+ What is an RSS feed?
Current feed - IEEE Spectrum
Video Friday is your weekly selection of awesome robotics videos, collected by your friends at IEEE Spectrum robotics. We also post a weekly calendar of upcoming robotics events for the next few months. Please send us your events for inclusion.
Enjoy today’s videos!
Behind the scenes at DARPA Triage Challenge Workshop 2 at the Guardian Centers in Perry, GA.
[ DARPA ]
Watch our coworker in action as he performs high-precision stretch routines enabled by 31 degrees of freedom. Designed for dynamic adaptability, this is where robotics meets real-world readiness.
[ LimX Dynamics ]
Thanks, Jinyan!
Featuring a lightweight design and continuous operation capabilities under extreme conditions, LYNX M20 sets a new benchmark for intelligent robotic platforms working in complex scenarios.
[ DEEP Robotics ]
The sound in this video is either excellent or terrible, I’m not quite sure which.
[ TU Berlin ]
Humanoid loco-manipulation holds transformative potential for daily service and industrial tasks, yet achieving precise, robust whole-body control with 3D end-effector force interaction remains a major challenge. Prior approaches are often limited to lightweight tasks or quadrupedal/wheeled platforms. To overcome these limitations, we propose FALCON, a dual-agent reinforcement-learning-based framework for robust force-adaptive humanoid loco-manipulation.
[ FALCON ]
An MRSD Team at the CMU Robotics Institute is developing a robotic platform to map environments through perceptual degradation, identify points of interest, and relay that information back to first responders. The goal is to reduce info blindness and increase safety.
[ Carnegie Mellon University ]
We introduce an eldercare robot (E-BAR) capable of lifting a human body, assisting with postural changes/ambulation, and catching a user during a fall, all without the use of any wearable device or harness. With a minimum width of 38 cm, the robot’s small footprint allows it to navigate the typical home environment. We demonstrate E-BAR’s utility in multiple typical home scenarios elderly persons experience, including getting into/out of a bathtub, bending to reach for objects, sit-to-stand transitions, and ambulation.
[ MIT ]
Sanctuary AI had the pleasure of accompanying Microsoft to Hannover Messe, where we demonstrated how our technology is shaping the future of work with autonomous labor powered by Physical AI and general purpose robots.
[ Sanctuary AI ]
Watch how drywall finishing machines incorporate collaborative robots and why Canvas chose Universal Robots platform.
[ Canvas ] via [ Universal Robots ]
We’ve officially put a stake in the ground in Dallas Fort Worth. Torc’s new operations hub is open for business—and it’s more than just a dot on the map. It’s a strategic launchpad as we expand our autonomous freight network across the southern U.S.
[ Torc ]
This Stanford Robotics Center talk is by Jonathan Hurst at Agility Robotics, on “Humanoid Robots: From The Warehouse to Your House.”
How close are we to having safe, reliable, useful in-home humanoids? If you believe recent press, it’s just around the corner. Unquestionably, advances in Al and robotics are driving innovation and activity in the sector; it truly is an exciting time to be building robots! But what does it really take to execute on the vision of useful, human centric, multi purpose robots? Robots that can operate in human spaces, predictably and safely? We think it starts with humanoids in warehouses, an unsexy but necessary beachhead market to our future with robots as part of everyday life. I’ll talk about why a humanoid is more than a sensible form factor, it’s inevitable; and I will speak to the excitement around a ChatGPT moment for robotics, and what it will take to leverage Al advances and innovation in robotics into useful, safe humanoids.
[ Stanford ]
For more than three years, an IEEE Standards Association working group has been refining a draft standard for procuring artificial intelligence and automated decision systems, IEEE 3119-2025. It is intended to help procurement teams identify and manage risks in high-risk domains. Such systems are used by government entities involved in education, health, employment, and many other public sector areas. Last year the working group partnered with a European Union agency to evaluate the draft standard’s components and to gather information about users’ needs and their views on the standard’s value.
At the time, the standard included five processes to help users develop their solicitations and to identify, mitigate, and monitor harms commonly associated with high-risk AI systems.
Those processes were problem definition, vendor evaluation, solution evaluation, contract negotiation, and contract monitoring.
The EU agency’s feedback led the working group to reconsider the processes and the sequence of several activities. The final draft now includes an additional process: solicitation preparation, which comes right after the problem definition process. The working group believes the added process addresses the challenges organizations experience with preparing AI-specific solicitations, such as the need to add transparent and robust data requirements and to incorporate questions regarding the maturity of vendor AI governance.
The EU agency also emphasized that it’s essential to include solicitation preparation, which gives procurement teams additional opportunities to adapt their solicitations with technical requirements and questions regarding responsible AI system choices. Leaving space for adjustments is especially relevant when acquisitions of AI are occurring within emerging and changing regulatory environments.
Gisele Waters
Currently there are several internationally accepted standards for AI management, AI ethics, and general software acquisition. Those from the IEEE Standards Association and the International Organization for Standardization target AI design, use, and life-cycle management.
Until now, there has been no internationally accepted, consensus-based standard that focuses on the procurement of AI tools and offers operational guidance for responsibly purchasing high-risk AI systems that serve the public interest.
The IEEE 3119 standard addresses that gap. Unlike the AI management standard ISO 42001 and other certifications related to generic AI oversight and risk governance, IEEE’s new standard offers a risk-based, operational approach to help government agencies adapt traditional procurement practices.
Governments have an important role to play in the responsible deployment of AI. However, market dynamics and unequal AI expertise between industry and government can be barriers that discourage success.
One of the standard’s core goals is to better inform procurement leaders about what they are buying before they make high-risk AI purchases. IEEE 3119 defines high-risk AI systems as those that make or are a substantial factor in making consequential decisions that could have significant impacts on people, groups, or society. The definition is similar to the one used in Colorado’s 2034 AI Act, the first U.S. state-level law comprehensively addressing high-risk systems.
The standard’s processes, however, do complement ISO 42001 in many ways. The relationship between both is illustrated below.
| IEEE 3119 clause | ISO/IEC 42001:2023 clause |
|---|---|
| 6 Problem definition |
4.1 Understanding the organization and its context
4.2 Understanding the needs and expectations of interested parties 6.1.4 AI system impact assessment |
| 7 Solicitation preparation process |
4.3 Determining the scope of the AI management system
4.4 AI management system and its processes |
| 8 Vendor evaluation process |
5. Leadership (commitment, policies, roles, etc.)
6.1.2 AI risk assessment 7.1 Resources 7.2 Competence 7.3 Awareness |
| 9 Solution evaluation process |
4.3 Determining the scope of the AI management system
4.4 AI management system and its processes 6.1 Actions to address risks and opportunities 7.4 Communication 7.5 Documented information |
| 10 Contract negotiation process |
6.1.3 AI risk treatment
9.1 Monitoring, measurement, analysis, and evaluation 10.2 Nonconformity and corrective action |
| 11 Contract monitoring process |
4.4 AI management system
6.3 Planning of changes 7.1 Resources 7.2 Competence 7.5.2 Creating and updating documented information 7.5.3 Controlling documented information 8.1 Operational planning and control 8.2 AI risk assessment 8.3 AI risk treatment 8.4 AI system impact assessment 9.1 Monitoring, measurement, analysis, and evaluation 9.2 Internal audit 9.3 Management review 10.2 Nonconformity and corrective action |
Source: IEEE 3119-2025 Working Group
International standards, often characterized as soft law, are used to shape AI development and encourage international cooperation regarding its governance.
