The document discusses how to build enterprise IoT projects iteratively using an example water pipe monitoring system. It describes how the system started as a proof of concept at a hackathon using basic sensors (Phase 1) and was iteratively improved by adding a web interface (Phase 2) and scaling to monitor multiple pipes (Phase 3). The key lessons are to start small, prove concepts work, and scale features over time while maintaining scope and deadlines. Iterative development allows projects to grow from initial prototypes to full enterprise solutions.
2. Talk Overview
How to start doing IoT
How to grow
Use Case Example: Save the Water Pipes
Very hackathon-centric, but some of the principles can be applied to commercial
applications as well
Always ask: who is my customer and what is the immediate need? Where and when
it’s possible: how can I design this so there’s room to grow while still meeting the
immediate need and deadline?
Won’t go too deep technically, describe the process of improvement
3. In the Beginning
April 2016, the “Blue Team” enters the Intel IoT Roadshow in Denver
We win first place with our freezing water pipe detection and
prevention system
Detects when water in a pipe is about to freeze
When temp gets into “red” zone, stepper motor opens a valve to
release water
When temp gets into “yellow” or safe zone, stepper motor closes the
valve
5. Technologies Used
Intel Edison board
Intel XDK: IoT Edition
Javascript-based IDE to build IoT applications
Intel libraries: Javascript wrappedArduino
Grove shield and Grove sensors
6. What Was Missing?
Notice there is no pipe, water temp sensor is placed directly into
bucket to simulate “freezing water temp in pipe”
Water temp sensor is taken out of the bucket and placed between our
fingers to “warm up the water temperature”
Another trick: water temp sensor measures voltage, not temperature.
We had to adjust our algorithms to get the results that we want
7. What Was Missing? (cont.)
All sensor data displayed to console and LCD
There is no website, mobile app, any software other than
JavaScript-wrappedArduino code (Intel hardware libraries)
Used sample sensor code within the IDE to get started
8. Why Did We Win?
Strength and simplicity of our idea
we focused on 1 problem and demonstrated its solution well
Strong hardware and sensor demo
We told a compelling story during our presentation
We tried not to do too much at once - only had 2 days to work on project start to
finish & first time working with Edison board and Intel IDE and technologies
The point of hackathons is to show a concept is possible, NOT to build a
market-ready product. We saw some teams fail because they lose sight of this
9. China US Young Maker Competition
Fresh from our success at the Roadshow, we entered the
China-US Young Maker Competition on hackster.io, also
sponsored by Intel
A couple of weeks after the Roadshow, we got back to work
We only had a handful of weekends to work together as a
team
Submission deadline: mid-June 2016
11. What Was Added
Used an Intel sample application to get started:
https://software.intel.com/en-us/creating-an-application-to-
communicate-using-web-sockets
Challenge: how to get the data “out” from the IDE console and into
a useful application
Asked: what did we want our “useful application” to do?
Local web server hosted on Intel Edison board used to communicate
between sensors and display results on website
13. How we worked together
Version Control: Git and Bitbucket
Project Management: Trello
Schedule in person team meetings and used communication tools as needed:
Hipchat
I filled the role of software lead and architect
Selected and largely modified the Intel sample application for the
foundation of our project
As team lead, ensured that we completed our project in time for submission by
the deadline
14. Selected To Go To China
Mid-August 2016, the “Blue Team” went to Beijing China to
compete in final rounds of competition
15. Project Enhancements
Simulate monitoring water pipes for multi-unit system instead of single unit
Solenoids replace stepper motors to power valves
Relay Shield added to power multiple solenoid valves
16. Project Enhancements
Website has additional page to show status of all units in real
time, plus existing page to show status of single unit in real time
Requires refactoring of JavaScript code in back end (used JS
objects)
17. Competition Results
Successful in implementing improvements to system
Placed 11th out of 64 teams overall
Great experience!
To learn more:
https://sunfishempire.wordpress.com/2016/08/27/
save-the-water-pipes-project-wins-excellence-
award-in-iot-china-contest/
18. Conclusions
We didn’t get to where we are today overnight, we did it
in steps. You should too!
Phase 1: Start with controlling sensors and displaying
sensor data to console
Phase 2: Use sample application to start building real
software application that displays data outside of IDE
Phase 3: Refactor code as needed to scale system
19. Conclusions Part II
DO Start Small, and add one feature at a time
DON’T Think you must begin with an Enterprise-level IoT
system right off the bat
DO Use version control, project management, etc., as soon as
feasible
DON’T Stay in “Hack-a-Thon Mode” forever!
DO Keep project requirements and scope in mind
DON’T Spend time you don’t have on features you don’t need