DJI Aeroscope data leak exposes 80k user IDs and the folly of RID

Jurgita Lapienytė  Chief Editor of Cybernews writes about their discovery of a 54Gb Aeroscope database held on an AWS server in America. That’s 90 million flights logged.

We knew Aeroscope was out in the wild, able to track most brands and used by several companies but did we know that the data from them became joined up?

Was this data from a single company with 66 Aeroscopes 53 of them in the USA?

I know one of you knows all the details of this, that’s the beauty of the sUAS News reader!

A good friend of sUAS News @d0tslash would love a copy of the data if you have access and Jurgita would like to know who the data belongs to.

This foreshadows data sets to come that will be collated by RID systems. RID is much much much cheaper and easier to decode. It’s an open standard.

For many years aviation regulators ruled the roost when it came to knowing where many aircraft were.

They owned the big expensive massive infrastructure required, radar systems.

The first primary, is just a signal bounced back from something metal and then secondary a small code added to identify the track and altitude. This is a very simplified explanation.

Then back in the last century folks started discussing creating a better system that would not require big rotating radar heads but just receivers. Taking advantage of new-fangled GPS to transmit aircraft position heading and altitude. Big companies were then able to sell some new receiver infrastructure on the ground. Hooray radar and new receiver sales ka-ching.

What was not foreseen was the rise of single-board computers, in particular the Raspberry Pi and SDR software-defined radios. One each of those clever coders and you have what was sold to regulators for millions of dollars for hundreds of dollars. Yes, it’s not as sensitive but because of its low cost, it scales much faster than government systems.

This is how services like FlightAware and FlightRadar24 came into being. They now sell their data to airports to drive flight status display boards and to companies, so they can keep a handle on their aircraft. It is very clever. Very low cost and now has much better coverage than official multimillion-dollar systems.

ADS-B is a horrible standard that is easily spoofed and does not really deserve to be in our drone digital sky.

Remote ID (RID) is very short-range, Bluetooth 5 and WiFi. But like the Raspberry Pi and ADS-B that problem can be mitigated with many many receivers.

Let us imagine we want to start capturing RID data for an entire city. Looking to hook some VC cash.

When we start out we chuck low-cost devices with a 400-metre range out into the wild, deploy 10 and realise it is going to take an age, we then buy better receivers and site them more appropriately and get the reception out to 750m and finally roof mount external antenna and all the way out to 1.5km.

It might look like this for Cincinnati.

We start selling data to LEO’s and Part 107 competitors so things are on the up and up and we make our ultimate receiver systems and get as far as 2km on a good day.

It’s become possible to cover a city at a great resolution for a lower cost than a single Aeroscope!

But blast this is the tech world and whilst we have been rushing the rollout of our receivers a competitor has found a way of leveraging a Bluetooth 5 mesh network that has become the latest and greatest big thing. Unlike LoRA the darling of mesh, BLE5 is built into phones so it was much easier to roll out. Hundreds of teenagers in Cincinnati are using the mesh to stop their parents from seeing their online chats and the coverage is vast. They have made sure there are paths to all the high schools.

Oh, bother.

ADS-B arrived as the great safety saviour of general aviation, and no thought at all was given to emerging technologies. It can be spoofed from simple COTS devices. I would bet that there is already firmware that does it for this.

The unintended consequences of RID and its collection of data have not been thought through.

Aviation regulators need to think long and hard, not having the wool pulled over their eyes by vendors who tell them things can never happen.

Can the range of Bluetooth 5 on a drone be imagined I wonder?

Oh and finally, remember the data was not leaked by DJI, it was an Aeroscope user. Another future security vector to think about.

Its a brave new world.