From Planning to Powering Up
(Part 1 of a two-part series. Part 2 can be found here).
On May 23, 2023, ACN held its first internal robotics hackathon. The day was themed around our beloved city of Montreal and focused on Viam Robotics’ products and platform.
Participants hailed from diverse development backgrounds, most with little to no prior robotics experience. As our team boasts machine learning and MLOps experts, we, of course, tried Viam’s machine learning modules, alongside other challenges. Our goal was to familiarize ourselves with the Viam platform and its possibilities while documenting any hurdles encountered. It was a refreshing departure from our usual work, which rarely involves tinkering with hardware components.
This post is about how we prepared for our hackathon.
Some of our ideas were:
- 🏎️ Navigating the rover from point A to point B with precision: We proposed a modification of this tutorial themed around the Montreal metro’s iconic “green line”, introducing “stops” along the line and incorporating the metro’s distinct audio signal.
- 👁️ Testing our image recognition skills by distinguishing between a sesame bagel and a poppy bagel: we hoped to use Viam’s ML service, which allows users to upload Tensorflow lite models, then use the vision service for detection, classification and segmentation.
- 🦾 Using the robotic arm to delicately grab a bagel without squishing it: Since we lacked the necessary time/components to get our mechanical arms fully working, we proposed that participants could help add an arm component to be controlled via the SDK.
Setup
- We acquired a total of five raspberry pi’s.
- Two were each mounted to a Viam rover base.
- A third was associated with a Yahboom DOFBOT arm, which is supported by Viam.
- The fourth was associated with a smaller arm made by Adeept. Given that that arm is not supported by Viam, it would involve manual configuration.
- Unfortunately, we did not have time to get this arm fully functional before the hackathon, so we offered participants the option to ssh into it and programmatically test all the servos and see if they could get them working.
- We hoped to use the fifth pi as an office assistant.
- We acquired various components to mount on this pi, to test how well they would work with the Viam platform. These included a webcam and a temperature/humidity sensor.
- We configured a private wifi network so we could access each raspberry pi device by a static IP, and installed viam-server to each raspi.
- We also had on hand a network switch, some ethernet cables, and a grab bag of assorted tools and random electronic parts, as well as extra microSD cards in case we ran out of space with video capture.
- We made sure all our batteries were charged ⚡!
- We provided participants with appropriate documentation.
- We took the rovers on a test drive using the web interface to make sure everything was connected properly, much to the concern of Rhubarb the office dog.
To learn more about how your organization could incorporate robotics and machine learning, reach out to see how we can help!
Thanks to Alexandre Rimthong and Florencia Herra Vega for the initiative in planning and making the event possible, and Becks Simpson and Mona Ghassemi for helping with setup.
