges in health, energy and the environment.

Professor Thompson explained: “The design draws inspiration from the human brain, using the natural wiggling and jiggling of atoms to process and store information. As the molecules pivot and bounce around their crystal lattice, they create a multitude of individual memory states.

“We can trace out the path of the molecules inside the device and map each snapshot to a unique electrical state. That creates a kind of tour diary of the molecule that can be written and read just like in a conventional silicon-based computer, but here with massively improved energy and space economy because each entry is smaller than an atom.

“This outside the box solution could have huge benefits for all computing applications, from energy hungry data centres to memory intensive digital maps and online gaming.”

To-date, neuromorphic platforms — an approach to computing inspired by the human brain — have worked only for low-accuracy operations, such as inferencing in artificial neural networks. This is because core computing tasks including signal processing, neural network training, and natural language processing require much higher computing resolution than what existing neuromorphic circuits could offer.

For this reason then, achieving high resolution has been the most daunting challenge in neuromorphic computing. The team’s reconceptualization of the underlying computing architecture achieves the required high resolution, performing resource-intensive workloads with unprecedented energy efficiency of 4.1 tera-operations per second per watt (TOPS/W).

The team’s breakthrough extends neuromorphic computing beyond niche applications in a move that can potentially unleash the long-heralded transformative benefits of artificial intelligence and augment the core of digital electronics from the cloud to the edge.

Project lead at IISc Professor Sreetosh Goswami said: “By precisely controlling the vast array of available molecular kinetic states, we created the most accurate, 14-bit, fully functional neuromorphic accelerator integrated into a circuit board that can handle signal processing, AI and machine learning workloads such as artificial neural networks, auto-encoders, and generative adversarial networks.

“Most significantly, leveraging the high precision of the accelerators, we can train neural networks on the edge, addressing one of the most pressing challenges in AI hardware.”

Further enhancements are coming, as the team works to expand the range of materials and processes used to create the platforms and increase the processing power even further.

Professor Thompson explained: “The ultimate aim is to replace what we now think of as computers with high-performance ‘everyware’ based on energy efficient and eco-friendly materials providing distributed ubiquitous information processing throughout the environment integrated in everyday items from clothing to food packaging to building materials.”

Subscribe to Paradigm!

Medium, Twitter, Telegram, Telegram Chat, LinkedIn, and Reddit.

Main sources

Research articles

Science Robotics

Science Daily

IEEE Spectrum

Main sources

Research articles

Science Robotics

Science Daily

IEEE Spectrum

--

--