Tag Archives: can

#442075 How ancient sea creatures can inform ...

Soft robotics is the study of creating robots from soft materials, which has the advantage of flexibility and safety in human interactions. These robots are well-suited for applications ranging from medical devices to enhancing efficiency in various tasks. Additionally, using different forms of robotic movement may also serve us well in exploring the ocean or space, or doing certain jobs in those environments. Continue reading

Posted in Human Robots

#442010 Robots that can autonomously build ...

A combined team of engineers from NASA Ames Research Center and KBR has designed and built a robot system that can autonomously build structures using specially designed lattice blocks. In their paper published in the journal Science Robotics, the group describes the robots and the lattice blocks they use to build structures and how they whole system works. Continue reading

Posted in Human Robots

#441944 This AI Trained on the Life Events of ...

The possibility of mapping out someone’s entire life in advance is both exciting and scary. A new artificial intelligence trained on the personal data of every person in Denmark can do just that.

Today’s deep learning-based AI systems are prediction machines. They work by ingesting vast amounts of data and using it to pick out statistical patterns that can be used to make informed guesses about previously unseen data.

Despite the uncannily fluent linguistic capabilities of AI chatbots, they operate in much the same way. They learn from huge amounts of text data and then attempt to predict what word comes next in a string of text.

What allowed the breakthrough in capabilities that we’ve seen in the last few years was a new deep learning architecture, known as a transformer, that can train on far more data than previous algorithms. It turns out that when you can train models on almost the entirety of the internet, their predictions become very sophisticated.

Now researchers have shown that they can use the same kind of techniques to train a model on a huge database of health, social, and economic information collected by the Danish government. The resulting AI was able to make highly accurate predictions about people’s lives, including how likely they are to die in a given time window and their personality traits.

“The model opens up important positive and negative perspectives to discuss and address politically,” Sune Lehmann from the Technical University of Denmark, who led the study, said in a statement. “Similar technologies for predicting life events and human behavior are already used today inside tech companies that, for example, track our behavior on social networks, profile us extremely accurately, and use these profiles to predict our behavior and influence us.”

The dataset the researchers used spans from 2008 to 2020 and includes all six million Danes. It features information on their income, job, social benefits, visits to healthcare providers, and disease diagnoses, among other things.

Getting the data into a format a transformer can understand took some work though. They restructured all the information in the database into what they call “life sequences,” with all the events associated with each individual organized in chronological order. This makes it possible to do next-event prediction in much the same way an AI chatbot does next-word prediction.

When trained on large numbers of these life sequences, the model can start to pick out patterns that connect disparate events in someone’s life and help it make predictions about the future. The researchers trained their model on the life sequences of people aged 25 to 70 between the years of 2008 and 2016 and then used it to make predictions about the next four years.

When they asked it to guess the likelihood of someone dying in that period, it outperformed the current state-of-the-art by 11 percent. They also got the model to make predictions about how people scored on a personality test, and the results outperformed models specifically trained for that task.

While the performance on those two tasks is impressive, in a paper describing the research in Nature Computational Science, the team points out that what’s really exciting about the model is the fact it can potentially be used to make all kinds of predictions about people’s lives. Previously, AI has normally been trained to answer specific questions about people’s health or social trajectories.

Obviously, this kind of research raises some thorny questions about privacy and human agency. But the researchers point out that private companies are almost certainly doing similar things with their own data, so it’s useful to understand what these kinds of techniques make possible.

And given the rapidly advancing capabilities of AI, it will be important to have public debates about what kind of AI-powered predictions we allow in both the private and public sphere, says Lehmann.

“I don’t have those answers,” he said in a press release. “But it’s high time we start the conversation because what we know is that detailed prediction about human lives is already happening and right now there is no conversation and it’s happening behind closed doors.”

Image Credit: Nat / Unsplash Continue reading

Posted in Human Robots

#441933 This Robotic Pack Mule Can Carry Your ...

This article is part of our exclusive IEEE Journal Watch series in partnership with IEEE Xplore.

The useful niche that quadrupedal robots seem to have found for themselves, at least for the moment, is infrastructure inspection. They’ve had a mild to moderate amount of success monitoring industrial sites, tracking construction progress, and things of that nature. Which is great! But when you look at what humans have historically relied on quadrupeds for, there’s a little bit of situational awareness (in the form of security), but the majority of what these animals have done for us is manual labor.

