BeyondATC

I brought it up on Discord. Unfortunately it seems they have no way of addressing this issue.

I added it as a wishlist item. Got upvoted pretty quickly. Clearly an in-demand feature. Its kind of essential - more so than traffic management (which seems a far more challenging issue to solve).

https://beyondatc.nolt.io/771

2 Likes

How is this now difficult to resolve when they stated in their video on ground traffic, that once that’s done ā€œair traffic is the easy partā€??

You misread me a little: as two features, pilot selectable approaches would be an easier problem to solve compared to the very complex challenge of traffic.

Of course they have invested a lot of dev time and resources on implementing traffic and at this point is may be relatively easier to do. I’m just surprised that pilot-selectable approaches was not considered a core feature and implemented prior to traffic.

No no, I got you. I am only baffled how they stated 7 Months ago that "once they are in the air … and that should be much easier "

21:00

I understand that every dev studio wants to jump on the bandwagon but can they please think about the implications of their statements before stating something??:roll_eyes:

Also I’m with you on the request an approach-thing

1 Like

Ah I see! Well, yes indeed…

I’m afraid I’ve shelved BATC for now. Causing more problems than solutions for me. About to try my 24 SI free trial…

You may be equally as frustrated with SI but you can at least request different runways/approaches with it.

I don’t know why both of these programs are focused so much on implementing traffic when the core/structural ATC stuff doesn’t work properly in many cases with both BATC and SI…

4 Likes

It may be that some of the core ATC has dependency that the core SIM in current state can’t do well (airspace, in a nutshell). So any 3rd party ATC will be limited.

Generally speaking I prefer GA and I prefer VFR so maybe SI will be better for me. I will say this for BATC - the voice recognition is excellent. Very high rates of accuracy. Unlike Pilot2ATC, which had me yelling with frustration at times…

The best part about BeyondATC is the voice interaction. They really nailed that well, but when I’m assigned a visual approach it doesn’t recognize when I tell them I have the runway in sight, so I have to use the menu for that response. I’m an American whose lived in southern California all my life, so I don’t really have any accent that it would have a hard time with.

Have the same problem actually. Perhaps a bug?

I think if BATC would allow you to call ā€˜airport in sight’ further out it would be so much better. For me, the last Visual approach I did had them vectoring me basically in a traffic pattern (clear skies) and I wasn’t ā€œallowedā€ to call airport in sight until I was 1-2 miles away. It should allow this 10-12 miles out and then you could maneuver as you need for the approach.

Stop calling ATC ā€œdude.ā€

/s

3 Likes

Agreed .. some ATC are Dudettes

This is starting to go Off-Topic folks. Let’s get back on track.

Because without traffic there is no need for ATC. Its main purpose is to manage traffic. Why would I need ATC if I am the only plane in the air and I am approaching an empty airport?

7 Likes

My point is that if they will also be controlling the AI traffic, they need to have the core ATC working first. Having ATC vectoring a 777 onto a 2 mile final at 2000’ isn’t going to work for AI any more than it works for me. Both ATC programs seem nearly incapable of providing proper/consistent vectors, approach clearances and procedures which need to be working first before being focused on AI traffic…IMO of course.

1 Like

It’s cool that AI can generate the voices, but the program would be far more powerful if AI could learn proper approach procedures for the airports. It would save the developers a lot of time in the long run.

But isn’t that part of the Developer’s job, ( assuming the ATC is going to use AI to control Traffic) to TEACH the ATC " to provide proper/consistent vectors, approach clearances and procedures"

How else is the AI to learn – not from a load of novice sim pilots ā€œdoing their thingā€ !@!

Not that simple, perhaps?

In the real world AI has to deal with a lot of random humans who may or may not follow the rules. A simulator is a controlled environment. For now, it seems the developers are having to tweak procedures manually, based on user feedback, but proper procedures are documented in the real world, so why not build an AI engine that takes in said documents to generate approach procedures for each region? You have a flight plan if you’re flying IFR, which as of now Beyond ATC requires, so it knows where you’re going. It should also know the conditions, and when traffic injection is introduced it will know everything about them too.

It’s a sure way to apply machine learning. And as AI engines are developed, you will likely see a lot of applications developed this way.