Hard laws for AI, or legally binding rules and obligations, are a work in progress around the world. In the United States, a patchwork of state legislation governs different aspects of AI, and the approach to national AI regulation is fragmented, with different federal agencies implementing their own guidelines.
Europe has led by passing the European Union’s AI Act, which began governing AI systems based on their risk levels when it went into effect last year.
But the world lacks regulatory hard laws with an international scope.
The IEEE 3119-2025 standard does align with existing hard laws. Due to its focus on procurement, the standard supports the high-risk provisions outlined in the EU AI Act’s Chapter III and Colorado’s AI Act. The standard also conforms to the proposed Texas HB 1709 legislation, which would mandate reporting on the use of AI systems by certain business entities and state agencies.
Because most AI systems used in the public sector are procured rather than built in-house, IEEE 3119 applies to commercial AI products and services that don’t require substantial modifications or customizations.
The standard is intended for:
The IEEE Standards Association has partnered with the AI Procurement Lab to offer the IEEE Responsible AI Procurement Training program. The course covers how to apply the standard’s core processes and adapt current practices for the procurement of high-risk AI.
The standard includes over 26 tools and rubrics across the six processes, and the training program explains how to use many of these tools. For example, the training includes instructions on how to conduct a risk-appetite analysis, apply the vendor evaluation scoring guide to analyze AI vendor claims, and create an AI procurement “risk register” tied to identified use-case risks and their potential mitigations. The training session is now available for purchase.
It’s still early days for AI integration. Decision makers don’t yet have much experience in purchasing AI for high-risk domains and in mitigating those risks. The IEEE 3119-2025 standard aims to support agencies build and strengthen their AI risk mitigation muscles.
There’s a mathematical concept called the kissing number. Somewhat disappointingly, it’s got nothing to do with actual kissing. It enumerates how many spheres can touch (or “kiss”) a single sphere of equal size without crossing it. In one dimension, the kissing number is 2. In two dimensions, it’s 6 (think The New York Times’ spelling bee puzzle configuration). As the number of dimensions grows, the answer becomes less obvious: For most dimensionalities over 4, only upper and lower bounds on the kissing number are known. Now, an AI agent developed by Google DeepMind called AlphaEvolve has made its contribution to the problem, increasing the lower bound on the kissing number in 11 dimensions from 592 to 593.
This may seem like an incremental improvement on the problem, especially given that the upper bound on the kissing number in 11 dimensions is 868, so the unknown range is still quite large. But it represents a novel mathematical discovery by an AI agent, and challenges the idea that large language models are not capable of original scientific contributions.
And this is just one example of what AlphaEvolve has accomplished. “We applied AlphaEvolve across a range of open problems in research mathematics, and we deliberately picked problems from different parts of math: analysis, combinatorics, geometry,” says Matej Balog, a research scientist at DeepMind that worked on the project. They found that for 75 percent of the problems, the AI model replicated the already known optimal solution. In 20 percent of cases, it found a new optimum that surpassed any known solution. “Every single such case is a new discovery,” Balog says. (In the other 5 percent of cases, the AI converged on a solution that was worse than the known optimal one.)
The model also developed a new algorithm for matrix multiplication—the operation that underlies much of machine learning. A previous version of DeepMind’s AI model, called AlphaTensor, had already beat the previous best known algorithm, discovered in 1969, for multiplying 4 by 4 matrices. AlphaEvolve found a more general version of that improved algorithm.
DeepMind’s AlphaEvolve made improvements to several practical problems at Google. Google DeepMind
In addition to abstract math, the team also applied their model to practical problems Google as a company faces every day. The AI was also used to optimize data-center orchestration to gain 1 percent improvement, to optimize the design of the next Google tensor processing unit, and to discover an improvement to a kernel used in Gemini training leading to a 1 percent reduction in training time.
“It’s very surprising that you can do so many different things with a single system,” says Alexander Novikov, a senior research scientist at DeepMind who also worked on AlphaEvolve.
AlphaEvolve is able to be so general because it can be applied to almost any problem that can be expressed as code, and which can be checked by another piece of code. The user supplies an initial stab at the problem—a program that solves the problem at hand, however suboptimally—and a verifier program that checks how well a piece of code meets the required criteria.
Then, a large language model, in this case Gemini, comes up with other candidate programs to solve the same problem, and each one is tested by the verifier. From there, AlphaEvolve uses a genetic algorithm such that the “fittest” of the proposed solutions survive and evolve to the next generation. This process repeats until the solutions stop improving.
AlphaEvolve uses an ensemble of Gemini large language models (LLMs) in conjunction with an evaluation code, all orchestrated by a genetic algorithm to optimize a piece of code. Google DeepMind
“Large language models came around, and we started asking ourselves, is it the case that they are only going to add what’s in the training data, or can we actually use them to discover something completely new, new algorithms or new knowledge?” Balog says. This research, Balog claims, shows that “if you use the large language models in the right way, then you can, in a very precise sense, get something that’s provably new and provably correct in the form of an algorithm.”
AlphaEvolve comes from a long lineage of DeepMind’s models, going back to AlphaZero, which stunned the world by learning to play chess, Go, and other games better than any human player without using any human knowledge—just by playing the game and using reinforcement learning to master it. Another math-solving AI based on reinforcement learning, AlphaProof, performed at the silver-medalist level on the 2024 International Math Olympiad.
For AlphaEvolve, however, the team broke from the reinforcement learning tradition in favor of the genetic algorithm. “The system is much simpler,” Balog says. “And that actually has consequences, that it’s much easier to set up on a wide range of problems.”
The team behind AlphaEvolve hopes to evolve their system in two ways.
First, they want to apply it to a broader range of problems, including those in the natural sciences. To pursue this goal, they are planning to open up an early access program for interested academics to use AlphaEvolve in their research. It may be harder to adapt the system to the natural sciences, as verification of proposed solutions may be less straightforward. But, Balog says, “We know that in the natural sciences, there are plenty of simulators for different types of problems, and then those can be used within AlphaEvolve as well. And we are, in the future, very much interested in broadening the scope in this direction.”
Second, they want to improve the system itself, perhaps by coupling it with another DeepMind project: the AI coscientist. This AI also uses an LLM and a genetic algorithm, but it focuses on hypothesis generation in natural language. “They develop these higher-level ideas and hypotheses,” Balog says. “Incorporating this component into AlphaEvolve-like systems, I believe, will allow us to go to higher levels of abstraction.”
These prospects are exciting, but for some they may also sound menacing—for example, AlphaEvolve’s optimization of Gemini training may be seen as the beginning of recursively self-improving AI, which some worry would lead to a runaway intelligence explosion referred to as the singularity. The DeepMind team maintains that that is not their goal, of course. “We are excited to contribute to advancing AI that benefits humanity,” Novikov says.
And it’s not a new one: From early telephones to modern cellphones, everyday liquids have frequently conflicted with devices that must stay dry. Consumers often take the blame when leaks and spills inevitably occur.
