First developed more than 100,000 years ago, clothing is one of humanity’s earliest — and most culturally significant — inventions, providing wearers not just protection from the environment and elements but also signifying social status, membership in a community and their role within that group. As robots increasingly move out of labs, off of factory floors and into our everyday lives, a similar garment revolution could soon be upon us once again, according to a new research study out of New York’s Cornell University.

“We believe that robot clothes present an underutilized opportunity for the field of designing interactive systems,” the team argues in What Robots Need From Clothing, which was submitted to the In Designing Interactive Systems Conference 2021. “Clothes can help robots become better robots — by helping them be useful in a new, wider array of contexts, or better adapt and function in the contexts they are already in.”

“I started by looking at how different materials would move on robots and thinking about the readability of that motion — like, what is the robot's intention based on the way materials move on the robot,” Natalie Friedman, a PhD student at Cornell Tech and lead author on the paper, explained to Engadget. “From there, I started thinking about all the different social functions that clothes have for people and how that could influence how the robot is viewed.”

While tomorrow’s robots may wear white button down dress shirts and black bow ties while serving hors d'oeuvres to party guests or wear candy stripes while working as nurses, it’s not simply a matter of tossing human clothing onto a robotic chassis. “What robot clothes are is integrally tied to what robots need from clothing. Robot clothing should analogously fulfill needs robots have, rather than just being human clothes on a robot,” the researchers wrote.

Robo-clothes could take any number of forms, depending on their wearer’s specific function. Robotic firefighters, such as the Thermite from Howe and Howe, might theoretically be issued heat-resistant overcoats akin to what humans wear but embedded with thermochromic ink to provide the robot’s operator an easy visual reference to the area’s ambient temperature or indicate that the robot is in danger of overheating. Conversely, search-and-rescue bots could wear waterproof garments when conducting oceanic operations and then strap on extra-grippy boots when searching for lost hikers in mountainous terrain or survivors of a building collapse.

"I think this work is important to helping engineers and technologists understand the functional importance of aesthetics and signaling in design,” Cornell Tech professor and co-author Wendy Ju, said in a recent blog. “It's not ‘just fashion’ - what the robot wears helps people understand how to interact with it in ways that are critical to safety and task execution."

Overall, the use of swappable attire could lead to more generalized robot designs as the specific capabilities the clothing provides don't have to be baked into the robot’s construction. “It is more difficult to build a new robot than to build new clothes,” Friedman said. “I think that clothes are going to influence robot design and robot designs are going to influence clothes. Maybe it'll start in one direction — clothes made to fit robots — but, in the future, I think that robots might be built to better fit in clothes.” She notes that Pepper, though recently discontinued by SoftBank, offers an online merch store with a wide variety of costumes and outfits for the robot to wear including outfits designating cultural, national, professional and religious affiliations.

NurPhoto via Getty Images

But clothing on robots isn’t just for their own benefit, it also serves to demystify and humanize these cutting-edge machines in the eyes of the people they’re working with. For example, clothing could help protect a robot’s sense of shame — or rather that of its user.

“The need for wire modesty — to cover up nudity — stems from anthropomorphic priggishness, since robots do not get embarrassed about wires poking out of them,” the researchers wrote. “However, both humanoid and non-humanoid robots have pragmatic reasons to maintain a clean and covered aesthetic, because exposed wires present a real risk to function. Any wire that is pulled out or cut will remove power or signal to a subsystem, and that can be risky to the robot and any people or objects in the environment.”

“I definitely see a future where [when robots] aren't wearing clothes, it might look a little funny,” Friedman added. “I mean we are just mapping our ideas onto robots, right? Robots don’t have consciousness, so they don't feel shame.”

However, putting clothes on robots could also prove problematic especially if the apparel style has been culturally appropriated. You can bet your bottom dollar that the first cannabis dispensary to dress an automated budtender in rastafarian garb is going to make headlines — and not the kind that are good for business — same as if you outfitted a Roomba with a Native American headdress. “Hawaiian shirts, for example, used to be a marker of ‘casual Friday’ office attire, but more recently are affiliated with the extremist ‘Boogaloo Boys,’” the researchers wrote.

