Jump to content

4.268 Routes for vehicles - getting them to stick to the roads.


Gibsonm

Recommended Posts

  • Members
1 hour ago, Gibsonm said:

But its not a sharp turn, the planned route is as per the purple line:

Sorry, I didn't look sharp enough.

Nevertheless, SOMEHOW the unit ended up on the wrong side of the wall - maybe it didn't overshoot but left the road for other reasons, and once that the wall prevents it from returning to the road (because it can't even sense that it's there), it's trapped. So, on the one hand effort should be made to prevent the unit from leaving the road just before the wall. But that's something that you can only make "less likely". And if it is a location where this happens with some regularity despite your efforts to prevent it, the next best option is to use a Navmesh supported route that still is on the road (you just can't use Shift+Click) until you have bypassed the problematic location. So if the unit deviates, it at least has the navmesh to calculate an alternate recovery route to the next waypoint that is guaranteed to be free of known obstacles, and where the unit can also handle dynamic (=moving) obstacles.

Link to comment
Share on other sites

10 hours ago, Ssnake said:

Sorry, I didn't look sharp enough.

Nevertheless, SOMEHOW the unit ended up on the wrong side of the wall - maybe it didn't overshoot but left the road for other reasons, and once that the wall prevents it from returning to the road (because it can't even sense that it's there), it's trapped. So, on the one hand effort should be made to prevent the unit from leaving the road just before the wall. But that's something that you can only make "less likely". And if it is a location where this happens with some regularity despite your efforts to prevent it, the next best option is to use a Navmesh supported route that still is on the road (you just can't use Shift+Click) until you have bypassed the problematic location. So if the unit deviates, it at least has the navmesh to calculate an alternate recovery route to the next waypoint that is guaranteed to be free of known obstacles, and where the unit can also handle dynamic (=moving) obstacles.

 

Here's the video:

 

https://u.pcloud.link/publink/show?code=XZ2wgGXZNAaH5x3iACmt4zBA5VhLDplGzDf7

 

2 units are OK.

 

1 visits the farm and continues

 

1 visits the farms and stays.

 

 

Edited by Gibsonm
Link to comment
Share on other sites

  • Members

At this point, I think that further iterations will yield diminishing results (don't want to waste too much of your time). We have an idea of where problems occur and will attempt to replicate them on our end, then see what we can improve on the Steel Beasts side of things.

Link to comment
Share on other sites

Well I have to get the scenario to work. :)

 

Iteration 4 far better.

 

Link to AAR: https://u.pcloud.link/publink/show?code=XZ2wgGXZNAaH5x3iACmt4zBA5VhLDplGzDf7

 

Only one issue, caused by "dual occupancy" and waypoint 237 is now fine (six units went past the farm gate and none were tempted to visit):

 

1796076526_Resultrun4.thumb.png.3e3e8ed3b81238e0a7a9091a58cdc37a.png

 

837555337_Dualoccupancy.thumb.png.0abd8d0d89dbb63e3d842de54ebb6a77.png

 

Going forward I think I'll reduce the delay at the bridges from 30 sec to 20 sec (just to provide a buffer) and perhaps pull the waypoint just before the bridge back slightly so if there is an issue, vehicles can recover.

 

If I only have bogged motorcycles, that's quite alright.

 

Frustration levels far lower now than at the start of this post!

 

Edited by Gibsonm
Link to comment
Share on other sites

  • Members
5 hours ago, Gibsonm said:

Here's the video:

Thanks.

After countless interruptions I was able to actually watch it. Since this is taken from the AAR, I presume that the scene happened in time accelerated mode?

 

I should stress one more time that time acceleration mode will create more breakdowns of pathfinding than will happen if the scenario runs in real-time. I understand that you want to test a large scenario in time accelerated mode to save some time, but I think this also results in a lot more cases of seemingly erratic behavior of computer-controlled units because the time steps in the simulation are bigger and therefore small errors get scaled up which can lead to big errors.

It may be too much to ask for a leap of faith here to just run with it, but I think during the actual exercise you won't see these kind of accidents. My recommendation would be to create a mini scenario of the problematic locations, and then have two or three dozen vehicles take the route and see if, without time acceleration, any of them will actually fail to reach the destination. If they all make it, you have very good chances that it'll work during the exercise too, without having to wait four hours for test results.

 

And then you can run the same mini test in time accelerated mode. If you now see erratic behavior again, chances are you have found the culprit.

Link to comment
Share on other sites

31 minutes ago, Ssnake said:

Thanks.

After countless interruptions I was able to actually watch it. Since this is taken from the AAR, I presume that the scene happened in time accelerated mode?

 

I should stress one more time that time acceleration mode will create more breakdowns of pathfinding than will happen if the scenario runs in real-time. I understand that you want to test a large scenario in time accelerated mode to save some time, but I think this also results in a lot more cases of seemingly erratic behavior of computer-controlled units because the time steps in the simulation are bigger and therefore small errors get scaled up which can lead to big errors.

It may be too much to ask for a leap of faith here to just run with it, but I think during the actual exercise you won't see these kind of accidents. My recommendation would be to create a mini scenario of the problematic locations, and then have two or three dozen vehicles take the route and see if, without time acceleration, any of them will actually fail to reach the destination. If they all make it, you have very good chances that it'll work during the exercise too, without having to wait four hours for test results.

 

And then you can run the same mini test in time accelerated mode. If you now see erratic behavior again, chances are you have found the culprit.

 

The testing of the mission was run in real time, not accelerated. This is why they take some 5 hours to run. :)

 

The video uploaded of that part of the AAR replay was run using the ">>" button (i.e. Fast Forward) just to keep the size of the video and its uploading speed reasonable.

 

Of course the full AAR file is provided so you can watch it yourself in real time. :)

 

However this issue (waypoint 237) is now resolved (at least for the last two "real time" runs) because I cleaned up the waypoint itself and moved it North to the bend in the road.

 

Link to comment
Share on other sites

On 1/19/2022 at 12:24 PM, Ssnake said:

Okay, understood.

 

I'm sorry if that sounded a bit snippy. I was just trying to be clear.

 

Iteration 5 complete with no issues. :)

 

Now to replicate those changes into the non test scenario with the remaining units, etc.

 

Look fwd to seeing those various bugs squashed in due course.

 

Edited by Gibsonm
Link to comment
Share on other sites

On 1/18/2022 at 3:24 AM, Ssnake said:

Bug 9802 was just fixed, which covers the problem that waypoints in or in the vicinity of highway bridge #5 can sometimes not be reached by wheeled vehicles. It's a start, I guess.

 

Sorry, a last minute query. I should have asked earlier ...

 

Do you happen to know if this bug only manifests itself when a type 5 bridge is beside another (i.e. in this case, the dual carriageway of the A7)?

 

Does it also occur if the road is a single, two lane, carriageway with a single highway bridge #5?

 

Just trying to determine if I only need to edit the routes that cross a highway bridge #5 on the A7 or if I need to edit all the routes that cross a highway bridge #5, regardless of the road type.

 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...