Rachel Plotnick, an associate professor of cinema and media studies at Indiana University Bloomington, studies the relationship between technology and society. Last year, she spoke to IEEE Spectrum about her research on how people interact with buttons and tactile controls. In her new book, License to Spill: Where Dry Devices Meet Liquid Lives (The MIT Press, 2025), Plotnick explores the dynamic between everyday wetness and media devices through historical and contemporary examples, including cameras, vinyl records, and laptops. This adapted excerpt looks back at analog telephones of the 1910s through 1930s, the common practices that interrupted service, and the “trouble men” who were sent to repair phones and reform messy users.
Mothers never liked to blame their babies for failed telephone service. After all, what harm could a bit of saliva do? Yet in the early decades of the 20th century, reports of liquid-gone-wrong with telephones reached the pages of popular women’s magazines and big-city newspapers as evidence of basic troubles that could befall consistent service. Teething babies were particularly called out. The Boston Daily Globe in 1908 recounted, for instance, how a mother only learned her lesson about her baby’s cord chewing when the baby received a shock—or “got stung”—and the phone service went out. These youthful oral fixations rarely caused harm to the chewer, but were “injurious” to the telephone cord.
License to Spill is Rachel Plotnick’s second book. Her first, Power Button: A History of Pleasure, Panic, and the Politics of Pushing (The MIT Press, 2018), explores the history and politics of push buttons. The MIT Press
As more Americans encountered telephones in the decades before World War II, those devices played a significant role in daily life. That daily life was filled with wet conditions, not only teething babies but also “toy poodles, the ever-present spittoon, overshoes…and even people talking while in the bathtub,” according to a 1920 article from the journal Telephony. Painters washed ceilings, which dripped; telephones sat near windows during storms; phone cords came in contact with moist radiators. A telephone chief operator who handled service complaints recounted that “a frequent combination in interior decoration is the canary bird and desk telephone occupying the same table. The canary bird includes the telephone in his morning bath,” thus leading to out-of-order service calls.
Within the telephone industry, consensus built around liquids as a hazard. As a 1913 article on telephone service stated ominously, “Water is one of the worst enemies.” At the time, cords were typically made from silk tinsel and could easily corrode from wetness, while any protective treatment tended to make them too brittle. But it wasn’t an elemental force acting alone or fragile materials that bothered phone workers. Rather, the blame fell on the abusing consumer—the “energetic housewife” who damaged wiring by scrubbing her telephone with water or cleaning fluid, and men in offices who dangerously propped their wet umbrellas against the wire. Wetness lurked everywhere in people’s spaces and habits; phone companies argued that one could hardly expect proper service under such circumstances—especially if users didn’t learn to accommodate the phone’s need for dryness.
In telephony’s infancy, though, users didn’t always make the connection between liquidity and breakdown and might not even notice the wetness, at least in a phone company’s estimation.
This differing appraisal of liquids caused problems when telephone customers expected service that would not falter and directed outrage at their provider when outages did occur. Consumers even sometimes admitted to swearing at the telephone receiver and haranguing operators. Telephone company employees, meanwhile, faced intense scrutiny and pressure to tend to telephone infrastructures. “Trouble” took two forms, then, in dealing with customers’ frustration over outages and in dealing with the damage from the wetness itself.
Telephone breakdowns required determinations about the outage’s source. “Trouble men” and “trouble departments” hunted down the probable cause of the damage, which meant sussing out babies, sponges, damp locations, spills, and open windows. If customers wanted to lay blame at workers’ feet in these moments, then repairers labeled customers as abusers of the phone cord. One author attributed at least 50 percent of telephone trouble to cases where “someone has been careless or neglectful.” Trouble men employed medical metaphors to describe their work, as in “he is a physician, and he makes the ills that the telephone is heir to his life study.”
Serge Bloch
Stories about this investigative work abounded. They typically emphasized the user’s ignorance and established the trouble man as the voice of reason, as in the case of an ill-placed wet umbrella leaned up against the telephone wiring. It didn’t seem to occur to the telephone worker that the umbrella user simply didn’t notice the umbrella’s positioning. Phone companies thus tried to make wetness a collective problem—for instance, by taking out newspaper announcements that commented on how many households lost power in a particular storm due to improper umbrella habits.
Even if a consumer knew the cord had gotten wet, they didn’t necessarily blame it as the cause of the outage. The repairer often used this as an opportunity to properly socialize the user about wetness and inappropriate telephone treatment. These conversations didn’t always go well: A 1918 article in Popular Science Monthly described an explosive argument between an infuriated woman and a phone company employee over a baby’s cord habits. The permissive mother and teething child had become emblematic of misuse, a photograph of them appearing in Bell Telephone News in 1917 as evidence of common trouble that a telephone (and its repairer) might encounter. However, no one blamed the baby; telephone workers unfailingly held mothers responsible as “bad” users.
Teething babies and the mothers that let them play with phone cords were often blamed for telephone troubles. The Telephone Review/License to Spill
Repair work often involved special tools meant to identify the source of the outage. Not unlike a doctor relying upon an X-ray to visualize and interpret a patient’s body, the trouble man relied on an apparatus known as the Telefault to evaluate breakages. The repairer attached an exploring coil to a telephone receiver and then generated an intermittent current that, when sent out over the malfunctioning wire, allowed him to hear the source of the fault. This wasn’t always an easy process, but linemen nevertheless recommended the Telefault through testimonials and articles. The machine and trouble man together functioned as co-testers of wetness, making everyday life’s liquidity diagnosable and interpretable.
Armed with such a tool, repairers glorified their own expertise. One wire chief was celebrated as the “original ‘find-out artist’” who could determine a telephone’s underlying troubles even in tricky cases. Telephone company employees leveraged themselves as experts who could attribute wetness’s causes to—in their estimation—uneducated (and even dimwitted) customers, who were often female. Women were often the earliest and most engaged phone users, adopting the device as a key mechanism for social relations, and so they became an easy target.
Phone repairers were constructing everyday life as a problem for uninterrupted service; untamed mouths, clumsy hands, and wet umbrellas all stood at odds with connectivity.
Though the phone industry and repairers were often framed as heroes, troubleshooting took its toll on overextended phone workers, and companies suffered a financial burden from repairs. One estimate by the American Telephone and Telegraph Company found that each time a company “clear[ed] wet cord trouble,” it cost a dollar. Phone companies portrayed the telephone as a fragile device that could be easily damaged by everyday life, aiming to make the subscriber a proactively “dry” and compliant user.
Telephone workers also quantified the cost of moisture incidents that impaired good service. According to an investigation conducted by an Easton, Pa., central office employee, a baby chewing on a cord could lead to 1 hour and 45 minutes of lost service, while a spilled pitcher of water would cause a whopping 8-hour outage. Other quantifications related to spilled whisky, mustard, wet hands, and mops. In a cheeky summary of this work, a reporter reminded readers that the investigator did not recommend “doing away with babies, sponges and wet bouquets” but rather offered his statistics “as an educational hint to keep the telephone cord away from dampness.”
Everyday sources of wetness, including mops and mustard, could cause hours of phone interruption. Telephony/License to Spill
A blossoming accessory market also emerged, which focused on moving phones away from sources of moisture. The telephone bracket, for example, clamped onto a desk and, like a “third arm” or “human arm,” would “hold [the phone] out of your way when not in use; brings it where you want it at a touch.” The Equipoise Telephone Arm was used in offices and on ships as a sort of worker’s appendage. One company’s advertisements promised that the Equipoise could prevent liquid messes—like overturned inkstands—and could stop cords from getting tangled or impeding one’s work.