Despite the potential drawbacks to putting pants on robots, doing so could help make the entire field of research more attractive to a new generation of roboticists. “I like to think about girls in robotics,” Friedman said. “When they're young, I think robotics seems like a really intimidating thing but I see clothes as kind of a way to welcome, you know, the stereotypically feminine... skills that women have. I see clothes as a way to welcome girls into [robotics].”

Alphabet has launched another company in its X moonshot factory, and this one may be its most ambitious robotics project to date. The just-opened firm, Intrinsic, plans to make industrial robots more accessible to people and businesses that couldn't otherwise justify the effort involved to teach the machines. You could see robotic manufacturing in more countries, for example, or small businesses that can automate production that previously required manual labor.

Intrinsic will focus on software tools that make these robots easier to use, more flexible and more affordable. To that end, the company has been testing a mix of software tools that include AI techniques like automated perception, motion planning and reinforcement learning. Company chief Wendy Tan-White has relevant experience, too. She started the "world's first" software-as-a-service site builder to make web development more accessible, launched early online banking and lending services, and helped nurture startups as a VP at X.

The technology is still early, but there are already promising results. During its development time at X, the team trained a robot to make a USB connection in two hours (instead of programming it over hundreds of hours) and had robotic arms build simple furniture (shown below). Automation wouldn't be "realistic or affordable" for efforts like these using existing technology, Intrinsic said.

Intrinsic/X

The new company still has much work to do. It's now focused more on creating a practical product and "validating" its tech. It's also hunting for partners in car manufacturing, electronics and healthcare that currently use industrial robots. If Intrinsic succeeds, though, it could make robotics more equitable and fill gaps in production. The company even suggests that its work could help the environment — the closer robotic factories are to people, the lower the emissions needed to transport goods to customers.

This might prove challenging. Rethink Robotics spent years developing collaborative robots that learn through simple human guidance, only to shut down as sales fell short. X moonshot companies also aren't guaranteed to succeed — look at Loon's fate as an example. Alphabet's money could help where companies like Rethink struggled, however, and Intrinsic is focused more on solving overall robotics problems rather than specific scenarios. This effort might stand a better chance than most.

Soft robots still tend to rely on hard electronics to function, but a new invention might reduce that need for unyielding chips. UC Riverside researchers have developed pneumatic computer memory that they used to help a soft robot play the piano.

Instead of conventional transistors and electric circuits, the "air-powered" memory relies on microfluidic valves that control airflow. Atmospheric pressure in a given valve represents a binary "0," while a vacuum indicates a "1." The researchers' memory has a complex-enough array of these valves to function like an 8-bit RAM chip — not exactly powerful, but good enough that a pair of soft robot hands can play "Mary Had a Little Lamb" at a slow but steady pace.

The absence of positive pressure makes this particularly safe — there's no danger of the memory exploding in mid-use.

The technology is far from ready for everyday use. Besides needed improvements to complexity and speed, a robot would need soft versions of processors and other components to completely eliminate the need for rigid electronics. The goal is clear, however. Pneumatic memory could at least reduce the need for chips in soft robots, and points to a future of completely flexible robotics that shouldn't hurt you if there's a collision.

Robots have plenty of potential to help people with limited mobility, including models that could help the infirm put on clothes. That's a particularly challenging task, however, that requires dexterity, safety and speed. Now, scientists at MIT CSAIL have developed an algorithm that strikes a balance by allowing for non-harmful impacts rather than not permitting any impacts at all as before. 

Humans are hardwired to accommodate and adjust to other humans, but robots have to learn all that from scratch. For example, it's relatively easy for a person to help someone else dress, as we know instinctively where to hold the clothing item, how people can bend their arms, how cloth reacts and more. However, robots have to be programmed with all that information. 

In the past, algorithms have prevented robots from making any impact with humans at all in the interest of safety. However, that can lead to something called the "freezing robot" problem, where the robot essentially stops moving and can't accomplish the task it set out to do. 

To get past that issue, an MIT CSAIL team led by PhD student Shen Li developed an algorithm that redefines robotic motion safety by allowing for "safe impacts" on top of collision avoidance. This lets the robot make non-harmful contact with a human to achieve its task, as long as its impact on the human is low.

