Menu
▾
▴
Re: [Flightgear-devel] Abnormal surface friction factor
|
From: Wayne B. <kil...@co...> - 2023-06-13 07:20:59
|
Hi all, I finally had a chance to look at this a little closer. This unreasonably high rolling friction factor is only being calculated under the JSBSim FDM. Under YASim it is not dividing the material value by .02 and you end up with a reasonable value. Also why would you use static values directly from the material settings but not the rolling values? Yasim From: David Megginson Sent: Thursday, June 1, 2023 6:57 PM To: FlightGear developers discussions Subject: Re: [Flightgear-devel] Abnormal surface friction factor With modern data sources like OpenStreetMap or high-res landcover rasters, many of those areas can match real life instead of just being randomly generated. but for lower res scenery, adding some random clearings might not be a bad idea. For example, newer updates of my scenery (like w080n40) show actual streams and creeks, so they don't need the fake repeating ones in the texture. I call these my "wet" builds, because they include even small areas of water. I find it makes a huge difference for the verisimilitude of the scenery. I'll be extending the wet builds to all of the areas as they get updated. D On Thu, Jun 1, 2023, 3:49 p.m. Fahim Dalvi <dal...@gm...> wrote: Hello all, … open areas in a woodland scene …. … open area in and amongst the tree covered sections …. …. A fire break in a woodland could possibly be enough to land on…. Just to chime in here, are these areas well defined out in the world, or more or less “random”? I can imagine if it is indeed random, its potentially possible to lets say post-process the WS30 landclass textures and generate these patches in with some programming (or even do it procedurally at runtime some day depending on weather, time of year etc). Best, Fahim On Jun 1, 2023, at 10:15 PM, David Megginson <dav...@gm...> wrote: But more generally, I agree that FlightGear is a long way from simulating bush flying realistically — I don't think any consumer soon does that yet. For example, skis and floats have very little relationship to real life: there are no water currents, shallows, wakes from other ships, shoaling water, whitewater around and over rocks, snowbanks or windrows, different textures and stickiness of snow, different grades and thicknesses of ice (or water gaps). Winds around mountains or tall hills are also still very basic. Etc. Even maintained grass strips don't feel much like grass, and we don't distinguish wet from dry grass (as far as I know) or long from short grass. It would be great if we could focus on a few scenery areas to at least provide more places for bush planes to land while we work on the other problems. David p.s. The closest I've come to bush flying IRL is landing on short grass and gravel strips and iced-over runways, encountering the occasional mountain wave in the mountains of eastern North America, and one lenticular cloud sighting over Mount Katahdin in Maine. 🙂 On Thu, Jun 1, 2023, 3:03 p.m. David Megginson <dav...@gm...> wrote: thanks for checking it out, Wayne. what we need is better definitions in OpenStreetMap. I used 30 m land cover raster because it was available for canada, the us, and Mexico, but there are finer grained rasters available. Still, OpenStreetMap can define very small features, like Miami Beach in my scenery gallery. as OSMers get more active in remote areas, we should see more of those appearing — in fact, feel free to add some to OSM yourself. One interesting challenge that would come from reducing the rolling friction in forested areas would be a strange situation where people could take off and land more easily through trees. That occurs because we don't actually detect collisions with our runtime scenery objects; if we did, then we could have lower rolling friction between trees, assuming you were going through scrub or grass or dirt or whatever the undergrowth was. But for now, you can just fly or taxi through trees as if they weren't there. D On Thu, Jun 1, 2023, 2:36 p.m. Wayne Bragg <kil...@co...> wrote: David, sorry that wasn’t your scenery, I forgot to point at it. This is though and it is basically the same over much of the big country. Inline bug report, ha, the white patch is the same cover as the surrounding area but missing texture for some reason and its rolling is 1 where the rest of it is 50? I think it was EvergreenBroadCover. I don’t know what happened there. Anyway, this is to my point that 50 is either just too high a value for any but the most “sticky” or “soft” ground or we need to be able to break these areas down with incredibly more robust resolution in order to facilitate the real life features as it relates to bush aircraft, ie: landing in the bush country. Wayne From: Wayne Bragg Sent: Thursday, June 1, 2023 12:26 PM To: FlightGear developers discussions Subject: Re: [Flightgear-devel] Abnormal surface friction factor David, here is an example of what I am referring to. Using your scenery