So I’ve been pretty excited about going to Hackathons. They are fun, you get to met new people, try different things, and have some fun. For those that don’t know, a Hackathon is sort of a contest between software folks to address a set of challenges to come up with something unique. Its an idea generation contest. It also provides a chance to learn what skills you don’t know, what you need to brush up on.

The biggest Hackathon of the year is the AT&T Hackathon that takes place in Las Vegas the weekend before the annual Consumer Electronics Show (CES). This year there were quite a few sponsors and a whole lot of prizes. After looking at the sponsor list, I came up with an early plan. I’ve been playing around with different tutorials on OpenCV (Open Source Computer Vision) and thought that would be a fun thing to play around with.

Looking at the challenges on the Hackathon list, I saw a challenge there for “It Can Wait”… an idea to reduce distracted driving. I also saw a challenge for best use of Intel Edison board. That got me to do a bit more research.

I discovered the Intel Garage videos, where they hook up a web camera to an Intel Edison, and are able to use OpenCV on that platform. I was able to reproduce these demos with an old USB webcam I had hanging around.

  • http://www.intel.com/content/www/us/en/do-it-yourself/garage-content/garage-facial-recognition-video.html
  • http://www.intel.com/content/www/us/en/do-it-yourself/garage-content/garage-webcam-hook-up-video.html
  • http://www.intel.com/content/www/us/en/do-it-yourself/garage-content/garage-streaming-web-video.html
  • http://www.intel.com/content/www/us/en/do-it-yourself/garage-content/garage-audio-voice-recognition-video.html

Awesome. At first glance it should be possible to add OpenCV to Intel Edison. Time for more research. I did discover a Github submission that uses OpenCV to track pupil gaze. Its not the entire thing, but its a very decent start towards making a proposal for a box that sits on the dashboard of a car, looks at the driver and detects when they are looking down at their mobile phone. Awesome idea. Form up a team before hand, make sure we have all the necessary cables and bits and we should be ready to develop something.

So we go to the Hackathon. One of our team members gets way sick. He’s clearly the smartest of all of us, but oh well, lets move on anyway. So we get to the hackathon, and we’re talking to a sponsor. He asks us what our plan is for the event. We tell him. He tells us, wait a second, wasn’t that the grand prize winner for 2015? He tells us, wait a minute, here is the Youtube video of the project. Oops. In fact, big oops. We quickly realize, there is no way we can repeat something that has been done before. Ouch.

So what next? We’re sitting around the table, moping about the issue. We’re visiting some of the sponsors, looking for good swag. One of the items we picked up for review was a Freedom ARM processor board from the folks at NXP / Freescale, a FRDMK64F . Its an Arduino style board with decent processor and memory, and the kit includes an accelerometer, magnetometer and gyroscope input card. Very cool. We start playing around with the board, as well as some items from other sponsors. We add a Ublox GPS sensor and an Konekt Dash (a digital cell phone / sim card transmit device) to the mix.

ShockWatchParts

And we think we have a feasible project. A guy I met in line decides to join us. Later two international students ask if they can join our team. Sounds great. We form a five person pick up team and by get started on the project around 5PM on Saturday. Its late but what do we have to lose. We break the project down to little steps, divide up the work and move it along. Most of the coding was done in C using the Mbed online IDE tool.

We’re making a Shock Watch device. Its a small battery operated box that detects GPS locations and monitors acceleration loads (shock G loads) Its placed in the back of an 18 wheeler truck to determine maximum loads placed on products being shipped. I’ve worked with plastics before, and heavy impact loads, particularly in cold weather, just don’t play well together. Think of an entire load of large screen TV’s being shipped across the midwest in January or February in sub-zero temperatures. Cold plastic doesn’t take impacts well at all. Our shock watch device monitors the G-forces, sends out a digital message to the cloud every five minutes. You can determine the location of rough road areas via GPS location. You can even monitor the location of the truck during transport. All data goes thru the Konekt server with the intent of storage at AT&T M2X data storage in the cloud.

Click here for presentation slides.

Its not our original idea, nor is it our original team. What makes this way cool is this project wins a $5000 kicker prize from a sponsor for unique use of their product, the Konekt Dash.

Lessons learned:

  • Its okay to abandon ideas to try something else out when you hit a road block.
  • Never give up.
  • Get organized.
  • Everybody has something to offer, no matter what their background. The two international students didn’t do a bit of coding. In hindsight, their input was instrumental in the teams getting to a prize. They kept us on track ensuring that critical team management events were done on time. They also created a wonderful Powerpoint slide show describing the project, and did so without prompting. That presentation really made our idea stand out.

TeamShockWatchAwesome

Thanks and Congrats to the “Shock Watch Awesome” team (L->R):
LB Corney, Las Vegas
Dirk Schmidhofer, Las Vegas
Daniel Kunkel, Seattle Area
Kanat Mustafin, International Student from Kazakhstan
Poy Yeung, International Student from China