"Developing algorithms to prevent physical harm without unnecessarily impacting the task efficiency is a critical challenge," said Li. "By allowing robots to make non-harmful impact with humans, our method can find efficient robot trajectories to dress the human with a safety guarantee."

For a simple dressing task, the system worked even if the person was doing other activities like checking a phone, as shown in the video above. It does that by combining multiple models for different situations, rather than relying on a single model as before. "This multifaceted approach combines set theory, human-aware safety constraints, human motion prediction and feedback control for safe human-robot interaction," said Carnegie Mellon University's Zackory Erickson.

The research is still in the early stages, but the ideas could be used areas other than just dressing. "This research could potentially be applied to a wide variety of assistive robotics scenarios, towards the ultimate goal of enabling robots to provide safer physical assistance to people with disabilities," Erickson said. 

Grubhub is teaming up with Russian tech giant Yandex to deliver food to students and others on US college campuses with the help of autonomous robots. The companies have agreed a multi-year partnership, and the robots will start dropping off orders on select campuses this fall. Grubhub works with more than 250 colleges across the country.

Yandex says its robots can access areas and navigate obstacles that cars cannot. It will be able to deliver food in mainly pedestrian areas and the robot delivery service will be integrated into Grubhub's app. When the robot gets close to its destination, the customer will receive a notification. They can retrieve their order by using the app to unlock a hatch on the robot.

The machines use the same self-driving tech as Yandex's autonomous cars, and they can operate in a variety of weather conditions. Yandex has been using the robots for its own food and grocery delivery services in Russia. The rovers have also been fulfilling restaurant orders in Ann Arbor, Michigan, since April.

We may have to say farewell to SoftBank's adorable humanoid robot Pepper. According to Reuters, the Japanese conglomerate has stopped the robot's production last year and is slashing jobs across robotics-related businesses in several countries. Apparently, there wasn't much demand for Pepper, and SoftBank only ever produced 27,000 units manufactured by Foxconn. 

Nikkei has also reported that Pepper's production was halted due to weak demand, but the SoftBank rep it talked to denied that the company is killing the robot entirely. "We plan to resume production if demand recovers," the spokesperson said.

While Pepper sold out in under a minute when it was released in Japan in 2015, the company only produced 1,000 machines for its consumer launch. Pepper was built as a social robot that can recognize faces and basic human emotions, so it can interact with people. Most of the units SoftBank produced are leased to corporate clients, and the company also placed the robot in its mobile phone stores in Japan.

With a price of over $1,600, though, Pepper a bit too expensive for most developers and small businesses. Reuters' sources said its sales suffered from limited functionality and unreliability. Further, SoftBank wasn't able to give the robot more features, because culture clashes between its French business in charge of robotics projects and its Tokyo management reportedly affected Pepper's development. 

After SoftBank shifted its focus to the cleaning robot Whiz, Pepper was sidelined. The company plans to cut about half of its 330 staff positions in France in September, Reuters said, and half of the sales staff positions in the US and Britain had already been cut.

If you want to “enhance your athletic training regimen,” or perhaps just have a bit of fun with robotically launched ping pong balls, then be sure to check out the Pingo apparatus shown in the video below. This robot moves back and forth on four DC motor-powered wheels, searching for targets with an ultrasonic rangefinder.

When something comes into view, Pingo adjusts its ping pong launching tube’s angle to match the target distance, then loads a ball and flings it into the air with a pair of spinning disks. 

The device is controlled by an Arduino Mega and uses a half-dozen DC motors, a pair of steppers, and even a servo to accomplish its mission.

When you step out in public, you’ll often be filmed by a number of cameras and perhaps even be analyzed by tracking software of some kind. The Watchman robot head by Graham Jessup, however, makes this incredibly obvious as it detects and recognizes facial movements, then causes a pair of eyeballs to follow you around.

The 3D-printed system — which is a modified version of Tjhazi’s Doorman — uses a Raspberry Pi Camera to capture a live video feed, along with a Raspberry Pi Zero and a Google AIY HAT for analysis.

