Drones aren’t grounded – decisions on the fly with Watson

By | 4 minute read | November 11, 2016

leadspace image of a drone inspection

Drones have really fired our imaginations in the last couple of years. When Amazon announced it was investigating using them for home delivery we all started to envisage a future with thousands of them zipping around us all the time carrying books, clothes, tea bags and all the other stuff we buy online. And as costs have come down they’ve become sought after Christmas gifts for kids and grown-ups.

Struggling to get drones off the ground

But along with all this has come fear, uncertainty and government intervention. We hear about amateur drone users flying them near airports, and the path to drone based home delivery seems a highly uncertain one. This week there was press about Alphabet Corps struggles in getting its drone business off the ground, with the story centering on disagreements between engineering and commercial teams over-using drones to deliver food to students on campus – a project dismissed as a stunt by the engineers who wanted to focus more on reliability and safety.

A valuable, commercial case for drones

Amid all this we’ve started partnering with a small company based in the Netherlands who have found real, valuable commercial use cases for drones. And what’s more Watson IoT is key to their success. Intrigued, I spent time with Jan Wouter Kruyt, VP of Strategic Partnerships at Aerialtronics to find out more.

The origin of Aerialtronics is pretty cool. Founder, Robin van de Putte, was doing promotional video shoots of luxury cars from helicopters. The financial crisis made the James Bond approach unaffordable, so Robin started experimenting using drones and got pretty good at it. Originally Aerialtronics bought all their drones in parts but the technology proved unreliable, so they started an engineering department. Today they’re on the third generation of their product.

In the 4.5 years since they got started, Jan told me, regulation and technology has changed a lot. Now, he says, it really is viable to operate drones to generate commercial value. He explained that the regulatory environment is complex but generally helpful. Every country has different air worthiness authorities, meaning there’s a myriad of different standards and rules to adhere to. At the same time, regulation is taking away a lot of the concerns around how to operate drones. Major corporations or government departments are generally in the business of complying with regulations so having known rules, even lots of them, has been good in giving businesses the confidence to invest and bring projects to life.

How are people using drones today?

Aware that my laundry detergent isn’t yet arriving by drone I was most interested in the use cases. Where are people actually using drones today and seeing success? It turns out there are three main areas.

First is public safety: creating better situational awareness for decision makers responding to disasters and emergencies.  Think police or fire departments or any kind of first-responder dealing with a complex, unfolding situation. Having a live, overhead view, and coupling that with views from body cameras, enables better decision making. Interestingly, drones can also be used as flying cell phone towers, bringing connectivity back into affected areas much faster than the portable towers used today in disaster relief.

Second is mapping and surveying, particularly to assess the quantity and value of raw materials in mining and agriculture. Drones can help measure the volume in stockpiles, the value of timber in a forest or the likely crop to be harvested from a tobacco field. Drones with infra-red cameras are being used to map crop health in agriculture.

Third is industrial inspection, and this is where Watson IoT comes in. This use case is very data intensive and dependent not so much on the real time view but deep post-flight analytics. Drones are being used to inspect oil rigs, power lines, phone masts, wind turbines and bridges. Traditionally companies have used cherry pickers, binoculars or had their employees climb masts after taking rope access training courses. It is time consuming, dangerous work, and the results are inconsistent. Drones are quicker and safer. But each flight can create gigabytes of photos and video, and the people analyzing it post-flight are struggling with data overload. Enter Watson IoT!

Aerialtronics has connected their drones to the Watson IoT platform. As Jan says: ‘we’re able to get very advanced analytics on the data in an automated way.’ With the IoT platform they can ingest and organize the data, and then call Watson’s visual recognition capabilities. This helps makes sense of the data they’ve captured. Specifically, they’ve trained Watson to look for corrosion, serial numbers, loose cables, and misaligned antennas. It automates the process of finding faults and defects, and it came together fast. Working together we were able to train Watson to look for many of these patterns in a few weeks.

In all the hype, and the current high number of news stories showing this technology still grounded, it was brilliant to hear how drones are out there in the world helping businesses make better decisions with Watson.  Find out more about Watson IoT’s work with Aerialtronics.