Humanoid robots have just begun stepping into Amazon warehouses and Mercedes-Benz automotive factories. Now, they are being recruited for an even more ambitious effort – the creation of artificial general intelligence with capabilities comparable to those of humans.

US computing firm Nvidia, which has become one of the world’s most valuable companies through its AI chip sales, recently announced several hardware and software products to boost humanoid robot training. The centrepiece is a “moonshot” initiative, called Project…

A mucus-excreting robot with a single large foot can effectively imitate the way snails crawl over surfaces – even steeply inclined ones.

“I always say that snails are like Michael Jackson to me. You don’t see how they move, but somehow gliding is happening,” said Saravana Prashanth Murali Babu at the University of Southern Denmark during a presentation at the American Physical Society’s March Meeting in Minneapolis, Minnesota, on 4 March.

Fascinated by the shelled molluscs, Saravana and his colleagues decided to build a version of a snail’s single large, soft foot and use it as the basis of a robot that moves like a snail.

During his presentation, Saravana explained that the team chose to build the foot from a soft material that could be inflated in segments by small pneumatic pumps. He says that while the chemical properties of snail mucus have previously been studied in detail, the way the snail’s foot moves had only been hypothesised based on biologists’ observations. These past studies proposed that different parts of a snail’s foot hit the ground, then detach from it before hitting it again, and they do so out of sync with each other. This creates a wave form across the foot that lets the snail glide forward on its mucus.

The researchers replicated this “pedal wave” motion in their experimental robot, which could also excrete mucus, and saw it successfully move forward and make turns without falling over. It even managed to move up steeply inclined surfaces in some experiments, says Saravana.

Although the bot is still at the experimental stage, Saravana hopes that it will become the first ever robot that propels itself like a snail. To make it more self-contained, the team is experimenting with putting the pumps in a snail-like shell on top of the robot. The shell – a slightly oversized plastic replica of a real snail shell – can also house electronics for remotely controlling the robot, and a syringe system for releasing mucus beneath the robotic foot, mimicking a real snail’s slimy trail.

The team’s final goal, however, is to make the robot’s inflatable foot even softer and more similar to real snails, whose bodies are mostly made of water. The researchers hope that a robot that successfully moves on mucus could inform the design of soft medical robots that could eventually move inside the human body where mucus is also abundant.

Topics:

A robot dog can use a leg to open doors, press buttons and pick up rucksacks while balancing on its other three legs.

Four-legged robots like Spot, the star of Boston Dynamics’ viral videos, normally need an arm attached to their body to open doors or pick up objects, but this can add significant weight and make it harder for the robot to squeeze through narrow spaces.

Philip Arm at ETH Zurich in Switzerland and his colleagues used a machine-learning model to teach an off-the-shelf robotic dog to use one of its legs to perform tasks while standing still or moving with the other three legs.

“We cannot do everything with the legs that we could do with an arm – right now, a hand is way more dexterous. But the point is really to make this work for applications where you maybe have mass constraints, or we don’t want to have that additional complexity, like for space exploration where every kilogram of such a robot counts,” says Arm.

To train the dog – an ANYmal robot made by ANYbotics – Arm and his team gave the machine-learning model the objective of finding a specific point in space with one of the robot’s legs. The model then worked out by itself how to control the remaining three legs and balance the robot while standing or walking.

Arm and his team could then control the robot remotely to carry out movements like picking up a backpack and putting it in a box, or collecting rocks. While the robot can currently only do these tasks while operated by a person, Arm hopes that future improvements will allow the dog to autonomously manipulate objects with its leg.

Topics:

A wheeled robot set loose on a college campus has figured out how to open all kinds of doors and drawers while rolling around in the real world.

The robot adapted to new challenges on its own – paving the way for machines capable of independently interacting with physical objects. “You want the robots to work autonomously… without relying on humans to keep giving examples at test time for every new kind of scenario that you’re in,” says Deepak Pathak at Carnegie Mellon University (CMU) in Pennsylvania.

Pathak and his colleagues initially trained the robot through imitation learning, providing visual examples of how to open objects such as doors, cabinets, drawers and refrigerators. They then turned it loose on the CMU campus to try opening doors and cabinets it had never encountered before. This required the robot to adapt to each new object using artificial intelligence that rewarded it for figuring things out.

The robot typically spent from 30 minutes to an hour learning how to consistently open each object, says Haoyu Xiong at CMU, who built the robot and scouted out the campus for a wide variety of test locations. The team included 12 training objects for practice and then eight additional objects as a test of the robot’s capabilities.

Although its initial success rate was about 50 per cent on average, the robot sometimes completely failed to open a new object when first starting out. By the end, its success rate rose to about 95 per cent.

In addition to learning on the fly, it had to be able to physically handle heavy doors,  says Russell Mendonca at CMU. Achieving both goals cost $25,000, he says, which is much less expensive than other robotic systems with adaptive learning capabilities.

The robotic demonstration outside the lab “marks a concrete step toward more general robotic manipulation systems”, says Yunzhu Li at the University of Illinois Urbana-Champaign. “Opening doors and drawers – a seemingly simple task for humans – is actually surprisingly difficult for robots,” he says.

