Welcome to the Bulletin by Remix Robotics, where we share a summary of the week's need-to-know robotics and automation news.

In today's email -

• Robot driver crashes NASCAR

• Big plans in the UK - a Drone Super Highway and our very own DARPA

• A robot arm is being sent to Mars

• Apple's growing robotic IP portfolio

• One useful tip for jobseekers

• Four interesting investments

This is our 10th Bulletin - Thanks for reading and if you're finding the content interesting please consider sharing it with someone who will too.

Snippets

Safety First -  NASCAR has implemented a robot driver to perform high-speed crashes in their next-generation car. “The challenge was getting this extremely complex machine to do a very precise test without a human driver piloting the car”. Luckily a robot was more than happy to drive straight into a wall at 130 mph.

Drone Super Highway - The UK plans to build a 165-mile drone highway connecting Oxford, Cambridge, Rugby and Milton Keynes. A network of sensors will provide a real-time moving map of the airspace and guide drones to their destinations. Expect UK drone delivery to become a reality very soon.

The UK's Moonshot Agency - The newly formed Advanced Research and Invention Agency (ARIA) has announced its first CEO, Ilan Gur and Chairman, Matt Clifford. ARIA is inspired by the US's DARPA and will use its £800 million R&D fund to invest in high-risk, high-reward science and technology. ARIA will have - "a much higher tolerance for failure than is normal, recognising that in research, the freedom to fail is often also the freedom to succeed.”

Attack of the Martian Robots - The European Space Agency has announced that they are sending a robot arm to Mars. The 2.5m arm will retrieve sample tubes from NASA's Perseverance rover and transfer them to a rocket destined for Earth. Due to the 14min delay between Earth and Mars, the system will be highly autonomous. Our resident space expert says the challenge will really start when samples re-enter the Earth's atmosphere at 20 km/s, double the speed of samples from the Moon. It's never been done before and he has a PHD so we'll trust him.

Apple Controls Everything - We've spoken at length about Apple's self-driving vehicle patents and their phone disassembly robots. Now, they're patenting a general controller for autonomous boats, aircrafts, drones, space vehicles, buses, and semi-trailer trucks. This may just be following their normal approach of patenting every single idea they come up with or it could indicate an even greater focus on robotics in the future.

Emotion = “disappointed” - Google's GoEmotion dataset came under criticism when Surge found that 30% of the data points had been mislabelled. So what? The dataset was meant to “understand context and emotion”, which could be used to “detect harmful online behaviour”. Errors here could lead to bias, errors and all manner of unethical behaviour. This blog presents a great case study on how not to build a dataset for AI, and why deep learning approaches are only as good as the data, they're given. Also, the data came from Reddit comments and the mistakes are pretty comical.

Top Tip for Job Seekers - This amazing DIY submarine has been floating around the internet. It's a fun project but students/job seekers should seek inspiration. We can speak from experience - If you're looking for a job in robotics and have something like this in your portfolio, we can guarantee it will go a very long way. That aside, it's a great Deep Dive into how submarines work!

The Big Idea

How China grew its Global Share of Robots by 100X

This week we were introduced to Rian Whitton the writer of a report on China's robotics strategy titled - "China Tests Our Dream of an Automated Future". It's incredibly well-researched and I recommend reading the whole piece at the Bismarck Brief. We'll review some of the key ideas in this week's Big Idea.

"From 2000 to 2020, China's worldwide share of new industrial robot installations grew from 0.4% to 44%.”

This ridiculous 100X growth didn't happen organically. The Chinese government engineered it in a concerted effort to cement themselves as the "World's preeminent manufacturing power for the foreseeable future". More than this, they hoped robotics would solve one of their biggest challenges  - declining growth rates. As China's middle class grows and its population ages, the Ministry of Education predicts a shortfall of 30 million manufacturing workers by 2025.

The article argues that you can have economic growth, low birth rates or low immigration but not all three. As birth rates decline and China remains reticent to increase immigration, the Party wants to have its cake and eat it. Using robotics to deal with the inevitable reduction in productivity.

What was China's strategy?

