Special Aircraft Service

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 ... 42 43 44 [45] 46 47 48 ... 51   Go Down

Author Topic: Weekly progress report  (Read 121831 times)

0 Members and 1 Guest are viewing this topic.

Pursuivant

  • member
  • Offline Offline
  • Posts: 711
Re: Weekly progress report
« Reply #528 on: June 26, 2021, 10:26:20 PM »

My design is for solid bones connected by rigid joints to hold all the aircraft parts in place. So when the aircraft gets destroyed you can just break the joint at the correct place and the aircraft falls apart as you would expect.

There's a hidden benefit here which you might not have considered yourself - the ability to bend those joints with or without breaking.

That lets failing parts bend before they break.

E.g., at 1:08 on this video:



If you wanted to get really fancy, you could have joints which deform under stress, so that vehicles crumple when they crash.
Logged

Stainless

  • moderator
  • member
  • Offline Offline
  • Posts: 1531
Re: Weekly progress report
« Reply #529 on: June 27, 2021, 12:00:32 AM »

yes , I need to look at how to do that.

One of the things I was thinking about was the testimony of a P51 pilot from world war two.

He dove in on a ME109 at low altitude and fired a 2 second burst from six o clock high

He says he saw the rounds hit either side of the cockpit by the wing root. Left guns hit left of the cockpit, right guns hit right of the cockpit.

The result was two small explosions then the wings folded upwards.

He flew alongside the doomed plane for a few seconds and could see the pilot look at him. Since the cockpit on the 109 folds out to the left, and the wing was blocking it, the pilot had no chance to bail out. He knew he was dead.

The P51 pilot says that one haunted him the rest of his life. Like most pilots he shoots down planes, not people.

I have talked to other people and they say it's impossible. The structure of the 109 means the wings would come off in that situation, not fold up.

I tend to believe the pilot more than the physics
Logged

Chaoic16

  • Modder
  • member
  • Offline Offline
  • Posts: 930
Re: Weekly progress report
« Reply #530 on: June 27, 2021, 05:09:38 AM »

Hello Stainless,

When I saw this video, I thought of you and the projects you are working on.  I thought these videos might have anything helpful:

Realistic Aircraft Physics for Games


Ocean waves simulation with Fast Fourier transform
Logged

Pursuivant

  • member
  • Offline Offline
  • Posts: 711
Re: Weekly progress report
« Reply #531 on: June 28, 2021, 11:46:51 PM »

I have talked to other people and they say it's impossible. The structure of the 109 means the wings would come off in that situation, not fold up.

The wings would have quickly come loose, but they could have hung on for a few seconds, attached by the skin and a partial spar, after they folded up. That would could have been long enough to keep the Bf-109 pilot from bailing out before his airplane went into an uncontrollable spin or got too low to allow a bail out.

I was also thinking of being able to fly a "bent" plane - like a Fw-200 with a "broken back" due to a hard landing, a plane with wings bent out of alignment due to serious abuse or severe weather, or a plane with bent propeller tips due to a flying too close to the ground.

Anyhow, it's another potentially very powerful tool once you get it working.
Logged

Stainless

  • moderator
  • member
  • Offline Offline
  • Posts: 1531
Re: Weekly progress report
« Reply #532 on: July 10, 2021, 01:22:49 AM »

So I have finally figured out how the taxiways work in FSX





You have an array of points , then an array of connections. A connection is just the indexes into the point array.

Which makes life hell.

This shows the problem



You can't just draw a thick line as you can see in the "simple" section above.

So what I have done is run through all the points all count the number of times they are referenced by paths. From this I can work out the basic shape of each intersection.

  • Start
  • Line
  • T junction
  • Crossroad
  • This is going to hurt

Now I have to add code for all of the combinations of the connection counts, so far I have done start - start and start - line which was complex enough

I had to form a ray along each edge of the start section, create a plane from the two points of the next section and get the intersection of the two to create a point. Then do the same for the other edge.

