Raspberry Pi and Arduino powered robot lets users make art from wine

Combining cloud computing and Internet of Things hardware, the Splash Machine allows users to create their own wine art online, streamed live and photographed for posterity.

Raspberry Pi and Arduino powered robot lets users make art from wine Swirl Machine

Spilling wine is typically something to be avoided, though the shapes made by wine spills inspired a team of engineers at marketing firm KPS3 to build the Splash Machine. This is a robot that allows users to make wine art online using a web-controlled, live streamed rig powered by a Raspberry Pi 3B+ and Raspberry Pi 4, alongside an Arduino Micro, and the mechanical devices needed to move all the associated parts.

The website for the Splash Machine website is a single-page React.js app stored in S3, served via CloudFront, with the platform running on a MySQL and Redis database, with a Node.js/Koa REST API server that communicates with an API worker that manages the queue and communicates with the machine through gRPC. A Lambda media worker is used for image processing, and Mux.com is used for live streaming the video captured on the Raspberry Pi 3B+.

SEE: The best alternatives to the Raspberry Pi (free PDF) (TechRepublic)

Visitors to the website can watch the current splash art being made, or register to join the queue to use the service—once their turn arrives, they are given a full-screen view and options to control their creation. Specialized watercolor paper is picked up by a servo equipped with a suction cup attached to a venturi (itself attached to an air compressor), which proceeds to hand off to a custom gripper to hold the paper in place for the wine to be spilled. 

From there, the Arduino-controlled stepper motor turns the lead screw, moving the paper into the splash zone, and fills the glass to the user-selected level, and begins to swirl the glass at the user-selected speed. Once the proper speed is attained, an actuator tips the glass forward, spilling the wine onto the paper. All of this is streamed by two cameras, with the final result photographed by a third. This photo is then processed in Lambda to clean up the photograph, detecting corners, applying filters and branding for the Santa Maria Valley Chamber of Commerce.

"The idea came on a car ride to Santa Maria Valley," said Rob Gaedtke, KPS3 president and CEO. "We were playing around with the idea of capturing slow motion swirls, that led to how interesting wine spills are, and that led to the machine. The fact that it was a spill fit perfectly into the vibe of their wine experience. If you spill a little, no one really cares. And because of who we are as a company… it clearly had to push our technical boundaries."

Building the Splash Machine took "about a year and 1,500 hours," Gaedtke said, with three prototypes that were controlled mechanically before the online component was added. The entire build cost approximately $2,000 with plastic and wood components costing more than the technology, Gaedtke said.

The company decided to utilize Amazon Web Services—as opposed to Microsoft Azure or Google Cloud Platform—due to prior experience with the platform. "We've used their services like EC2, ELB, CloudFront, and S3 on other projects and have had good experiences. We also had the opportunity with this to try out some of their other services that we've wanted to mess around with but haven't like Lambda functions and Elastic Beanstalk."

For more, check out "21 of our favorite Raspberry Pi projects" and "Oracle: This 1,060 Raspberry Pi supercomputer is 'world's largest Pi cluster'" at ZDNet.

Also see

vlcsnap-2019-09-20-10h58m30s759.png

Image: KPS3