In a paper published last month in IEEE Robotics and Automation Letters, roboticists from the Robotic Systems Lab at ETH Zurich are aiming to address the fact that “legged robots are still too weak, slow, inefficient, or fragile to take over tasks that involve heavy payloads.” Their new robot that is none of these things is Barry, which can efficiently carry up to 90 kilograms so that you don’t have to.

If you go back far enough, a bunch of the initial funding for quadrupedal robots that enabled the commercial platforms that are available today was tied into the idea of robotic pack animals. Boston Dynamics’ BigDog and LS3 were explicitly designed to haul heavy loads (up to 200 kg) across rough terrain for the U.S. Military. This kind of application may be obvious, but the hardware requirements are challenging. Boston Dynamics’ large quadrupeds were all driven by hydraulics, which depended on the power density of gasoline to function, and ultimately they were too complex and noisy for the military to adopt. The current generation of quadruped robots, like Spot and ANYmal, have a payload of between 10 and 15 kg.

Barry manages to carry 50 percent of the payload of LS3 in a much smaller, more efficient, and quieter form factor. It’s essentially a customized ANYmal, using unique high-efficiency electric actuators rather than hydraulics. The robot itself weighs 48 kg, and can handle unmodeled 90 kg payloads, meaning that Barry doesn’t have to know the size, weight, or mass distribution of what it’s carrying. It’s a key capability, because it makes Barry’s payload capacity actually useful, as the paper’s first author Giorgio Valsecchi explains: “When we use a wheelbarrow, we don’t have to change any settings on it, regardless of what we load it with—any manual adjustment is a bottleneck in usability. Why should a ‘smart’ robot be any different?” This is really what makes Barry’s payload capacity actually real-world useful, and also means that if you want to, you can even ride it.

Barry: A High-Payload and Agile Quadruped Robot

youtu.be

Barry’s heroic payload is enabled by its custom actuators. While the standard approach for developing powered robotic joints involves choosing the smallest motor capable of producing the required peak power, Barry focuses on motor efficiency instead. “It turns out that the ideal solution is to have the biggest possible motor,” Valsecchi says. “It is a bit counterintuitive, but bigger motors are more efficient, they consume less energy when performing the same task. This results in a robot with more payload capabilities and a lower cost of transport.” Barry is actually quite efficient: with a cost of transport of just 0.7, it can operate with a payload for over two hours and travel nearly 10 km.

The commercial potential for a robot like Barry is obvious, and Valsecchi is already thinking about several use cases: “carrying raw materials on construction sites to prevent injuries and increase productivity, carrying equipment in search and rescue operations to free up rescuers from excessive loads… The same technology could be used to design a walking wheelchair, and we actually got some requests for this specific use case. Once we started showing the robot with a big box on top, people realized a lot of things could be done.”

At the moment, Barry doesn’t yet have much in the way of perception, so giving the robot the ability to intelligently navigate around obstacles and over complex terrain is one of the things that the researchers will be working on next. They’re also starting to think about potential commercial applications, and it certainly seems like there’s a market for a robot like this—heck, I’d buy one.

The preserved 200 year old body of the original Barry.Photo via Wikipedia by PraktikantinNMBE and reproduced under CC BY-SA 4.0.

Barry, by the way, is named after a legendary St. Bernard who saved the lives of more than 40 people in the Swiss Alps in the early 1800s, including by carrying them to safety on his back. “Being able to ride the robot was one of our ambitions,” Valsecchi tells us. “When we managed to accomplish that I thought we did well enough to tribute the original Barry by using his name, to convey our vision of what robots could become.” Barry the dog died in 1814 (apparently stabbed by someone he was trying to rescue who thought he was a wolf), but his preserved body is on display at the Natural History Museum in Bern.
Barry: A High-Payload and Agile Quadruped Robot, by Giorgio Valsecchi, Nikita Rudin, Lennart Nachtigall, Konrad Mayer, Fabian Tischhauser, and Marco Hutter from ETH Zurich, is published in IEEE Robotics and Automation Letters. Continue reading

Posted in Human Robots

#441891 Octopus-inspired robot arm can grab and ...

A team of engineers at Beihang University, working with a colleague from Tsinghua University, both in China, has designed, built and tested a haptically controlled octopus robot arm that is capable of grasping, lifting and carrying objects on land and underwater. In an article published in the journal Science Robotics, the group describes how they built their robot, how it works and how well it performed when tested under a variety of scenarios. Continue reading

Posted in Human Robots