This setup passes info on to an Arduino Uno that actuates the eyeballs via a 16-channel servo shield and a number of servos. The device can follow Jessup up, down, left, and right, making for a very creepy robot indeed!

This is a guest post from Surrogate, a team of developers building games that people play in real-life over the internet.

We introduced this concept last year, and have launched 3 games so far. Our final game of 2019 was SumoBots Battle Royale — where players from anywhere in the world can fight real robots on a battle royale style arena. The aim of the project was to have the game run semi-autonomously, meaning that the bots could self-reset in between the games, and the arena could run by itself with no human interaction. This was our most complex project to date, and we wanted to share some parts of the build process in more detail, specifically, how we’ve built these robots and hooked them online for people to control remotely.

Robot selection

We’ve started our process by choosing which robots we’d want to use for the game. There were a couple of requirements for the robots when making the evaluation:

• Are able to withstand 24/7 collision
• Easily modifiable and fixable
• Can rotate on the same spot
• Must have enough space to fit the electronics

After looking at a lot of different consumer robots, maker projects, and competitive fighting bots, we’ve decided to use the JSUMO BB1 robots for this game. We liked the fact that these bots have a metal casing which makes them very durable, all parts are easily replaceable and can be bought separately, and it has 4 independent motors (motor shields included), one for each wheel, which allows it to rotate on the same spot.

We were pretty skeptical of being able to fit all the electronics into the original casing, but we decided to go with this robot anyways, as it had the best overall characteristics. As this robot is easily modifiable, we can always 3D print an extra casing to fit all the parts.

What is the board?

Now that we’ve decided on the robot, it was the time to define what electronics should we use in this build. As usual, it all starts with the requirements. Here’s what we need for the game to run smoothly:

• The robot should be able to recover from any position
• Can stay online while charging
• Supports WiFi network connection and offers reliable connectivity
• Easily programmable and supports OTA updates
• Can control 4 motors simultaneously

Based on these requirements we had the following electronics layout in mind:

We had to find a board that is energy efficient, can send commands to motors, supports parallel charging and has a small footprint on the robot size. With so many requirements, finding the perfect board can be a challenge.

Arduino to the rescue

Fortunately, Arduino was there to help us out. They offer a rich selection of boards to fit every possible robotics project out there and have very detailed documentation for each of the boards. 

More importantly, Arduino is known for its high quality, something that is crucial for semi-autonomous types of applications. Coming from an embedded software background and having to work with all sorts of hardware, we often see that some features or board functionalities are not fully finished which can lead to all sorts of unpleasant situations.

After looking at the Arduino’s collection of boards we quickly found a perfect candidate for our project, the Arduino MKR1000 WiFi. This board fits all of our main requirements for the motor controls, is easily programmable via Arduino IDE, and due to its low power design is extremely power efficient, allowing us to have a lower capacity battery. Additionally, it has a separate WiFi chip onboard, which solely focuses on providing a reliable WiFi connection, something that is very important in our use case.

Now that we’ve decided on the “brain” of our robot, it was time to choose the rest of the components.

Robust hardware means working software

Something to keep in mind is that when working with hardware, you should always try to avoid any possible risks. This means that you should always over-do your minimal hardware requirements where possible. The reason is — if your hardware doesn’t work as intended, your whole software stack becomes unusable too. Always chose reliable hardware components for mission-critical applications.

Some of our electric components might look a bit overkill, but due to the nature of our projects, they are a critical requirement.

Avoiding the battery explosions

As there is a lot of robot collision involved in the game, we decided to go with a high safety standard battery solution. After evaluating multiple options on the market, we decided to go with the RRC2040 from RRC (Germany). It has a capacity of 2950 MaH that allows us to run the robots for up to 5 hours on a single charge. It has an internal circuitry for power management, protection features and it supports SMBUS communications (almost like I2C), and is certified for all of the consumer electronics battery standards. For charging, we used RRC’s charging solution designed specifically for this battery and that offers the possibility to feed power to the application while the battery is being charged.