Topics:

A tiny, bipedal robot that combines muscle tissue with artificial materials can walk and turn by contracting its muscles.

While biohybrid robots that crawl and swim have been built before with lab-grown muscle, this is the first such bipedal robot that can pivot and make sharp turns. It does this by applying electricity to one of its legs to make the muscle contract, while the other leg remains anchored. The muscle acts as a biological actuator – a component that converts electrical energy into mechanical force.

At the moment, the robot, which is only 3 centimetres tall, cannot support itself in air and has a foam buoy to help it stand up in a water tank. The muscles are grown from rat cells in a laboratory.

“This is still basic research,” says team member Shoji Takeuchi at the University of Tokyo, Japan. “We are not at the stage where this robot itself can be used anywhere. To make it work in the air, many more related issues would need to be solved, but we believe it can be done by increasing the muscular strength.”

The robot is still extraordinarily slow by human standards, moving just 5.4 millimetres per minute. It also takes over a minute to turn 90 degrees, with an electric stimulation every 5 seconds.

Takeuchi hopes the team can make the robot faster by optimising the pattern of electrical stimulation and improving the design.

“The next step for the biohybrid robot would be to develop a version with joints and additional muscle tissues for more sophisticated walking capabilities,” he says. “Thick muscles would also need to be built to increase strength.”

To walk in air rather than water, the robot would also need a nutrient supply system to keep the muscle tissue alive.

Victoria Webster-Wood at Carnegie Mellon University in Pennsylvania says the study is an interesting proof of concept for biohybrid robots.

“These types of biohybrid robots are useful tools for studying engineered muscle tissue and investigating how to control biological actuators,” says Webster-Wood. “As the force and control capabilities advance through this type of scientific research, the ability to apply these actuators to more complex robots will increase.”

Topics:

 

A humanoid robot can relay video and touch sensations to a person wearing haptic feedback gloves and a virtual reality (VR) headset hundreds of kilometres away, offering a way for people to attend events without travelling.

The iCub 3 robot is a 52-kilogram, 125- centimetre-tall robot with 54 points of articulation across its aluminium alloy and plastic body. Its head contains two cameras where a human’s eyes would be, and an internet-connected computer where the brain would go. Along with the cameras, sensors covering its body send data to the robot’s “brain”. These sensations are then replicated on a suit and VR headset worn by a remote human operator.

When the operator reacts to what they see and feel, the suit’s sensors pick up the movements and the robot matches them. “The key is to translate every signal and bit of numeric data that can be sent through the network,” says Stefano Dafarra at the Italian Institute of Technology, who was part of the iCub 3 team. There can be a small delay of up to 100 milliseconds to capture and transmit the visual footage, but the operator can mitigate this by moving slightly slower than normal.

The team has demonstrated the robot at the Venice Biennale, where it wandered through an exhibition while its operator stood 290 kilometres away in Genoa.

Dafarra hopes people will use the iCub 3 to attend events remotely, reducing the need to travel. But at present, a fall could be hugely damaging to the robot, and it’s uncertain whether it could stand up again on its own, he says.

“iCub 3 is an interesting robot and offers clear advantages from the previous iteration,” says Jonathan Aitken at the University of Sheffield, UK, whose laboratory owns a prior version of the robot. However, he is disappointed that the team wasn’t clear in its research about the data transmission requirements of the new version of the robot. “It would be good to know just how much data was required, and what the upper and lower bounds were,” he says.

Topics:

A robot that can grow around trees or rocks like a vine could be used to make buildings or measure pollution in hard-to-reach natural environments.

Vine-like robots aren’t new, but they are often designed to rely on just a single sense to grow upwards, such as heat or light, which means they don’t work as well in some settings as others.

Emanuela Del Dottore at the Italian Institute of Technology and her colleagues have developed a new version, called FiloBot, that can use light, shade or gravity as a guide. It grows by coiling a plastic filament into a cylindrical shape, adding new layers to its body just behind the head that contains the sensors.

“Our robot has an embedded microcontroller that can process multiple stimuli and direct the growth at a precise location, the tip, ensuring the body structure is preserved,” she says.

This fine control of the tip’s direction means the robot can easily navigate unfamiliar terrain, says Dottore, by wrapping itself around trees or using the shaded parts of leaves as signposts.

FiloBot grows at around 7 millimetres per minute. While slower than many conventional robots, this gentle progress could mean it doesn’t disrupt sensitive natural environments, she says.

The team doesn’t have an exact use for the robot at present, but hopes it could be deployed to collect data in places that are hard for humans to reach, like treetops.

Topics:

An artificial skin is helping a robot to recognise the difference between picking up inanimate objects and living sea creatures such as starfish and shellfish. That sense of touch could prove useful in cleaning up the ocean, doing underwater exploration or even carrying out deep-sea mining on the seafloor.

The artificial skin’s sense of touch harnesses what is known as the magnetoelastic effect – changes that occur in the magnetic field of materials as they are pushed and pulled. This phenomenon is unaffected by underwater…