out of Whitehorse CYXY. There are two or three different covers adjacent to that area that are rolling 50. Broad swaths of ground that you cannot land on. Only a road, town or grassland, etc. Not even river bottoms, they take on the same landcover that is designated as 50 rolling. That is not accurate at all. In all that acreage are thousands of areas you should be able to land a bush equipped aircraft on. I submit, 50 rolling friction over that vast of an area is the wrong way to go here. See: https://i.imgur.com/pDuhCIR.jpg https://i.imgur.com/NDxvzsw.jpg Wayne From: David Megginson Sent: Thursday, June 1, 2023 5:08 AM To: FlightGear developers discussions Subject: Re: [Flightgear-devel] Abnormal surface friction factor In fact, it is feasible to have that kind of resolution for global scenery. We've seen it in the improved European scenery, and I've been building much-more-detailed scenery for Canada and the US based on 30m landcover rasters and OpenStreetMap (it includes most of Canada and Mexico and all of CONUS; Alaska is coming soon): https://davidmegginson.github.io/fgfs-scenery-build/ Depending on the source, the landcover rasters will capture anything bigger than either (nominally) 10×10m or 30×30m, which should take care of most places a bush plane might like to land. There is no limit on the resolution in OpenStreetMap, though I filter out very small details to keep the scenery from becoming too large. Many beaches, sandbars, clearings, etc show up in mine. Between the two of them, it's already possible to have much better scenery than the default terrasync scenery for North America, which is currently based on low-resolutiin and inaccurate data sets from more than two decades ago. D On Wed, May 31, 2023, 7:01 p.m. Wayne Bragg <kil...@co...> wrote: >This sounds like a bug in the scenery, not in the code. Better-resolution scenery for Alaska would include sand and/or gravel cover for riverbeds and basins. In a perfect world, maybe, but I don’t think it would be feasible to get that kind of resolution over the entire world scenery. At any rate, 50 is just to high for anything other than mud. Think about all the open areas in a woodland scene. It would be unlikely that you could ever account for that by trying to cut in different ground cover types for all those possibilities. You’d have to define every open area in and amongst the tree covered sections. A fire break in a woodland could possibly be enough to land on. Better to err on the side of not limiting access rather than shutting out those areas to bush craft. Leaf litter doesn’t absolutely mean soft ground. Tundra tires give you a really large footprint and for the most part should work in much of that terrain. I don’t remember exactly all the cover types I tested and programed for in the scripting I did for the c172p and j3 but I had it dialed in pretty accurately. In much of the cover types, the standard gear would not roll easily or at all, but changing to the tundra tires allowed you to land and takeoff in those areas. Same with snow, I have it tuned to not allow standard gear to work, but switch to tundra tires and your good to go up till a few inches. Then the skies and floats work. This type of flying (and landing) is loads of fun, I don’t want to loose it. Same as on ice, that was working well until we made the change to shader ice visuals and had to change some code to get that working again. I think we left it in a working state, but there is more we can do in that area also. Erik, if there is anything you need or want or if you need any help reworking the ground reactions, let me know, I’ll do everything I can to help make this a robust system. Wayne From: David Megginson Sent: Wednesday, May 31, 2023 1:42 PM To: FlightGear developers discussions Subject: Re: [Flightgear-devel] Abnormal surface friction factor On Wed, May 31, 2023, 7:02 a.m. Wayne Bragg <kil...@co...> wrote: > for what it is worth, I would not consider EvergreenBroadCover to be terrain friendly to airplanes. Typically I would agree, yet the resolution of much of our scenery isn’t detailed enough to support such a broad assumption over such a large area. All the riverbeds and basins are defined as this landcover, likely many other areas. This sounds like a bug in the scenery, not in the code. Better-resolution scenery for Alaska would include sand and/or gravel cover for riverbeds and basins. _______________________________________________ Flightgear-devel mailing list Fli...@li... https://lists.sourceforge.net/lists/listinfo/flightgear-devel _______________________________________________ Flightgear-devel mailing list Fli...@li... https://lists.sourceforge.net/lists/listinfo/flightgear-devel _______________________________________________ Flightgear-devel mailing list Fli...@li... https://lists.sourceforge.net/lists/listinfo/flightgear-devel _______________________________________________ Flightgear-devel mailing list Fli...@li... https://lists.sourceforge.net/lists/listinfo/flightgear-devel |
