Attack Vector add-ons

[Ndangin]

Good day all,

I am very new to FlightGear and require some assistance if possible.
I would like to use FlightGear for a comparative analysis of Drone Vulnerabilities and for that I would need to be able to load drones into the simulator as well as various cybersecurity attacks e.g DDoS, GPS Spoofing and Jamming on the drones. Is there any way I can do this?

I have installed the simulator on both Ubuntu 14.04 and Windows 10 so either option would be very helpful. Any and all help would be immensely appreciated.
Installed flightGear 2018.3.1
Thank you

[Hooray]

Hi & welcome,

yeah, that would definitely be possible - the question is what exactly you have in mind, i.e. in terms of soft/hard requirements, programming languages/libraries involved etc.

For instance, you mention "add-ons" in your title, which kinda implies using FG's native add-on mechanism (via Nasal scripting). While that is definitely do-able, if you have certain scaling requirements, you will find that a fgms/multiplayer based implementation should be superior. However, the Nasal based option is fairly lightweight and there isn't much of a barrier to entry - especially not for people with a programming/scripting background.

So there are basically two main ways to "inject traffic" into the fgfs world: 1) using Nasal scripting to create scripted AI objects and 2) using custom multiplayer packets to create arbitrary traffic.

How to proceed from here will definitely depend in your short-term goals.

You can find many relevant pointers in the following topics that covered "fgfs traffic injection" a while ago:

web GIS talking to FG to display traffic in (near) real time
Displaying additional aircraft on certain position

Obviously, this will only cover the very first requirement, i.e. adding/updating models at runtime - modeling different attacks will then be a different matter (think making updates less frequent, more random, more erratic, less reliable etc) - equally, GPS spoofing could be simply modeled by using a randomized offset for the corresponding APIs

Approaching thing that way would mean that you could prototype simple attacks using roughly 50 lines of Nasal code per attack vector, and then take it from there to model more complex behavior.

Obviously, if you are doing all this out of fgfs, you can ignore this part and simply do the traffic injection part.
Even though it would definitely be interesting to have some of this contributed back into fgfs, especially because we have a growing community of RC/UAV pilots and DIY-hardware enthusiasts.