Note: the Arduino MKR1000 has a pretty neat charging solution on the board itself. You can connect the battery to the board directly as the main power source, and you charge it directly through the MKR1000’s micro USB port. We really wanted to use it to save space and have a more robust design, but due to the large capacity of our battery, we couldn’t use it at full potential. In our future projects with smaller scale robots, we definitely plan to use the board’s internal charging system, as it works perfectly for 700-1800 MaH power packs.

Bot recovery

For the bot to be able to recover from falling on its head, we’ve implemented a flipping servo. We didn’t want to have any risk of not enough torque, so we went with DS3218, which is capable of lifting up to 20KG of weight. Here’s how it works:

Hooking everything together

Now that we’ve decided on all of the crucial elements of this setup, it was time to connect all the elements together. As the first step, we figured what would be the best step way to locate all the pieces within the bot. We then 3D-printed a casing to protect the electronics. With all of the preliminary steps completed, we’ve wired all of the components together and mounted them inside of the casing. Here’s how it looks:

It was really convenient for us that all the pins on the board could be connected just by plugging them in, this avoids a lot of time spent on soldering the cables for 12 robots and more importantly, allowed us to cut out the risk of bad soldering that usually can’t be easily identified.

Arduino = Quick code

Arduino MKR1000 offered us the connectivity we needed for the project. Each sumo robot hosts their own UDP server using MKR1000 WiFi libraries to receive their control commands for a central control PC and broadcasting their battery charge status. The user commands are translated to 3 different PWM signals using Arduino Servo library for the flipping, left and right side motor controllers. The board used has support for hardware PWM output which was useful for us.  Overall we managed to keep the whole Arduino code in a few hundred lines of code due to the availability of Servo and Wifi libraries.

The out of the box ArduinoOTA support for updating the code over the WiFi came in handy during the development phase, but also anytime we update the firmware for multiple robots at the same time. No need to open the covers and attach a USB cable! We created a simple Bash script using the OTA update tool bundled in Arduino IDE to send firmware updates to every robot at the same time.  

To summarize

It’s pretty amazing that we live in the age where you can use a mass market tiny, small form factor board like the Arduino MKR1000 and have so much functionality. We’ve had a great experience developing our SumoBots Battle Royale game using the MKR1000 board. It made the whole process very smooth and streamlined, the documentation was right on point, and we never had to hit a bottleneck where the hardware wouldn’t work as expected.

More importantly, the boards have proven to be very robust throughout the time. These SumoBots have been used for more than 3000 games already, and we haven’t seen a single failure from the Arduino MKR1000. For a game where you literally slam the robots in to each other at a high speed, that’s pretty impressive to say the least.

We look forward to working with Arduino on our future games, and we can’t wait to see what they will be announcing in 2020!

Omni wheels normally contain a number of rollers arranged on their circumference, allowing them to slide left and right and perform various tricks when combined with others. The rollers on UCLA researchers Junjie Shen and Dennis Hong’s OmBURo, however, are quite different in that they are actually powered, enabling a single wheel to accomplish some impressive feats on its own.

These powered rollers give OmBURo the ability to move in both longitudinal and lateral directions simultaneously, balancing as a dual-axis wheeled inverted pendulum. 

Control is accomplished via an Arduino Mega along with an IMU and encoders for its two servo motors —one tasked with driving the wheel backwards and forwards, the second for actuating the rollers laterally via helical gears and a flexible shaft. 

As seen in the video below, the robot can follow different paths via remote control, and even balance on an inclined plane. More informaton on the impressive build is available in the Shen and Hong’s research paper here.

A mobility mechanism for robots to be used in tight spaces shared with people requires it to have a small footprint, to move omnidirectionally, as well as to be highly maneuverable. However, currently there exist few such mobility mechanisms that satisfy all these conditions well. Here we introduce Omnidirectional Balancing Unicycle Robot (OmBURo), a novel unicycle robot with active omnidirectional wheel. The effect is that the unicycle robot can drive in both longitudinal and lateral directions simultaneously. Thus, it can dynamically balance itself based on the principle of dual-axis wheeled inverted pendulum. This letter discloses the early development of this novel unicycle robot involving the overall design, modeling, and control, as well as presents some preliminary results including station keeping and path following. With its very compact structure and agile mobility, it might be the ideal locomotion mechanism for robots to be used in human environments in the future.