Although telephone companies put significant effort into reforming their subscribers, the increasing pervasiveness of telephony began to conflict with these abstinent aims. Thus, a new technological solution emerged that put the burden on moisture-proofing the wire. The Stromberg-Carlson Telephone Manufacturing Co. of Rochester, N.Y., began producing copper wire that featured an insulating enamel, two layers of silk, the company’s moisture-proof compound, and a layer of cotton. Called Duratex, the cord withstood a test in which the manufacturer submerged it in water for 48 hours. In its advertising, Stromberg-Carlson warned that many traditional cords—even if they seemed to dry out after wetting—had sustained interior damage so “gradual that it is seldom noticed until the subscriber complains of service.”
Serge Bloch
Western Electric, another manufacturer of liquid-friendly cords, claimed its moisture-proof and “hard-knock proof” cord could handle “rough” conditions and wore its coating like the Charles Dickens character Tony Weller in The Pickwick Papers, with his many layers of clothing. The product’s hardiness would allow the desk telephone to “withstand any climate,” even one hostile to communication technology.
Telephone companies that deployed these cords saw significant cost benefits. A report from Bell Telephone noted that in 1919, when it installed 1,800,000 of these protected cords, it began saving US $90,000 per year (about $1.6 million in today’s dollars). By 1926, that same report concluded, the company had saved $400,000. But something else significant had shifted in this transition that involved far more than developing a moisture-proof solution. The cultural balance tilted from encouraging consumers to behave properly to insulating these media technologies from their everyday circumstances.
This subtle change meant that the burden to adapt fell to the device rather than the user. As telephone wires began to “penetrate everywhere,” they were imagined as fostering constant and unimpeded connectivity that not even saliva or a spilled drink could interrupt. The move to cord protection was not accompanied by a great deal of fanfare, however. As part of telephone infrastructure, cords faded into the background of conversations.
Excerpted from License to Spill by Rachel Plotnick. Reprinted with permission from The MIT Press. Copyright 2025.
By 2030, there will be a global shortage of 85 million workers, many of them in technical fields, according to the World Economic Forum. Many industries that need to employ technical workers will be impacted by the shortage, which is projected to cost them up to US $8.5 trillion in unrealized revenue.
Many technical roles now require university degrees. However, as companies consider how to overcome the worker shortage, some are reevaluating their higher education requirements for certain roles requiring specialized skills.
Those jobs might include technician, electrician, and programmer, along with other positions that compose the skilled technical workforce, as described by SRI International’s Center for Innovation Strategy and Policy.
Positions that don’t require higher education widen the pool of candidates.
Even if they eliminate the need for a degree, organizations will still need to rely on some kind of credential to ensure that job candidates have the skills necessary to do the job. One option is the skills-based microcredential.
Microcredentials are issued when learners prove mastery of a specific skill. Unlike traditional university degrees and course certificates, microcredential programs are not based on successfully completing a full learning program. Instead, a student might earn multiple microcredentials in a single program based on demonstrated skills. A qualified instructor using an assessment instrument determines if a learner has acquired the skill and earned the credential.
The IEEE microcredentials program offers standardized credentials in collaboration with training organizations and universities seeking to provide skills-based credentials outside formal degree programs. IEEE, as the world’s largest technical professional organization, has decades of experience offering industry-relevant credentials and expertise in global standardization.
IEEE microcredentials are industry-driven professional credentials that focus on needed skills. The program allows technical learning providers to supply credentials that bear the IEEE logo. When a hiring organization sees the logo on a microcredential, it confirms to employers that the instruction has been independently vetted and the institution is qualified to issue the credential. Credentials issued through the IEEE program include certificates and digital badges.
Training providers that want to offer standardized microcredentials can apply to the program to become approved. A committee reviews the applications to ensure that providers are credible, offer training within IEEE’s fields of interest, have qualified instructors, and have well-defined assessments.
The IEEE program offers standardized credentials in collaboration with training organizations and universities seeking to provide skills-based credentials outside formal degree programs.
Once a provider is approved, IEEE will work with it to benchmark the credentialing needs for each course, including the skills to be recognized, designing microcredentials, and creating a credential-issuing process. Upon the learner’s successful completion of the program, IEEE will issue the microcredentials on behalf of the training provider.
Microcredentials are stackable; students can earn them from different programs and institutions to demonstrate their growing skill set. The microcredentials can be listed on résumés and CVs and shared on LinkedIn and other professional networking websites.
All IEEE microcredentials that a learner earns are stored within a secure digital wallet for easy reference. The wallet also provides information about the program that issued each credential.
Researchers who are developing electrolyzers for hydrogen production are increasingly turning to a membrane platform originally used in fuel cells to scale up their technology. Their strategy: use anion-exchange membranes, which could be more cost-effective and combine the best features of conventional proton-exchange membranes and alkaline approaches.
Anion-exchange membrane (AEM) technology enables the selective transport of negatively charged ions between cathode and anode. In a hydrogen fuel cell, the membrane helps facilitate the chemical reactions needed to generate electricity. In hydrogen electrolysis, the membrane helps split water by separating hydrogen from oxygen.
So far, AEM has only been deployed at a small scale. But several renewable hydrogen companies are poised to change that. On 7 May, Ithaca, N.Y.–based Ecolectro announced a partnership with Framingham, Mass.–based Re:Build Manufacturing to deploy advanced AEM electrolyzers in the United States. And in March the French tire company Michelin and several French research institutions launched a multiyear collaboration to develop more durable versions of these membranes as part of Michelin’s expansion into renewable markets.
These companies, and several others globally, are betting on AEM technology to fulfill the long-sought promise of “green” hydrogen produced with renewable energy. “This has long been considered the potential savior to a lot of issues with other types of electrolysis that we’ve been trying to scale,” says Lindsey Motlow, a physicist and research director at Darcy Partners, a market intelligence firm in Houston.
Scaling up green hydrogen comes with challenges that have rendered it less competitive than other hydrogen production methods. The field relies on electrolyzers, which use electricity to split water molecules to release hydrogen. Most employ either a proton-exchange membrane (PEM), which uses precious metal catalysts and polymer membranes to split the molecules, or alkaline electrolysis, which works with an electrolyte solution.
PEM can quickly ramp up and down in response to variable energy sources like wind and solar power, but it requires iridium, which is in limited supply. Alkaline electrolysis is less capital intensive and more established at larger scales, but it lacks efficiency and its harsh, alkaline solution complicates system design.
That has led groups to turn to AEM, which substitutes nickel and steel for PEM’s costly metals. And while it does use an alkaline solution, AEM has better efficiencies than alkaline electrolysis, at least at the lab scale, Motlow says.
Saerbeck, Germany–based Enapter and Austin, Texas–based Agastya offer commercial megawatt-scale AEM electrolyzers used in industry for chemical reactions and heating. In China, Shandong-based Hygreen Energy in September 2024 launched a kilowatt-scale AEM electrolyzer for plug-and-play use in industrial parks, community buildings and transportation. However, these demonstrations remain limited in scale and maturity. AEM technology has not yet been proven at commercial scale for continuous industrial hydrogen supply.
