Thanksgiving Update - 11/30/15

Short update here - it's the holiday and we're gonna enjoy it!

Overview

A good week, overall. We attended the kickoff of Launch:Alaska, which is Alaska's first startup accelerator. Their goal is to tap into Alaskan resilience and self-sufficiency to solve big problems in the Arctic. We also had a chance to present to the group at the after-party and share what we're working on.

We're working hard developing our initial service offerings, which we outlined in the previous update. We're going to highlight one each update email and explain the work we are doing around it, provide an example of the work we can do, and explain our target markets in the state. Opportunities for hardware and differentiation will also be explored.

However, before we can really dive into that we need to figure out how to operate in this weather. The last 10 days has brought sub-zero temperatures, then 2 feet of snow, and now cold rain. Besides becoming a trying winter, it is also put a huge damper on us flying a non-waterproof drone that also isn't supposed to operate under 32 degrees F.

$150,000 vs $1,500

Why are we going to all this effort? Why not just purchase a rugged drone? Well, we don't have the money - but it also isn't a sound investment when it comes to the quadcopters on the market. There is a huge gap in the world of multi-rotor drones: those below $10,000 and those above $50,000. Very little hardware of note exists in the middle.

So we have set out to make a $1,500 drone fly like a $150,000 drone. Check out the comparison I put together below. I think you'll agree that there isn't a $149,000 amount of effort required to match efficacy:

Winter/Water Proofing

The DJI Phantom 3 has many vents and a very delicate camera hanging off the bottom.

We are going to take the scrappy approach here. As the SPX folks on this email know, there is a fine line between scrappy and crappy. However, in order to minimize weight and maximize efficacy, I think we can get away with shielding the brushless motors, blocking the vents up, and potting the electronics inside the craft in a resin. As for the camera, I will cover it with a superhydrophobic coating and bag it in a waterproof enclosure.

As far as winterizing goes, it will be a similar plan. Blocking the vents means the generated heat stays inside the craft. I plan to do this with a lightweight fleece or fabric. Yes, like a coat. Rather than a plastic bag around the camera, I am thinking that a fabric bag that pulls on will work. I also think that a couple handwarmers placed in the bag will keep the camera above 32 degrees.

I have an Arduino datalogger that I will use to test this ideas as soon as the temp drops again.

On a scale of scrappy to crappy, how does this idea sound? I think it has merit and can look sharp if done right. Most importantly, it's cheap and lightweight.

The Spotlight

Photogrammetry sucks in the winter because everything is uniform and white. Luckily one of our close friends thinks we can hack together a functional LiDAR system on the cheap. This, along with drone winterization, are our hardware focuses. Hopefully we'll have good news and data to share in the next couple months on this front.

Got this picture during mapping of the Swanson River Bridge. It came out a lot nicer than my first attempt at the 3D model:

Got this picture during mapping of the Swanson River Bridge. It came out a lot nicer than my first attempt at the 3D model:

You can see that there are two sets of guardrails on the right side of the image, and a weird growth jutting out of the middle of the road. Practice will fix these issues, but just goes to show that it will indeed take time.

You can see that there are two sets of guardrails on the right side of the image, and a weird growth jutting out of the middle of the road. Practice will fix these issues, but just goes to show that it will indeed take time.

Challenges and Questions

Challenges: Photogrammetry is not trivial! This is good for business, but sucks for time investment. We will keep working to improve our methods and accuracy through the winter.

Our next update will include a deeper dive with lots of examples so you can see our work thus far in this area!

Questions: None this time - will likely have some once we get data back on winterization methods.

Again - feel free to reach out and share your thoughts or say hello!

Ben