simulator VS MSFS2020-PilotPathRecorder

I've been trying for a while now and I started thinking it may not be possible. If anyone has managed to train a self-driving car simulator using openai gym on google colab(preferably), or on any remote server (AWS, GCP, ...) please let me know. So far, I tried carla, airsim, svl, deepdrive and they are all equally useless unless run locally with a gui. I'd really appreciate if someone suggests some way that actually can make it possible.

here

There are many ways to approach the planning and recording your progress, but I will advocate for simplicity here. Google Earth is free and widely used app. The reason I use it for planning is because you can examine the terrain in 3D and determine if it will be interesting enough to fly over. But it also becomes a great visual record of the trip once you use a simple app like PilotPathRecorder that is a freeware and it will generate a Path that can be easily imported into Google Earth as KML file.

This was the track of my attempt in MSFS 2020 to land a Cessna 172 at KIAD 1R using the localizer with DoNotBeOnFire's fog preset active. I tracked the approach using Pilot Path Recorder, then recorded the track in Google Earth using OBS.

I recently completed a scenic tour of the US in the DC Designs F-15. I designed the route around national parks, major US cities, and places I've been before. The speed of this plane enabled me to cover hundreds of miles in 30-60 minute flights, and finish the trip in a reasonable amount of time. The flights were recorded with pilot path recorder and imported into Google Earth. VFRmap was also used for navigation. I took close to 500 screenshots from the trip, but narrowed it down to a few below.