This week's World Robot Conference 2021 in Beijing, China was an e-Town industry expo that saw over 110 corporate and academic research exhibitors – largely Chinese – fill three halls with more than 500 different B2B and B2C robotic technologies.
The 2021 event revealed a much greater business and industrial focus than previous years’ more consumer-oriented solutions, with many products revealing mature, sector-specific applications, particularly in manufacturing, public service roles, and logistics.
Energy, biomedicine, and aviation were among the other industries represented, while robotics for extreme environments, such as space, offshore engineering, deep mining, and nuclear decommissioning remain development hotspots in the US, UK, and Europe.
In Beijing, technologies on display included the inevitable humanoids and androids; mobile robots for use in factories and warehouses; service robots for public liaison and other tasks; robotic arms and hands; exoskeletons, power suits, and other wearable devices, such as ‘e-knees’, which enhance, strengthen, or protect the body during manual labour; industrial robots; collaborative robots (cobots), which work alongside humans in factories; surgical robots; social care devices; and intelligent logistics solutions.
Most of these developments are being mirrored in other countries, including the US, Japan, the UK, and the world’s most highly automated country, South Korea.
5G, autonomy, artificial intelligence, sensors, computer-vision, end-effectors, additive manufacturing (3D printing), digital twins, data analytics, and cloud-based platforms are among the enabling technologies underpinning robotics, as China automates faster than other nations to maintain its status as a low-cost, highly skilled manufacturing centre.
According to China’s National Bureau of Statistics, industrial robot sales in China increased by over 194% between 2016 and 2020, from 72,000 units to 212,000. Meanwhile, its service robotics companies earned roughly $8.2 billion in 2020 alone, a year-on-year uptick of 41%.
Local technology incubators and accelerators have helped the industry grow within China, as they are doing in the US, UK, Japan, and elsewhere. Meanwhile, the pandemic has spurred numerous sectors to invest in robotics, automation, AI, and other Industry 4.0 technologies, to keep essential services running safely and efficiently.
Aside from the headline-grabbing humanoids, androids, and robot dogs on display in Beijing – photo opportunities which attract less and less attention as public curiosity wanes – more mature and useful concepts are beginning to dominate.
For example, Chinese company Siasun Robotics (the Siasun Robot & Automation Co Ltd) demonstrated a robotic factory floor, an integrated manufacturing facility of cobots, robotic arms, and smart devices working collaboratively to produce goods, linked by a software-based industrial control platform. Countries such as the UK, where manufacturing output has been falling and productivity is low, would do well to emulate some of these concepts.
The China Academy of Space Technology showcased a robot arm to assist Chinese astronauts (taikonauts) in satellite deployment and other tasks on China’s new space station. Robot arms have also been deployed on China’s moon landers and are already in service in the US and international space programmes, of course.
Space robotics is another key UK focus, particularly in Scotland, where Edinburgh’s Heriot-Watt University is one of several robotics centres and hosts the UK’s Robotarium.
Another popular attraction at the Chinese event was Beijing-based Tinavi Medical Technologies, which demonstrated robotic orthopaedic surgery on a model of a human spine. Robotic surgery is a reality in some US and UK hospitals, particularly at the world’s largest medical facility, the Texas Medical Center (TMC).
The humanoid problem
But what about the humanoid machines so beloved of the media?
Aside from their obvious appeal as brand ambassadors, as memes, as the public faces of robotics, or as expensive toys, humanoid robots remain impressive engineering and design achievements that lack convincing applications in most industries.
Even as the likes of Tesla and SpaceX supremo Elon Musk promise a new world of domestic robots via the supposed Teslabot, the utility and affordability of such machines (real-world or investor-baiting vapourware) have yet to be demonstrated, particularly as AI remains a long way from general intelligence.
Even if human-like Artificial General Intelligence (AGI) were to be developed, any sci-fi style humanoids that embody it may never fulfil a useful enough function to justify their cost, development, and rapid obsolescence, especially in a gig economy in which public acceptance of the technology would be a critical factor.
However, an alternative view is that true AGI could only emerge alongside robotics, as advanced AI would need to learn about the world by interacting with it physically, as a human child does.
In the present day, Amazon might be able to push a domestic robot as a loss-leader to spur uptake of a range of own-brand services, content, products, and skills – as it has done with its smart home devices – but the utility of providing these functions in robot form is unclear.
The company has long been rumoured to be developing Vesta, an Alexa-based robot, but this has yet to materialise, despite chatter about it increasing in 2018-19.
As the CEO of a US service robotics company told me in 2019, “No one makes money from selling robots. You make money by selling the service that robots provide.” In his company’s case, that is cleaning the floors of shopping malls.
If the robotics industry itself were a mall, it would be sensible to regard most humanoid machines as the mannequins that keep people window shopping.
An example of the underlying problem is Softbank’s Pepper humanoid in Japan, production of which was suspended some months ago. While the robot is familiar to many alongside the smaller NAO models (originally developed by France’s Aldebaran Robotics, which Softbank acquired), its usefulness in real-world applications has rarely been proved.
Pepper’s friendly physical design is undermined by its limitations in human interaction, relying on apps activated by trigger phrases and a touchscreen interface, both of which rapidly lose people’s interest and engagement. In Japan, robots are a pop-culture phenomenon, yet Pepper still failed to find a big enough market there.
Although it is possible to link Pepper and NAO devices to a natural language service such as IBM Watson in the cloud, that concept has never caught on commercially.
Another example is Boston Dynamics, makers of the Atlas and Spot robots that are familiar to millions from videos and memes, alongside its Stretch and Pick industrial solutions. The company has changed hands several times, perhaps because no one is clear what many of its technologies – impressive though they are – are actually for.
Aside from some specific applications, such as working alongside astronauts on Mars – which NASA’s Valkyrie robot is being developed to do – or collaborating with humans in other hazardous environments, it is hard to envisage the utility of most humanoid devices. Beyond a slow, evolutionary fusing of man and machine.
An alternative view would be to regard exoskeletons and other technologies that enhance our physical capabilities as a Trojan horse for humanoid robots. At some point, the question arises: why not just get a machine to do x, rather than make humans more like machines to do it? At that point Atlas – named after the mythical titan who carried the Earth on his shoulders – begins to make more sense.
Robots that can run across challenging terrain, jump, and somersault may grab public attention and alarm apocalyptic thinkers, but most teenage parkour fans, tricking street gymnasts, free climbers, and skaters can perform far more dazzling and impressive feats than that. Where are the jobs for them? As the China event proved this month, robotics needs to stay focused, real, practical, and fulfil real human needs.