Realistic Ground Friction

[Sasquatch_Punter]

I'm not sure why, but every surface in this game feels like a skating rink. You crash, and your parts go sliding down a gentle slope for miles. Kerbals glide along the ground at high speeds instead of ragdolling properly. Even rovers can't handle more than 10m/s before their wheels start acting drunk...

In all honesty, driving is one of the least fun and most annoying aspects of the game for me, when it could be so much more fun. Imagine driving high-speed rovers across Duna or even Minmus without having to contend with the sliding bug? Or wheels that are responsive to braking and never take 30 seconds to go from 10 m/s to 0.

Imho the ground-related physics need love too.

[Hagen von Tronje]

I can absolutely say that rover wheels will hold up just fine and drive perfectly straight up to nearly 50m/s, at which point they break. Nor do they take very long to accelerate, unless you've brought way less wheel than you need for your vehicle.

Claw's bugfixes have a brakes patch, if that's what you need.

But if it's realism you want, you'd be disappointed in the results. Fair to say Curiosity doesn't go 25m/s (the standard top speed for rover wheels on a flat with no additional thrust), which is just about highway speed in real life. Rovers in game can drive offroad faster than the speed limit anywhere but the Autobahn, exactly how much hotrod does the game need? Doesn't seem to be a ground friction thing either, considering I've climbed both up and down 40 degree slopes and more on nothing but wheels. Now, if you're in a low-gravity environment and have no downforce, expect terrible traction. That IS realistic.

Also fair to say I disagree with your assessment that rovers are unfun, I've enjoyed driving them all over the solar system. If yours are going 10m/s and operate like the driver is drunk, that's maybe a design issue you should iron out. Strongly suggest mirror symmetry to prevent bad alignment on the wheels, might cause a lot of issues like you describe.

PS - Assuming you aren't bound to the idea that rover wheels = rover, rovers can go 100m/s+ with ease on Duna. Try landing gear if you insist on setting a land speed record.

Edited August 12, 2015 by Hagen von Tronje

[Red Iron Crown]

Good news, the wheels are being completely reworked for the transition to Unity 5, apparently they will behave in a much more realistic manner.

[mythbusters844]

Kerbal Foundries has a stock wheel patch that replaces the game's wheel module with the one custom made by the Kerbal Foundries devs, and it's incredibly fun. I spent an entire day just driving over the surface of the Mun and I didn't flip over at all (except once when I did a sharp turn at 40 m/s).

All you need for the patch is ModuleManager and the KerbalFoundries mod (you only need the plugin and sounds). Note that this will break craft in flight using rover wheels and saved crafts using them might not work.

I can absolutely say that rover wheels will hold up just fine and drive perfectly straight up to nearly 50m/s, at which point they break. Nor do they take very long to accelerate, unless you've brought way less wheel than you need for your vehicle.

Claw's bugfixes have a brakes patch, if that's what you need.

We must be playing different games then, because before I got the KerbalFoundries wheel patch, my rovers would spontaneously flip out for no reason even when driving at 10 m/s and with a center of mass below the ground itself.

But if it's realism you want, you'd be disappointed in the results. Fair to say Curiosity doesn't go 25m/s (the standard top speed for rover wheels on a flat with no additional thrust), which is just about highway speed in real life. Rovers in game can drive offroad faster than the speed limit anywhere but the Autobahn, exactly how much hotrod does the game need? Doesn't seem to be a ground friction thing either, considering I've climbed both up and down 40 degree slopes and more on nothing but wheels. Now, if you're in a low-gravity environment and have no downforce, expect terrible traction. That IS realistic.

The only reason why Curiosity drives that slow is because it's a science rover. If it was, say, for transportation, then it would go much faster, and maybe even 25 m/s.

Also fair to say I disagree with your assessment that rovers are unfun, I've enjoyed driving them all over the solar system. If yours are going 10m/s and operate like the driver is drunk, that's maybe a design issue you should iron out. Strongly suggest mirror symmetry to prevent bad alignment on the wheels, might cause a lot of issues like you describe.

I've tried everything, but my rovers always spontaneously flip out in stock.

Edited August 12, 2015 by mythbusters844

[Sasquatch_Punter]

@HvT The lunar landspeed record is currently about 5m/s, yet slopes still cause lateral slide when trying to drive straight on the Mun at that speed. The problem is that traction gradients aren't implemented very well; you can see this in effect when your rover takes a small turn and you have a sudden unexplained loss of traction, followed by your wheels catching the ground and flipping your 3-ton rover on its ass.

A light, 1-ton rover should have more than enough mass to get traction on Duna, but inevitably they start sliding and the wheels start lifting off the ground. This isn't a small, isolated incident, either; the forum search turned up a bunch of threads about this exact issue.

@RIC, I didn't hear about that. That's good news!

Edited August 12, 2015 by Sasquatch_Punter

[Hagen von Tronje]

What can I say except that I've had a totally different experience. I've also documented much of it, and it can be read in my signature link. The Elcano uses mods but none of them alter physics, and the rover performs much the same when merely driving with all mod parts stripped, as I try to use mod parts in ways more interesting than merely "make it easier."