To achieve their goals, they had a simple strategy -

1. Increase the use of robotics in manufacturing

2. Increase the use of service robots in other industries, including agricultural, hospitality, hospitals and the home

3. Build a homegrown robotics industry to meet this need independently

4. Achieve this at unprecedented speeds using unprecedented subsidies

To achieve this, the government pledged to spend $672 billion on industrial policy goals, earmarked $135 billion for “machine substitution” of human labour in Larger Guangdong province alone and set out to build 40 robotic innovation parks across the country.

Was it successful?

Yes and No. China became the world's single biggest market for robotics and has maintained a significant trade deficit with the West, $355 billion with the US in 2021. That said, Growth rates have been declining and COVID seems to have impacted China far worse than the US.

Their foray into service robots has not gone quite as well. The technologies are less mature, the working environments are less controlled than manufacturing, making it very challenging to find solutions that bring real value. The amount of funding being poured into the space has mainly resulted in disappointment. Many companies are at the cutting edge of what's possible and the industry has a long way to go before it matures. As a result, current technologies are not going to solve China's labour shortages challenges in non-manufacturing sectors in the short term.

Concerning gaining robot independence, the results are also mixed. 100s of new robotics companies have launched in China since 2016 and became the world leader in publishing robotics patents. Unfortunately, these products and patents have not translated to real-world success. China's largest industrial robot brands have failed to secure large contracts and gain much traction locally or abroad. As a result, China's homegrown industry is still heavily dependent on government subsidies for survival.

To make up for this, China has tried to import expertise from Germany and Japan by purchasing or investing in foreign robotics companies (the Chinese Midea Group purchased Kuka in 2020). This isn't a sustainable long-term solution so China must focus on developing the education of its workforce. Becoming an automation superpower requires a huge pool of technically skilled engineers and China is currently behind other advanced manufacturing nations.

In summary

"China can lead in the narrow field of industrial robots and supplant countries like Germany and Japan in high-end manufacturing."]

and

“China can complete its transition from a consumer of industrial robots to a world-leading producer.”

But

“Automating all major sectors of the economy with service robots and avoiding a hard demographic landing, is a much greater challenge”

What's our take on this? Part of our team was based in Shenzhen for a year and saw this growth first-hand. China's results have been mixed but we would argue that there is a lot to be learnt on how to become a manufacturing superpower -

• Education is key. This is constantly being repeated but technology is not the solution. Trained engineers, apprentices, trades-people AND technology, are the solution

• Governments need to support manufacturers with implementing robotics and automation. It doesn't need 100 billion in subsidies but if we want to compete, we need to pay for it.

• Innovation needs to be fostered in a balanced way. Funding should be available but it needs to be discerning and competitive. Constraints can be good and limitless cash leads to grifts.

Interesting Investments

Intelligent City raised $22M -The housing technology company applies automation and robotics to the design and manufacture of mass timber buildings. (Founded 2008, $30M raised to date)

Syrius Robotics raised $7.4M - The Byte-dance (they own TickTok) backed Chinese start-up raised funds in a Series B round to expand its production of autonomous mobile robots for warehouses (Founded 2018, $40M raised to date)

ForSight Robotics raised $55M - Pioneers in ophthalmic robotics surgery, the Israeli-based company raised a Series A round to advance the world's first surgical platform for fully robotic cataract surgery (Founded 2020, $65M raised to date)

Chipotle invests in Hyphen (undisclosed) - Hyphen produces an automated system for preparing and fulfilling meals in the kitchen. The well-known fast food giant has launched a $50M venture labelled Cultivate Next, which makes early-stage investments into “strategically aligned companies”. (Founded 2015, $22M raised in the previous round)

Video

The story of Tesla's Gigapress

This video explains how and why Tesla casts the Model Y chassis in a vast six thousand-ton press. Casting has been a standard process for engine blocks and other automotive components but it has never been used for anything so large and structural.

It was impressive to see the size of the challenge in developing such a unique machine but the motivation for doing so is almost more interesting. Even though each step of the old process is individually quite simple, the sheer number of steps makes it very complex. The Gigapress converts 70 small parts into a single part and reduces the need for around 150 robots. This is very consistent with Tesla's 5 step design process.

Tweet of the Week

P.S - This week Remix presented at the TLA London Robotics networking event. We recommended it highly  - check out their next event in Nov!