Originally aimed at encouraging applicants for computer science at Cambridge, more than 14 million Rasberry Pi units have now sold worldwide.
Two weeks ago, the team behind the innovative microcomputer won the Royal Academy of Engineering MacRobert Award for engineering innovation. As well as helping thousands of people learn to code, the Pi is used for a dazzling array of exciting projects.
Following its award win five years after being launched, we’re looking at some of its best applications, from brewing beer to enabling artificial intelligence (AI) systems.
***
Sorting out a pickle
One day, Makoto Koike had an epiphany. Reading about Google’s AlphaGo AI beating the human world Go champion, he realised the huge potential for machine learning and deep learning to revolutionise work at his parent’s cucumber farm, according to the Raspberry Pi blog.
Raspberry Pi Trading CEO Eben Upton says the “whimsical” gourd-sorting process is one of his favourite Pi applications. Koike’s microcomputer photographs the cucumbers with a webcam, before sending the pictures to TensorFlow on the Google Cloud, which sorts them into nine categories and passes them back through the Pi to a sorting machine.
The system saves Koike’s mother eight hours every day during the peak harvest. It is “a nice demonstration of the intersection of the Internet of Things (IoT), machine learning and agricultural automation,” says Upton to Professional Engineering.
Engineer’s assistant
Tucked away in a quiet corner of the green Hampshire countryside, IBM inventors are hard at work engineering the future. A £35 microcomputer could seem out of place next to expensive kit running the latest high-tech demonstrations, but the Raspberry Pi fits in perfectly.
The computer monitors and transmits flow rate from a water pump, sending the information to IBM’s “cognitive computing” system. The system beams the figures on to an augmented reality (AR) headset, hovering conveniently above the pump in bright green letters. The concept could be used to quickly provide engineers with real-time information during a job.
“Systems like the Raspberry Pi are good for innovating and for use as sensor platforms because they are affordable, adaptable, physically small, and they can run a standard operating system,” says “master inventor” Kevin Brown to PE. “Its affordability helps because it reduces the barrier of entry to innovation, and its adaptability allows us to plug in GPS modules, temperature sensors, accelerometers or any custom-built sensors.”
Anti-squirrel AI
AI is working its way into more and more of our everyday lives, from Facebook chatbots to smart home devices. It has also found its way on to Raspberry Pi computers, like in this project from Microsoft researcher Ofer Dekel.
To protect flowers and encourage birds to visit his garden, Dekel trained a computer vision program to detect squirrels, which eat the flower bulbs and bird seed. He loaded the code on to a Pi, which now triggers the sprinklers every time a bushy-tailed rodent appears.
Dekel and colleagues are now pushing to embed similar software onto tiny microcontrollers, to make AI tools more accessible than ever before.
Remote education
Education was always the Pi’s primary aim. CEO Upton says his “favourite” application of the computer is Rachel Pi. The adapted device allows users in remote parts of the developing world to access curated educational content from online collections, such as Wikipedia, without using the internet. About 10 computers or tablets can use the device simultaneously.
The Pi also helps teach thousands of people how to code, through classes run by the Raspberry Pi Foundation. “We see technical skills - not just coding - as a powerful engine for social mobility for the individual, and economic growth for the societies in which those people live,” says Upton. “The potential impact of educational initiatives can be enormous, as the payoff travels with a child for their entire life. Ultimately, I'll be happy if one former child looks back on Raspberry Pi in the same way that I look back on my BBC Micro.”
Raspberry Pints
Queuing for hours, getting your elbows wet in a sticky puddle of warm cider on the bar, shouting your order over the sound system and then getting the wrong lager anyway – sometimes getting a pint can be more effort than it’s worth, and that’s before the £5+ pricetag.
Any of this could be behind a recent homebrewing boom in recent years. Hobbyists around the world use the Pi to tinker with and improve much of the process, and companies have sprung up to meet specific needs. CraftBeerPi monitors data and allows remote control of equipment, and BrewPi uses Pi computers to enable temperature control during fermentation for a thoroughly high-tech pint.