I suppose I just don't have the same assumptions. You took a "small turn", the loss of traction may be sudden but it's hardly unexplained. Besides loss of traction to turning itself, it's possible you altered the slope on which you're driving, which could lead to wheels losing contact almost immediately by simple leverage. If you're zoomed out or it's dark, or the terrain is just hard to see, you can easily drive over small incline changes without even noticing, as well.

I just feel most of it is exactly as I would expect, or even more forgiving than I would expect. On the real moon or Mars I expect most attempts at 50m/s+ speeds would be met with rapid unplanned disassembly as the reality of loose substrate and randomly scattered rocks sets in. Maybe there are some sites where it would be possible, but it's hardly the performance I'd expect as a rule. The best information I could find indicates that the real lunar rovers top speed was something like 15 kph (according to curator of the Smithsonian, this was the only direct estimate of speed rather than extrapolation from distance/time I could find).

I don't mean to be contrary, I simply ask what evidence there is that any unrealistic behavior KSP rovers have is anything other than generous to the player. Is there any good source to suggest that strapping rockets to landing gear on Mars would have better results than it does in KSP? Cause that's quite effective in the game - if real life is actually way better than that and we're getting the gimped version of extraplanetary rocket sleds, I'd definitely like to see some pics or video of that.

Also because I don't mean to be unhelpful - try adding downthrusters of some kind, or at least a torque wheel. Maybe also check to make sure you aren't trying to drive around in all-wheel drive/all wheel steering mode, that can make you flippy all by itself and I still sometimes forget to fix that the first time I send out a rover. SAS works in a rover just fine and will act on the wheels even if you have no torque, and that can save you a lot of trouble when driving alongside slopes. Give them another chance man, check out my Minmus notes and you'll maybe decide they have some practical application after all!

PS - My first attempt at stock munar rovers were extremely light and had no downthrust, and performed not TOO much differently than you describe, though I could still go 15-20m/s on relatively flat terrain, though braking in time to slow for a hill was a problem. The design changes I suggest changed all that. For merely driving reasonable distances (read: non-circumnavigational), I think a reasonably compact rover could accomplish a great deal, though due to the effect of downforce, bigger rovers are actually favored if anything.

Edited August 12, 2015 by Hagen von Tronje

[Sasquatch_Punter]

I think it's also due to the way the wheel module handles surface geometry; riding over an edge onto a very shallow decline causes momentary "air", and the wheels end up bouncing around as they fight to regain traction. On heavier rovers the suspension is loaded moreso to both decrease this air and push the wheels into the declining surface, therefore you have better handling of edges. The problem is that this isn't all that realistic. Neither is the sliding that kerbals experience when they hit the ground at high speed; regolith isn't smooth, it piles up like sand and slows objects immediately.

[Snikersnee]

I think it's the weight. If I were to make a 1 ton rover, it would flip like crazy under kerbin's gravity. Let alone the weight a kerbal can push. Although I do think the traction needs to be reworked. Lets hope for The Unity 5 update.

[Xavven]

IMO there's a balance that needs to be achieved between realism and fun when it comes to rover traction. IRL a rover on a low gravity moon is going to have to go very slowly, because less gravity = less tire traction, and on top of that there are no paved roads but rather loose, dusty regolith, which means even less traction. That said, I don't think it would be fun AT ALL to explore Mun in a rover going 15kph. Realistic, yes. But boring as heck, and I'd rather just EVA a kerbal and jetpack to where I need to go.

So to up the fun factor, we need unrealistically good traction that lets us speed at 30m/s across rocky, uneven terrain. Tank treads, maybe? Anyway, I'm with the OP, but let's see what update 1.1 gives us and go from there.

[NWM]

The main thing, the quality of the ground doesn't mean anything: ice, grass, dust or stone are same...

Edited August 24, 2015 by NWM

[Claw]

Lots of reasons and things...but the only one I'll mention is that spontaneously flipping rovers are most often caused by either not disabling the rear wheel steering, or because torque and steering controls coincide. That's not meant to say stock wheels are perfect, but once I got used to building with them, my rovers stopped flipping "on their own."

Yes, the brakes are broken. Either never (ever) right click on the wheel, ot use my fixes.

Sliding is definitely another issue...

The main thing, the quality of the ground is doesn't mean anything: ice, grass, dust or stone are same...

This would actually be pretty easy to simulate with the current system. We'll see if they do anything with this in 1.1...

Cheers,

-Claw

[Val]

My solution is simple.

Less gravity = more wheels.

Less mass = better traction, for the same number of wheels.

(KSP wheels and surfaces seem like inflexible structures, where you don't get better traction with more mass, like you do, to a degree, in real-life)

If a design drives well on Kerbin with 4 wheels, then it need proportionally more wheels to drive equally on a lower gravity body.

Half the gravity = double the number of wheels.

Edit: Those are my design guides, for successful rover designs.

Edited August 23, 2015 by Val