Ecolectro’s AEM electrolyzer stack uses a PFAS-free, iridium-free membrane platform.Ecolectro
The partnership between Ecolectro and Re:Build aims to reduce the high costs that have hindered the scale-up of green hydrogen for industrial use. In addition to sourcing cheaper materials for the electrolyzer components, Ecolectro is outsourcing the manufacturing to Re:Build’s plants in New York and Pennsylvania. For the membranes, Ecolectro will use a proprietary blend of chemicals with a nickel catalyst for better durability.
Ecolectro is taking it one step at a time, says cofounder and CEO Gabriel Rodríguez-Calero. The company’s first commercial-scale units, to be developed this year at Re:Build’s design plant in Rochester, N.Y., will be 250 to 500 kilowatts. Rodríguez-Calero says his team plans to reach megawatt scale in 2026.
To deploy beyond lab scale, powering AEM with renewables faces significant engineering hurdles. The high efficiencies at the lab scale assume a steady flow of electricity powered by fossil fuels, but the ability to quickly respond to fluctuations in renewable energy hasn’t been tested widely. Membrane durability is another challenge, because materials must withstand AEM’s harsh, alkaline conditions. Fluorinated polymer membranes are an efficient option, but they pollute water and introduce forever chemicals.
To solve the membrane issue, Michelin in Clermont-Ferrand, France, and its research partners launched a collaboration they call Alcal’Hylab. Researchers will develop a new, more durable membrane using a mix of chemicals alongside a cost-effective metal catalyst—a similar model to Ecolectro’s. Alcal’Hylab’s goal is to deploy this membrane in a 25-kW AEM electrolyzer stack by 2027.
“It’s difficult to find a structure of a polymer that is really compatible with these operating conditions for a long time,” says Jacques Maddaluno, director of chemistry at the French National Centre for Scientific Research, which will host the collaborative lab. “You get very good results at time zero, but it degrades very, very quickly.”
Despite the many research groups working on the problem, skepticism around green hydrogen remains. The scientific and economic hurdles to developing it at an industrial scale do not lend themselves to a worthwhile investment, even for a company like Michelin, says Joseph Romm, physicist at the University of Pennsylvania and author of The Hype About Hydrogen: False Promises and Real Solutions in the Race to Save the Climate (Island Press, 2025). “The fact that they are making deals with research organizations tells you how far they have to go,” he says.
True, green hydrogen has yet to live up to its hype, says Rodríguez-Calero of Ecolectro. “I think the pace of adoption of some of this new hydrogen market has been slower than what a lot of people hoped,” he says. He sees Ecolectro as a meaningful step toward competing with fossil-fuel-derived hydrogen for industrial users that need to produce it on site.
But to go beyond these kinds of point-to-point replacements, green hydrogen still struggles to compete with renewable electricity. The industry also lacks the infrastructure to transport hydrogen long distances. Says Romm: “The biggest problem for AEM is that hydrogen doesn’t just have one problem.”
The main assumption about humanoid robotics that the industry is making right now is that the most realistic near-term pathway to actually making money is in either warehouses or factories. It’s easy to see where this assumption comes from: Repetitive tasks requiring strength or flexibility in well-structured environments is one place where it really seems like robots could thrive, and if you need to make billions of dollars (because somehow that’s how much your company is valued at), it doesn’t appear as though there are a lot of other good options.
Cartwheel Robotics is trying to do something different with humanoids. Cartwheel is more interested in building robots that people can connect with, with the eventual goal of general-purpose home companionship. Founder Scott LaValley describes Cartwheel’s robot as “a small, friendly humanoid robot designed to bring joy, warmth, and a bit of everyday magic into the spaces we live in. It’s expressive, emotionally intelligent, and full of personality—not just a piece of technology but a presence you can feel.”
This rendering shows the design and scale of Cartwheel’s humanoid prototype.Cartwheel
Historically, making a commercially viable social robot is a huge challenge. A little less than a decade ago, a series of social home robots (backed by a substantial amount of investment) tried very, very hard to justify themselves to consumers and did not succeed. Whether the fundamental problems with the concept of social home robots (namely, cost and interactive novelty) have been solved at this point isn’t totally clear, but Cartwheel is making things even more difficult for themselves by going the humanoid route, legs and all. That means dealing with all kinds of problems from motion planning to balancing to safety, all in a way that’s reliable enough for the robot to operate around children.
LaValley is arguably one of the few people who could plausibly make a commercial social humanoid actually happen. His extensive background in humanoid robotics includes nearly a decade at Boston Dynamics working on the Atlas robots, followed by five years at Disney, where he led the team that developed Disney’s Baby Groot robot.
In humanoid robot terms, there’s quite a contrast between the versions of Atlas that LaValley worked on (DRC Atlas in particular) and Baby Groot. They’re obviously designed and built to do very different things, but LaValley says that what really struck him was how his kids reacted when he introduced them to the robots he was working on. “At Boston Dynamics, we were known for terrifying robots,” LaValley remembers. “I was excited to work on the Atlas robots because they were cool technology, but my kids would look at them and go, ‘That’s scary.’ At Disney, I brought my kids in and they would light up with a big smile on their face and ask, ‘Is that really Baby Groot? Can I give it a hug?’ And I thought, this is the type of experience I want to see robots delivering.” While Baby Groot was never a commercial project, for LaValley it marked a pivotal milestone in emotional robotics that shaped his vision for Cartwheel: “Seeing how my kids connected with Baby Groot reframed what robots could and should evoke.”
The current generation of commercial humanoids is pretty much the opposite of what LaValley is looking for. You could argue that this is because they’re designed to do work, rather than be anyone’s friend, but many of the design choices seem to be based on the sort of thing that would be the most eye-catching to the public (and investors) in a rather boringly “futuristic” way. And look, there are plenty of good reasons why you might want to very deliberately design a humanoid with commercial (or at least industrial) aspirations to look or not look a certain way, but for better or worse, nobody is going to like those robots. Respect them? Sure. Think they’re cool? Probably. Want to be friends with them? Not likely. And for Cartwheel, this is the opportunity, LaValley says. “These humanoid robots are built to be tools. They lack personality. They’re soulless. But we’re designing a robot to be a humanoid that humans will want in their day-to-day lives.”
Eventually, Cartwheel’s robots will likely need to be practical (as this rendering suggests) in order to find a place in people’s homes.Cartwheel
Yogi is one of Cartwheel’s prototypes, which LaValley describes as having “toddler proportions,” which are the key to making it appear friendly and approachable. “It has rounded lines, with a big head, and it’s even a little chubby. I don’t see a robot when I see Yogi; I see a character.” A second prototype, called Speedy, is a bit less complicated and is intended to be more of a near-term customizable commercial platform. Think something like Baby Groot, except available as any character you like, and to companies who aren’t Disney. LaValley tells us that a version of Speedy with a special torso designed for a “particular costume” is headed to a customer in the near future.
As the previous generation of social robots learned the hard way, it takes a lot more than good looks for a robot to connect with humans over the long term. Somewhat inevitably, LaValley sees AI as one potential answer to this, since it might offer a way of preserving novelty by keeping interactions fresh. This extends beyond verbal interactions, too, and Cartwheel is experimenting with using AI for whole-body motion generation, where each robot behavior will be unique, even under the same conditions or when given the same inputs.
