Forums

Full Version: Nashville Street Circuit (WIP thread)
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
(10-17-2021, 04:45 AM)Dennis Wrote: [ -> ]Samsepi0l, could you make also a test? Maybe you can drive more constant also through the corners.

Did you do this on a straight at cleveland since it is so wide? If this was a straight, the SG file won't have a radius so how would we provide a radius to your formula?

I will be glad to do this for you and provide both a replay and a trafo file. I can also make a video.

I have a nice wheel and pedals so keeping a consistent speed isn't too hard.

I just want a clear understanding of what you want me to repeat so that I get it right the first time.
(10-17-2021, 05:33 AM)samsepi0l Wrote: [ -> ]
(10-17-2021, 04:45 AM)Dennis Wrote: [ -> ]Samsepi0l, could you make also a test? Maybe you can drive more constant also through the corners.

Did you do this on a straight at cleveland since it is so wide?  If this was a straight, the SG file won't have a radius so how would we provide a radius to your formula?

I will be glad to do this for you and provide both a replay and a trafo file.  I can also make a video. 

I have a nice wheel and pedals so keeping a consistent speed isn't too hard.

I just want a clear understanding of what you want me to repeat so that I get it right the first time.
I made it on my track, it has a track width of 30m on the oval part. I mentioned Cleveland, because it's suited for the straights, but as I heard, it has weird corners.
The factor c is 1 on straights, so no radius needed.
You wrote earlier, that you also make a video to look for the speed. So you could also check the result, if you want.
The task is to drive with constant speed around the track (or a part of it), in snake line on the straights an in corners with large radius, and on the racing line in corners with small radius.
Time to update this thread with a new video of a full lap of the AI. It's a little rough in some places but crucially, it goes through the chicane OK (with a bit of a bobble on exit) and has no problem making it through the narrow section up the hill. The lap was based on the replay that samsepi0l created, with a couple other LP files that I put together quickly.

@samsepi0l I thought I'd post the LP files I've made so far, to give you something to work with. In particular, I generated MINPANIC and MAXPANIC using a tool that I am working on - hopefully it works well enough for now. The CSV files are meant to work with the CSV to LP converter I posted on the other thread.
(10-20-2021, 05:38 PM)checkpoint10 Wrote: [ -> ]@samsepi0l I thought I'd post the LP files I've made so far, to give you something to work with. In particular, I generated MINPANIC and MAXPANIC using a tool that I am working on - hopefully it works well enough for now. The CSV files are meant to work with the CSV to LP converter I posted on the other thread.

I took your LPs and put them in my track directory, 1009 records. I converted them to txt via EDIT45....

25 77.70 0.00 -10.13
26 72.82 0.00 -10.01
27 68.54 0.00 -9.97
28 64.38 0.00 -9.97
29 61.03 0.00 -10.02
30 56.18 0.00 -10.10
31 52.86 0.00 -10.14
32 48.81 0.00 -9.84
33 47.26 0.00 -8.68
34 45.44 0.00 -6.63
35 43.16 0.00 -3.38
36 40.26 0.00 1.69
37 35.77 0.00 8.52
38 31.85 0.00 17.78
39 29.18 0.00 29.75
40 30.34 0.00 43.26
41 34.53 0.00 55.23
42 37.79 0.00 64.43
43 40.12 0.00 70.79

44 43.16 0.00 74.92
45 46.55 0.00 77.54
46 49.09 0.00 78.55
47 50.35 0.00 78.48
48 50.92 0.00 77.21
49 51.35 0.00 74.53

I would have never guessed the lateral movement we would need would be this high... 13 ft movement between records... wow

EDIT: I think I see why you have some "bumping" going on - your max/minrace is not outside of the race.lp in many spots, but I'll fix all of those.

At this point, I am specifically interested in "seeing" where your max/min panic and max/minrace files are showing up in the track... so I think what I am going to do, is rename them to RACE one at a time, put some crazy-high values in for all of the max/min stuff and watch the AI so I can see how accurate they look.

I always make sure I tune the max/min/panic stuff first because it will cause problems with race.lp if they aren't right.
Hmm... I may have bad news.

I wanted to see where the race.lp was without the effect of bouncing off of any other LPs. The bump you saw in your video was due to maxrace.

I put the entire maxpanic file at 140ft, maxrace at 135ft. The maximum position in your race.lp is 132, so I overkilled it.
I put the entire minpanic file at -35ft, minrace at -30ft. The minimum position of your race.lp is -25 (overkill again)

I then watched some replays of your AI. Only 1 opponent on the track. Exiting the chicane it acts like the car hits a wall in almost the exactly same spot as you have, however it bounces violently back to the right and ends up on the inside of the runoff area in the entrance portion of the chicane. I went in and reduced the speed in the chicane to about 30mph, no effect.

I tried increasing the max values from 140 and 135 to 160 and 150, but there was also no difference.

I'm wondering if this is just too much lateral movement for the physics of the game. I'm going to try and screw with the positions of the where the line is in the LP. I don't know what to do besides this.

There are some other problems with the LP that I am not worried about because they seem very solvable- however the one thing of interest is the car takes a HUGE wide turn going onto the right approaching the 2nd bridge which I thought was strange.
I made some position adjustments and it does not crash as hard- so the car does not do a sumersault into the escape-road....but it most definately is bouncing off of something, which I don't understand because max/race/panic are WAY off to the left.

Is it possible that the game thinks the pit-wall is there? Maybe this is a track.txt problem?

If we ever get this damn thing to work, I'm going to nickname this chicane "the monkey-business" chicane.
I GOT IT!!!

it was a track.txt problem.  Value 11 isn't going to work.

This is how I fudged it:
PIT  252000 58800723 61411954 59578188 62306778 428250 2886755 14 59218188 59578188 2107806
PIT2  336000 58800723 61411954 64101421 2036460 523000 2886755 14 59218188 59578188 2107806

When I saw it not bounce back- I did one of these in my living room:



hahah
Thanks for looking into this. I was playing around with the LPs last night and saw all kinds of strange behavior at the chicane, but I thought what I posted had the AI going through the chicane the cleanest (when the car was alone). I was also mostly testing at 96% difficulty which may have slowed the cars down. I did notice that when other cars were there, they tend to do strange things like jumping to the other side of the track. So I was wondering if maybe it's MAXRACE/MINRACE or other LPs that were affecting it. I didn't do a very serious job at MAXRACE/MINRACE/PASS1/PASS2 so any of those could be causing issues.

Another thing I noticed, when I was chasing other AI cars, they would sometimes disappear momentarily from the road ahead. I am wondering if this is because of the sections where I've created additional track on the other side (like on the bridge, near the Firestone turn, or run-off near the chicane), or if it's because of the edits to the .3D file from fixing the elevation see-through glitch.

As for possible issues with track.txt - generally I am sure we need to tinker with those. I haven't really done anything except get the pits in the approximate places but I think all of it needs some refinement. I also wonder if it is possible to get the cars to start on the bridge, but keep the start/finish line where it is.

Edit: I saw you just solved the problem. Great!
(10-21-2021, 09:55 AM)samsepi0l Wrote: [ -> ]There are some other problems with the LP that I am not worried about because they seem very solvable- however the one thing of interest is the car takes a HUGE wide turn going onto the right approaching the 2nd bridge which I thought was strange.
That turn has a super small radius for its centerline and the track is actually way to the left of center, and it was the turn where my replay to CSV converter was giving me negative speed, which I actually don't understand how that could be. But anyway I manually put in the speeds there. I did notice the line seemed unnatural and probably needs some attention.
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30