Cloud Derby

This hands on acitivity is designed to help you learn Google Cloud Platform (GCP) in a fun way. You will build self driving robot that plays ball game against other robots. Robots are built on Dexter Industries’ GoPiGo car. We provide all the instructions to build hardware and software along with project source code and annotated images in GitHub, but you are free to experiment and add or change anything you like. You will be using various GCP services:

Cloud Derby Workshop is a one-day hands-on training session where participants, grouped in teams, will assemble and connect a Cloud car to Google Cloud. Once connected, the car will be integrated as an IoT device sending images and sensor information and receiving commands from Google Cloud service. The control logic is based on Google Machine Learning using pre-trained vision models. The teams will compete by modifying the logic to optimize colored ball ‘round up’. Pre-built artifacts and assembly instructions will be provided, along with the Github code for self-driving control logic. Winning team awarded a special prize!

Duration: Total time ~ 5 hours

• Overview, team building, setup: 0.5 Hours
• Robot Building using instructions / onsite help: 3.5 Hours
• Tournament: 1 Hour

Attendees will bring

• Bring their own laptops to the session
• Must have basic technical knowledge for programming and shell scripting
• Some knowledge of GCP services will be beneficial, but not required for all team members

DevFestDC + Google Cloud Engineering Team will provide

• Robot hardware kits + assembly instructions
  Source code in Git repo
• Instructions and coaching for setting up GCP project, robot software and connections to the cloud
• Wifi connectivity and venue for the derby
  GCP account

Google Services Used and Skills learned

• IoT Core, TensorFlow, Pub/Sub, Cloud Storage, IAM, Compute Engine, AppEngine, Cloud Functions
• How end-to-end IoT services operate from device to logic control and analysis
• How ML services are integrated as a key part of IoT design and control

Cloud Derby Agenda

1:30 PM     –    Overview, Team Building and Setup
2:00 PM    –     Project Creation and Setup
2:30 PM     –    Machine Learning (Image Preparation and Training)
3:45 PM     –    Machine Learning Inference
4:30 PM     –    Car Assembly and IoT Core Setup
5:00 PM     –    Voice Control with Google Assistant and Dialogflow
6:00 PM     –    Cloud Derby Tournament
7:00 PM      –   Winners Announced and Closing