While Cartwheel is starting with a commercial platform, the end goal is to put these small social humanoids into homes. This means considering safety and affordability in a way that doesn’t really apply to humanoids that are designed to work in warehouses or factories. The small size of Cartwheel’s robots will certainly help with both of those things, but we’re still talking about a robot that’s likely to cost a significant amount—certainly more than a major appliance, although perhaps not as much as a new car, is as much as LaValley was willing to commit to at this point. With that kind of price comes high expectations, and for most people, the only way to justify buying a home humanoid will be if it can somehow be practical as well as lovable.
LaValley is candid about the challenge here: “I don’t have all the answers,” he says. “There’s a lot to figure out.” One approach that’s becoming increasingly common with robots is to go with a service model, where the robot is essentially being rented in the same way that you might pay for the services of a housekeeper or gardener. But again, for that to make sense, Cartwheel’s robots will have to justify themselves financially. “This problem won’t be solved in the next year, or maybe not even in the next five years,” LaValley says. “There are a lot of things we don’t understand—this is going to take a while. We have to work our way to understanding and then addressing the problem set, and our approach is to find development partners and get our robots out into the real world.”
Cartwheel
Cartwheel has been in business for three years now, and got off the ground by providing robotics engineering services to corporate customers. That, along with an initial funding round, allowed LaValley to bootstrap the development of Cartwheel’s own robots, and he expects to deliver a couple dozen variations on Speedy to places like museums and science centers over the next 12 months.
The dream, though, is small home robots that are both companionable and capable, and LaValley is even willing to throw around terms like “general purpose.” “Capability increases over time,” he says, “and maybe our robots will be able to do more than just play with your kids or pick up a few items around the house. I see all robots eventually moving towards general purpose. Our strategy is not to get to general purpose on day one, or even get into the home day one. But we’re working towards that goal. That’s our north star.”
Jeremy is a 31-year-old autistic man who loves music and biking. He’s highly sensitive to lights, sounds, and textures, has difficulty initiating movement, and can say only a few words. Throughout his schooling, it was assumed he was incapable of learning to read and write. But for the past 30 minutes, he’s been wearing an augmented-reality (AR) headset and spelling single words on the HoloBoard, a virtual keyboard that hovers in the air in front of him. And now, at the end of a study session, a researcher asks Jeremy (not his real name) what he thought of the experience.
Deliberately, poking one virtual letter at a time, he types, “That was good.”
It was not obvious that Jeremy would be able to wear an AR headset, let alone use it to communicate. The headset we use, Microsoft’s HoloLens 2, weighs 566 grams (more than a pound), and the straps that encircle the head can be uncomfortable. Interacting with virtual objects requires precise hand and finger movements. What’s more, some people doubt that people like Jeremy can even understand a question or produce a response. And yet, in study after study, we have found that most nonspeaking autistic teenage and adult participants can wear the HoloLens 2, and most can type short words on the HoloBoard.
Nonspeaking autistic people can use the HoloBoard to type independently.
The HoloBoard prototype that Jeremy first used in 2023 was three years in the making. It had its origins in an interdisciplinary feasibility study that considered whether individuals like Jeremy could tolerate a commercial AR headset. That study was led by the three of us: a developmental psychologist (Vikram Jaswal at the University of Virginia), an electrical and software engineer (Diwakar Krishnamurthy at the University of Calgary), and a computer scientist (Mea Wang, also at the University of Calgary).
Our journey to this point was not smooth. Some autism researchers told us that nonspeaking autistic people “do not have language” and so couldn’t possibly communicate by typing. They also said that nonspeaking autistic people are so sensitive to sensory experiences that they would be overwhelmed by augmented reality. But our data, from more than a half-dozen peer-reviewed studies, have shown both assumptions to be wrong. And those results have informed the tools we’re creating, like the HoloBoard, to enable nonspeaking autistic people to communicate more effectively.
Autism is a lifelong neurological condition that affects people in very different ways. It’s most commonly associated with social differences, but many autistic people also have difficulty with communication. In fact, about one-third of autistic children and adults are nonspeaking: Even after years or decades of speech therapy, they cannot communicate effectively using speech. We don’t yet know why, but it may be related to the significant motor challenges associated with producing speech. As with autism in general, nonspeaking autistic people have a range of abilities and language skills: Some are comfortable typing, while others struggle to communicate at all.
Nonspeaking autistic people may also appear inattentive, engage in impulsive behavior, and score poorly on standard intelligence tests (many of which require spoken responses within a set amount of time). Historically, these challenges have led to unfounded assumptions about these individuals’ ability to understand language and their capacity for symbolic thought. To put it bluntly, it has sometimes been assumed that someone who can’t talk is also incapable of thinking.
Most attempts to provide nonspeaking autistic people with an alternative to speech have been rudimentary. Picture-based communication systems, often implemented on an iPad or tablet, are frequently used in schools and therapy clinics. If a user wants a cookie, they can tap a picture of a cookie. But the vocabulary of these systems is limited to the concepts that can be represented by a simple picture.
When asked what he thought of a HoloBoard session, a user typed out a positive review. Ethereal Research Group
There are other options. Some nonspeaking autistic people have learned, over the course of many years and guided by parents and professionals, to communicate by spelling words and sentences on a letterboard that’s held by a trained human assistant—a communication and regulation partner, or CRP. Part of the CRP’s role is to provide attentional and emotional support, which can help with conditions that commonly accompany severe autism and that interfere with communication, including anxiety, attention-deficit hyperactivity disorder, and obsessive-compulsive disorder. Having access to such assisted methods of communication has allowed nonspeaking autistic people to graduate from college, write poetry, and publish a best-selling memoir.
But the role of the CRP has generated considerable controversy. Critics contend that the assistants can subtly guide users to point to particular letters, which would make the CRP, rather than the user, the author of any words produced. If nonspeaking autistic people who use a letterboard really know how to spell, critics ask, why is the CRP necessary? Some professional organizations, including the American Speech-Language-Hearing Association, have even cautioned against teaching nonspeaking autistic people communication methods that involve assistance from another person.
And yet, research suggests that CRP-aided methods can teach users the skills to communicate without assistance; indeed, some individuals who previously required support now type independently. And a recent study by coauthor Jaswal showed that, contrary to critics’ assumptions, most of the nonspeaking autistic individuals in his study (which did not involve a CRP) knew how to spell. For example, in a string of text without any spaces, they knew where one word ended and the next word began. Using eye tracking, Jaswal’s team also showed that nonspeaking autistic people who use a letterboard look at and point to letters too quickly and accurately to be responding to subtle cues from a human assistant.
So how can technology help nonspeaking autistic people communicate? It’s not unusual for researchers to look at a platform technology like AR and imagine how it could be used to help a group of people. However, the ultimate success of any such project isn’t judged by technical innovation or elegance. Rather, the main criterion for success is whether or not the end result is used and useful. An amazing technology that is, say, too delicate or expensive to escape the laboratory is of limited value. And a raft of innovations that miss the mark in meeting the needs of the people it’s supposed to help is similarly limited.
Our focus then was not on improving underlying AR hardware and system software, but finding the most productive ways to adapt it for our users.
We knew we wanted to design a typing system that would allow users to convey anything they wanted. And given the ongoing controversy about assisted communication, we wanted a system that could build the skills needed to type independently. We envisioned a system that would give users more agency and potentially more privacy if the tool is used outside a research setting.
