The Raspberry Pi family of computers had another stellar year in 2017, with more than 17 million boards sold and the $10 Raspberry Pi Zero W released.

In its annual review for the year, the charitable foundation behind the tiny board, the Raspberry Pi Foundation, highlighted successes ranging from its educational outreach helping 150,000 children to 8.5 million people learning via its online projects.

SEE: Raspberry Pi 3 Model B+: An insider’s guide (free PDF) (TechRepublic)

But beyond these admirable achievements, the Pi boards themselves are perhaps most famous for their versatility, and the report also celebrates the makers who have cooked up the most inspirational Pi-powered creations during the past year, ranging from the sublime to the ridiculous. Here’s the foundation’s choice:

Stent-testing robot

Medical engineering is just one of the important areas where researchers have found a use for the Pi.

This stent-testing robot was designed by Dr Henry J Feldman, chief information architect at Harvard Medical Faculty Physicians.

The stents in question are rubber tubes that are used to keep open patients’ airways. New stent designs have to be tested 300000 times to be certified for use, and Feldman developed a system that uses a computer vision and machine learning to identify when a stent fails during testing. It works by running the image-recognition library OpenCV on the Pi, which analyses footage of the test captured by the Pi’s camera module. The Pi is augmented by a HAT [attached hardware board], which controls the gripper crushing the stent during tests.

The system provides data and images that identify the point of failure more precisely, which in turn allows designers to produce a more resilient stent.

Read more about the project here.

C-Turtle

This inspirational use of the low-cost computer could one day literally save lives.

Traditionally clearing landmines has required expensive robots, due to the need for the bots to survive multiple blasts.

Scientists at Arizona State University used the Pi to create a very different kind of bot. The C-Turtle is robotics on a shoestring, with a shell and turtle flippers made out of cardboard and running on a $5 Raspberry Pi Zero. Despite its low cost the bot is adaptable, using machine learning to teach itself how to push itself forward across different terrain on its flippers.

Costing just $70, the bots are cheap enough to be considered disposable, and the researchers hope that fleets of the bots could eventually be used to detect and map landmines.

Read more about the project here.

BitBarista

BitBarista is an autonomous coffee machine, designed to function with the minimum of human supervision.

The Delonghi coffee machine has been augmented with a Raspberry Pi, allowing it to sell coffee in return for bitcoin, and to pay punters bitcoin for basic maintenance, such as filling up its water supply or restocking it with coffee beans.

Built by a team at the University of Edinburgh’s Centre for Design Informatics, this automation project also helps spread social awareness by asking customers to order new coffee bean supplies, based on the information it shows them about quality, price, and the environmental and social impact of ordering different beans.

The attached Pi enhances the machine in various ways, allowing it to read signals from coffee machines sensors, controls its functions, and connect to the Internet. It is also fitted with a small camera to read smart codes, so it can make Bitcoin payments directly to customers’ Bitcoin wallets.

Read more about the project here.

Robocod

Of course, not every Pi-powered project is serious, some are downright silly, while still being technically quite impressive.

A case in point is Robocod, a project which uses the Pi to create the world’s first vehicle driven by a fish.

The movement is controlled by the fish, computer vision tracks its location and as it swims around, the Pi moves the tank in the corresponding direction.

If you want to give your goldfish a taste of the open road, check out the code repository for the project on GitHub, and find out more about the endeavor here.

Read more about the Raspberry Pi 3 Model B+