Next is two to two which is very similar, but the rest of the combinations need more thought.

I have also started getting the airport into UE5, which is going ok.






Logged

Pursuivant

  • member
  • Offline Offline
  • Posts: 711
Re: Weekly progress report
« Reply #533 on: July 16, 2021, 07:39:50 PM »

So I have finally figured out how the taxiways work in FSX

Wow. Whatever "intuitive" is, that's the opposite. I wonder why the FSX coders programmed it that way? (No need to reply: rhetorical question.)
Logged

Stainless

  • moderator
  • member
  • Offline Offline
  • Posts: 1531
Re: Weekly progress report
« Reply #534 on: July 18, 2021, 02:02:29 AM »

Realised the way Iwas doing this is inefficient.

SO working on a new technique. Again.

Now after I have read in all the points , I read in all the paths and store all the connections made to each point.

Then I iterate over the points generating new points based on the number of connections.

If a point only has one connection, it is trivial. You create a direction from this point to the connected point, cross it to generate a right vector, and multiply that by half the width. Then I use this to generate two points and store then in the point structure.

If it has two points, I start off the same. Generating a direction to the next point and the right vector, but this time I do the same for the other connection. so I end up with four points and two directions. I then use these points and directions to generate four lines and calculate the intersection points of these lines, and store those two intersections.

After that it gets more complex, but basically I generate directions from the current point to all connected points. I then sort these so they are counter clockwise. Then it's the same process of generating right vectors, generating lines, computing intersections, and storing the points.

This image attempts to explain the process



Then I iterate over the paths using the generated points to build triangles.

So you end up with something like this




It can handle changes in width as you go along no problem, and in the fist pass doesn't look too bad.

However I am thinking that by storing the direction of each line and doing a bit more logic , I can make it better by using splines




Logged

Stainless

  • moderator
  • member
  • Offline Offline
  • Posts: 1531
Re: Weekly progress report
« Reply #535 on: October 16, 2021, 05:53:10 AM »

Got dragged into input handlers



The system uses a blueprint class that contains a list of inputs needed for a particular aircraft

It can connect to anything HID compliant, so all of your custom made switch boxes will work with it

The stuff at the bottom is just debug display, once I have finished the default display it will go

Logged

SAS~Storebror

  • Editor
  • member
  • Offline Offline
  • Posts: 23588
  • Taking a timeout
    • STFU
Re: Weekly progress report
« Reply #536 on: October 16, 2021, 11:49:17 AM »

Nice!
I love it when things are that shiny behind the scenes.

]cheers[
Mike
Logged
Don't split your mentality without thinking twice.

Stainless

  • moderator
  • member
  • Offline Offline
  • Posts: 1531
Re: Weekly progress report
« Reply #537 on: October 24, 2021, 02:20:54 AM »

This week I solved a problem that has bothered me for a while.

I need fully functioning sector control rooms. One of the key items in the control room is the squadron tally board. It displays the current state of all the squadrons.

I have tried a few techniques for this, none of them really worked. So I decided to mix 2D and 3D.

The display is actually a UMG widget with a blueprint behind it to control the status of all the lights, then I wrap it in a world object.

To add one to a room, you just drag one into the world and set the squadron number and the default state, then place it where you want it.




Now I just need to model the rest of the room


Logged

Stainless

  • moderator
  • member
  • Offline Offline
  • Posts: 1531
Re: Weekly progress report
« Reply #538 on: October 31, 2021, 05:02:11 AM »




Barrage balloon indicators
Logged

Stainless

  • moderator
  • member
  • Offline Offline
  • Posts: 1531
Re: Weekly progress report
« Reply #539 on: November 14, 2021, 09:46:12 AM »

Room coming together with my own player controller



Still a lot ot do and I have to figure out UE5's bloody awful lighting
Logged
Pages: 1 ... 42 43 44 [45] 46 47 48 ... 51   Go Up
 

Page created in 0.09 seconds with 27 queries.