Geoff Ondrich [left] uses the Meta Quest 3 headset to type letters independently via the HoloBoard system. The augmented-reality system can be configured to use either hand tracking or eye tracking to determine which letter the user intends to press. Madison Imber
Augmented reality has various features that, we reasoned, make it attractive for these purposes. AR’s eye- and hand-tracking capabilities could be leveraged in activities that train users in the motor skills needed to type, such as isolating and tapping targets. Some of the CRP’s tasks, like offering encouragement to a user, could be automated and rolled into an AR device. Also, AR allows users to move around freely as they engage with virtual objects, which may be more suitable for autistic people who have trouble staying still: A HoloBoard can “follow” the user around a room using head tracking. What’s more, virtual objects in AR are overlaid on a user’s actual environment, making it safer and less immersive than virtual reality (VR)—and potentially less overwhelming for our target population.
We carefully considered our choice of hardware. While lightweight AR glasses like the Ray-Ban Meta AI glasses and Snap’s AI Spectacles would have been less cumbersome for users, they don’t have the high-fidelity hand-tracking and gaze-tracking we needed. Headsets like the HoloLens 2 and Meta’s Quest 3 provide greater computing power and support a broader range of interaction modalities.
We aren’t the first researchers to consider how AR can help autistic people. Other groups have used AR to offer autistic children real-time information about the emotions people show on their faces, for example, and to gamify social- and motor-skill training. We drew inspiration from those efforts as we took on the new idea of using AR to help nonspeaking autistic people communicate.
Our efforts have been powered by our close collaboration with nonspeaking autistic people. They are, after all, the experts about their condition, and they’re the people best suited to guide the design of any tools intended for them. Everything we do is informed by their input, including the design of prototypes and the studies to test those prototypes.
When neurotypical people see someone who cannot talk, whose body moves in unusual ways, and who acts in socially unconventional ways, they may assume that the person wouldn’t be interested in collaborating or wouldn’t be able to do so. But, as noted by Anne M. Donnellan and others who conduct research with disabled people, behavioral differences don’t necessarily reflect underlying capacities or a lack of interest in social engagement. These researchers have emphasized the importance of presuming competence—in our case, that means expecting nonspeaking autistic people to be able to learn, think, and participate.
Thus, throughout our project, we have invited nonspeaking autistic people to offer suggestions and feedback in whatever manner they prefer, including by pointing to letters on a physical letterboard while supported by a CRP. Although critics of assisted forms of communication may object to this inclusive approach, we have found the contributions of nonspeakers invaluable. Through Zoom meetings, email correspondence, comments after research sessions, and shared Google docs, these participants have provided essential input about whether and how the AR technology we’re developing could be a useful communication tool. In keeping with the community’s interest in more independent communication, our tests of the technology have focused on nonspeakers’ performance without the assistance of a CRP.
A user selects a letter on the HoloBoard by “pushing” it toward a virtual backplate. Successful activation is accompanied by a click and a recorded voice saying the letter aloud.Ethereal Research Group
In early conversations, our collaborators raised several concerns about using AR. For example, they worried that wearing a head-mounted device wouldn’t be comfortable. Our first study investigated this topic and found that, with appropriate support and sufficient time, 15 of 17 nonspeakers wore the device without difficulty. We now have 3D-printed models that replicate the shape and weight of the HoloLens 2, to allow participants to build up tolerance before they participate in actual experiments.
Some users also expressed concern about the potential for sensory overload, and their concerns made us realize that we hadn’t adequately explained the difference between AR and VR. We now provide a video before each study that explains exactly what participants will do and see and shows how AR is less immersive than VR.
Some participants told us that they like the tactile input from interacting with physical objects, including physical letterboards, and were concerned that virtual objects wouldn’t replicate that experience. We currently address this concern using sensory substitution: Letters on the HoloBoard hover slightly in front of a semitransparent virtual backplate. Activating a letter requires the user to “push” it approximately 3 centimeters toward the backplate, and successful activation is accompanied by an audible click and a recorded voice saying the letter aloud.
Our users’ needs and preferences have helped us set priorities for our research program. One person noted that an AR communication system seemed “cool,” but worried that the motor skills required to interact in AR might not be possible without practice. So from the very first app we developed, we built in activities to let users practice the motor skills they needed to succeed.
Participants also told us they wanted to be able to customize the holograms—not just to suit their aesthetic preferences but also to better fit their unique sensory, motor, and attentional profiles. As a result, users of the HoloBoard can choose its color scheme and the size of the virtual letterboard, and whether the letters are said aloud as they’re pressed. We’ve also provided several ways to activate letters: by pressing them, looking at them, or looking at them while using a physical clicker.
We had initially assumed that users would be interested in predictive text capabilities for the HoloBoard—having it autofill likely words based on the first letters typed. However, several people explained that although such a system could theoretically speed up communication, they would find it distracting. We’ve put this idea on the back burner for now; it may eventually become an option that users can toggle on if they wish.
To make things easier for users, we’ve investigated whether the HoloBoard could be positioned automatically in space, dynamically adjusting to the user’s motor skills and movement patterns throughout a session. To this end, we used a behavioral cloning approach: During real-world interactions between nonspeakers and their CRPs, we observed the position of the user’s fingers, palms, head, and physical letterboard. We then used that data to train a machine learning model to automatically adapt the placement of a virtual letterboard for a specific user.
So many assumptions are made about people who cannot speak, including that they don’t have anything to say.
Many nonspeaking participants who currently communicate with human assistance see the HoloBoard as providing a way to communicate with more autonomy. Indeed, we’ve found that after a 10-minute training procedure, most users of the HoloBoard can, like Jeremy, use it to type short words independently. We recently began a six-month study with five participants who have regular sessions in building their typing skills on the HoloBoard.
One of the most common questions from our nonspeaking participants, as well as from parents and professionals, is whether AR could teach the skills needed to type on a standard keyboard. It seems possible, in theory. As a first step, we’re creating other types of AR teaching tools, including an educational AR app that teaches typing in the context of engaging and age-appropriate lessons.
We’ve also begun developing a virtual CRP that can offer support and feedback as a user interacts with the virtual letterboard. This virtual assistant, named ViC, can demonstrate motor movements as a user is learning to spell with the HoloBoard, and also offers verbal prompts and encouragement during a training session. There aren’t many professionals who know how to teach nonspeakers typing skills, so a virtual CRP could be a game changer for this population.
Although nonspeakers have responded enthusiastically to our AR communication tools, our conversations and studies have revealed a number of practical challenges with the current technology.
For starters, most people can’t afford Microsoft’s HoloLens 2, which costs US $3,500. (It’s also recently been discontinued!) So we’ve begun testing our software on less expensive mixed-reality products such as Meta’s $500 Quest 3, and preliminary results have been promising. But regardless of which device is used, most headsets are bulky and heavy. It’s unlikely that someone would wear one throughout a school day, for example. One idea we’re pursuing is to design a pair of AR glasses that’s just for virtual typing; a device customized for a single function would weigh much less than a general-purpose headset.
