Flyingcarpet.network early stage idea.

— Blog post for the non-technical people :) —

The dreaded elevator pitch. Keep it snappy. The best solutions are elegant and brief. Occam’s razor and all that. That’s what they tell you. But how do you condense your life’s work into a sentence. The culmination of your childhood, your study, your career and all of the experiences along the way into a nice little sentence. Blockchain, AI, big data, the internet of things (IoT) — our project encompass all of today’s buzzwords. We even get to use words like decentralisation, machine learning and tokenisation with a straight face. And I want to share it all with the world — or at least anyone who will listen — in all of its details.

But back to the elevator pitch — moment of truth — this is where we currently stand;

‘A communication protocol for IoT devices like drones, enabling visual data to flow throughout a tokenised ecosystem, incentivising participation.’

Fabulous. But what does that even mean?

So to start from the beginning,

2009 — Union College, upstate New York. Just about to complete my Masters in Electrical Engineering, specialising in artificial intelligence and decentralised systems. Putting the final touches on my thesis on deep reinforcement learning for robot swarms.

Me (left), the robot (in my hand), my thesis advisor (best teacher ever :))
Saturday night in the workshop / living room

At the same time, I decided to put all of that theory into practice and actually build an autonomous robotics system. I built a system that enabled robots to autonomously gather photos around the campus. I implemented reinforcement learning to teach the system to sell these photos to humans in exchange of electricity. It learned how to survive.

Testing out the Flyingcarpet autonomous drone charging station in my backyard

This concept stuck with me. With the innovations in blockchain technology (#Ethereum, State channels etc), and with IoT and AI in the position to bring real value to our society, I built a wireless charging station for drones in early 2017. Combined with my work in automating the flight and data collection, these stations not only allowed drones to operate autonomously, but also enabled them to do so over extended periods of time. Rich analytics and insights could now be derived from this system, without the need for ongoing human input. The seed was sown.

So once again, what’s it all mean? Stick with me…

We are building more than a market place — its an ecosystem. One where you can create your own solution to data collection. Or you can seek out someone to solve that for you. You can even collaborate with someone to capture and distribute data you didn’t even know you had access to (or was valuable). But beyond the marketplace itself, we are creating the framework for it all to happen. And its open-source.

Developers can design algorithms that enable any IoT device to not only collect data but to analyse, synthesise and extract this data — faster, cheaper and more accurately than current methodologies allow.

Want to map the melting of glaciers? Specifically — the presence of dusts, soot and microscopic algae that might be accelerating the melting process? Let me know, we can help. Using modern imaging systems and pattern recognition software, the drone can provide researchers with 3-D models of the ice surface, as well as multispectral imagery that can be used to identify and map the various materials that make up the glaciers. This information can then be used to monitor precisely how the glaciers are changing over time.

Dr Joe Cook, Arctic Researcher

These services are run through our Virtual AI Engine that validates the task’s proficiency and safety and, after a human validation phase that follows, services are pushed into the services marketplace. This validation process ensures that services are accurate — and safe.

Our Virtual AI Engine

As these services carry out tasks, the data is fed into a secondary data exchange, opening up the market to all sorts of external commercial opportunities. For example, crop yield estimates from a farmer can be sold, via the data exchange, to futures market traders on Wall Street who make coconut futures predictions, potentially earning more for the farmer than the value of their entire crop.

You lost me — give me a real world example.

The highlands of Papua New Guinea — August 2017 we brought it into the real world — and put it to the test.

Taken during our drone flight over the coconut plantation

We built an algorithm that enabled a drone to autonomously count the number of coconuts in a coconut plantation. The aim was to increase the accuracy and reduce the costs of crop yield predictions for the farmer. Traditionally the farmer would walk around with a clipboard and a pen. This process took days, taking time away from farm work and the results were far from accurate. From a 20 minute autonomous flight we were able to accurately collect data from the entire plantation and provide an accurate count and translate that into crop yield predictions. This information could then not only be used by the farmer to optimise distribution of fertilizers, water, etc., but sold via the data marketplace to a futures market commodity trader in London to make better predictions of market movements.

Me (left), coconut farmer Namaliu Jr. (center), and Ben Jackson of Abt Associates (right).

Now what?

My colleagues and I are continuing to build this platform so that we can share it with the world — working towards the testnest launch and releasing our whitepaper which will outline the project in greater detail.

So that’s the story so far. Subscribe to our Mailing List below and join us on Telegram to stay up to date on the next chapter.