Shonagh Rae
We’ve also encountered technical challenges. For example, the HoloLens 2’s field of view is only 52 degrees. This restricts the size and placement of holograms, as larger holograms or those positioned incorrectly may be partially or entirely invisible to the user. So when participants use their fingers to point at virtual letters on the HoloBoard, some letters near the edges of the board may fall outside the visible area, which is frustrating to users. To address these issues, we used a vertical layout in our educational app so that the multiple-choice buttons always remain within a user’s field of view. Our systems also allow a researcher or caregiver to monitor an AR session and, if necessary, adjust the size of virtual objects so they’re always in view.
We have a few other ideas for dealing with the field-of-view issue, including deploying devices that have a larger field of view. Another strategy is to use eye tracking to select letters, which would eliminate the reliance on hand movements and the problem of the user’s pointing fingers obscuring the letters. And some users might prefer using a joystick or other handheld controller to navigate and select letters. Together, these techniques should make the system more accessible while working within hardware constraints.
We have also been developing cross-reality apps, which allow two or more people wearing AR headsets to interact within the same virtual space. That’s the setup we use to enable researchers to monitor study sessions in real time. Based on our development experience, we created an open-source tool called SimpleShare for the development of multiuser extended-reality apps in a device-agnostic way. A related issue is that many of our users make sudden movements; a sudden shake of a head can interfere with the sensors on the AR headset and upset the spatial alignment between multiple headsets. So our apps and SimpleShare instruct the headset to routinely scan the environment and use that data to automatically realign multiple devices, if necessary.
We’ve had to find solutions to cope with the limited computing power available on AR headsets. Running the AI model that automates the custom placement of the HoloBoard for each user can cause a lag in letterboard interactions and can cause the headset to heat up. We solved this problem by simplifying the AI model and decreasing the frequency of the model’s interventions. Rendering a realistic virtual CRP via a headset is also computationally intensive. In our virtual CRP work, we’re now rendering the avatar on an edge device, such as a laptop with a state-of-the-art GPU, and streaming it to the display.
As we continue to tackle these technology challenges, we’re well aware that we don’t have all the answers. That’s why we discuss the problems that we’re working on with the nonspeaking autistic people who will use the technology. Their perspectives are helping us make progress toward a truly usable and useful device.
So many assumptions are made about people who cannot speak, including that they don’t have anything to say. We went into this project presuming competence in nonspeaking people, and yet we still weren’t sure if our participants would be able to adapt to our technology. In our initial work, we were unsure whether nonspeakers could wear the AR device or interact with virtual buttons. They easily did both. In our evaluation of the HoloBoard prototype, we didn’t know if users could type on a virtual letterboard hovering in front of them. They did so while we watched. In a recent study investigating whether nonspeakers could select letters using eye-gaze tracking, we wondered if they could complete the built-in gaze-calibration procedure. They did.
The ability to communicate—to share information, memories, opinions—is essential to well-being. Unfortunately, most autistic people who can’t communicate using speech are never provided an effective alternative. Without a way to convey their thoughts, they are deprived of educational, social, community, and employment opportunities.
We aren’t so naïve as to think that AR is a silver bullet. But we’re hopeful that there will be more community collaborations like ours, which take seriously the lived experiences of nonspeaking autistic people and lead to new technologies to support them. Their voices may be stuck inside, but they deserve to be heard.
Keysight visited 6G researchers at Northeastern University who are working to overcome the challenges of high-speed, high-bandwidth wireless communication.
They shared concepts from their cutting-edge research, including overcoming increased path loss and noise at higher frequencies, potential digital threats to communication channels, and real-time upper-layer network applications.
During this event, you will gain insights into the following 6G topics:
Video Friday is your weekly selection of awesome robotics videos, collected by your friends at IEEE Spectrum robotics. We also post a weekly calendar of upcoming robotics events for the next few months. Please send us your events for inclusion.
Enjoy today’s videos!
Today I learned that “hippotherapy” is not quite what I wanted it to be.
The integration of KUKA robots into robotic physiotherapy equipment offers numerous advantages, such as precise motion planning and control of robot-assisted therapy, individualized training, reduced therapist workload and patient-progress monitoring. As a result, these robotic therapies can be superior to many conventional physical therapies in restabilizing patients’ limbs.
[ Kuka ]
MIT engineers are getting in on the robotic ping-pong game with a powerful, lightweight design that returns shots with high-speed precision. The new table-tennis bot comprises a multijointed robotic arm that is fixed to one end of a ping-pong table and wields a standard ping-pong paddle. Aided by several high-speed cameras and a high-bandwidth predictive control system, the robot quickly estimates the speed and trajectory of an incoming ball and executes one of several swing types—loop, drive, or chop—to precisely hit the ball to a desired location on the table with various types of spin.
[ MIT News ]
Pan flipping involves dynamically flipping various objects, such as eggs, burger buns, and meat patties. This demonstrates precision, agility, and the ability to adapt to different challenges in motion control. Our framework enables robots to learn highly dynamic movements.
[ GitHub ] via [ Human Centered Autonomy Lab ]
Thanks, Haonan!
An edible robot made by EPFL scientists leverages a combination of biodegradable fuel and surface tension to zip around the water’s surface, creating a safe—and nutritious—alternative to environmental monitoring devices made from artificial polymers and electronics.
[ EPFL ]
Traditional quadcopters excel in flight agility and maneuverability but often face limitations in hovering efficiency and horizontal field of view. Nature-inspired rotary wings, while offering a broader perspective and enhanced hovering efficiency, are hampered by substantial angular momentum restrictions. In this study, we introduce QuadRotary, a novel vehicle that integrates the strengths of both flight characteristics through a reconfigurable design.
[ Paper ] via [ Singapore University of Technology and Design ]
I like the idea of a humanoid that uses jumping as a primary locomotion mode not because it has to, but because it’s fun.
[ PAL Robotics ]
I had not realized how much nuance there is to digging stuff up with a shovel.
[ Intelligent Motion Laboratory ]
A new 10,000-gallon [38,000-liter] water tank at the University of Michigan will help researchers design, build, and test a variety of autonomous underwater systems that could help robots map lakes and oceans and conduct inspections of ships and bridges. The tank, funded by the Office of Naval Research, allows roboticists to further test projects on robot control and behavior, marine sensing and perception, and multivehicle coordination.
“The lore is that this helps to jump-start research, as each testing tank is a living reservoir for all of the knowledge gained from within it,” said Jason Bundoff, lead engineer in research at U-M’s Friedman Marine Hydrodynamics Laboratory. “You mix the waters from other tanks to imbue the newly founded tank with all of that living knowledge from the other tanks, which helps to keep the knowledge from being lost.”
If you have a humanoid robot and you’re wondering how it should communicate, here’s the answer.
[ Pollen ]
Whose side are you on, Dusty?
Even construction robots should be mindful about siding with the Empire, though there can be consequences!
- YouTube
[ Dusty Robotics ]
This Michigan Robotics Seminar is by Danfei Xu from Georgia Tech, on “Generative Task and Motion Planning.”
Long-horizon planning is fundamental to our ability to solve complex physical problems, from using tools to cooking dinners. Despite recent progress in commonsense-rich foundation models, the ability to do the same is still lacking in robots, particularly with learning-based approaches. In this talk, I will present a body of work that aims to transform Task and Motion Planning—one of the most powerful computational frameworks in robot planning—into a fully generative model framework, enabling compositional generalization in a largely data-driven approach.
