## speed-dreams-devel

 [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-04-28 16:50:30 ```Hi Kristof My name is Max and i found that you working on simu for speed dreams project and thats a reason why i'm writing you this letter I think it will nice feature to add tire temperature and wearing for car simulation model The main problem is that i'mnot a physic But anyway i have few ideas about tires: 1 Temperature Its very important parameter for tire, for an example in formula one teams used special device to warm tire before going on track. Also for first 1-2 laps while tires is not in working temperature they cant work as they must So i think we can calculate temperature based on summary forces that applied to the tire, we can use simple formula like this : y=e^x/20 where x is temperature, and y is total force applied to tire 2 tire wear Its much more complicated, but i found equations based on track temperature, tire temperature, wheel stiffness and so on please check this link http://www.dtic.mil/dtic/tr/fulltext/u2/699806.pdf pages 2 and 3 Will be nice to hear your opinion about this Regards Max ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-04-28 22:54:47 ```Hello Max, welcome on the SD board! And thanks for the paper, it looks interesting. Tire temperature and wear is planned for the future, in two steps. The second step, that is my long term goal is to have a nice tire model based on physics considerations, where grip, wear and temperature all came from a single model of the tire material. This will be part of our new simulation engine, which is expected in a few years time. The first step can be done in short term, which is to add a heuristic model to our current tire model. Here are my first ideas how it would look like (to SD devs: it is a slightly enhanced version of what was in Quo Vadis). Tire temperature: - The work of the friction force can be easily calculated, and a certain portion of it heats up the tire. - A hot tire is cooled by cold air around it. - In rain, some water gets to the tires surface, adding additional cooling. - All of these can be put into basic thermodynamic equations (energy balance), resulting in a changing temperature. I think these parameters are relatively close with different tires, so they hopefully can be kept constant. - The grip will depend on the temperature, first I would use a parabolic dependence, as it is the simplest function giving a maximum. Here the optimum temperature and a parameter describing how fast grip decreases with temperature are dependent on the type of tire. For example rain tire in dry weather can easily overheat, resulting in reduced grip; whereas dry tire in rain will be under its optimal temperature, again resulting in reduced grip. Tire degradation: - I would base this also on the work done by the friction force, simply multiply the work with a constant to keep in a suitable range, and sum up these to get total degradation number. - As tire wear increases with tire temperature, I would increase this multiplier with temperature. I have no idea how this is in real life, so first I would use a linear function, then look into the paper you sent me to find a better formula :). - With increased total degradation number grip decreases slowly and linearly. At a certain value there is a sudden drop, when basically all rubber worn off, and the wheel runs on the canvas. Maybe another drop, when the tire gets a puncture. - This grip - degradation curve can be described by another 2-3 pairs of parameters, describing at which degradation how much grip we have, and the curve is linear within a segment (like the torque curve). The coding of this is much easier then to find good parameters. What are your aims with SD? I mean you want to use it for some research, or want to make it better? I ask because for a research project you can hack in anything, but if the aim is to improve SD we need to respect the development rules. Anyway, if you are interested, I can put in such a simple model as an optional thing, then you have like 10 parameters to play with and find realistic values. Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-04-29 09:47:46 ```Hi Kristof Thank you for your reply Me and my friend has hobby project, were you playing as driver's race engeneer On this game you need to setup a racing car and plan a strategy for race We dont need any 3d engine but quality of racing car sumulation is very important We spend enough time for searching good physic engine and finally stopped on speed dreams Tire physics is very important for us and we would help to make your car physics model better But problem is that we are not as good in physics as in programming So if you will be so kind to provide few equations about tires we can try to program it as part of the simu library Or you can do it by yourself, and we will take care of finding good values for parameters We understand that speed dreams has own rules for development process so we are ready to follow them Regards Max On 29/4/13 02:54 , Kristóf Kály-Kullai wrote: > Hello Max, > > welcome on the SD board! And thanks for the paper, it looks interesting. > > Tire temperature and wear is planned for the future, in two steps. The > second step, that is my long term goal is to have a nice tire model > based on physics considerations, where grip, wear and temperature all > came from a single model of the tire material. This will be part of > our new simulation engine, which is expected in a few years time. > > The first step can be done in short term, which is to add a heuristic > model to our current tire model. Here are my first ideas how it would > look like (to SD devs: it is a slightly enhanced version of what was > in Quo Vadis). > > Tire temperature: > - The work of the friction force can be easily calculated, and a > certain portion of it heats up the tire. > - A hot tire is cooled by cold air around it. > - In rain, some water gets to the tires surface, adding additional cooling. > - All of these can be put into basic thermodynamic equations (energy > balance), resulting in a changing temperature. I think these > parameters are relatively close with different tires, so they > hopefully can be kept constant. > - The grip will depend on the temperature, first I would use a > parabolic dependence, as it is the simplest function giving a maximum. > Here the optimum temperature and a parameter describing how fast grip > decreases with temperature are dependent on the type of tire. For > example rain tire in dry weather can easily overheat, resulting in > reduced grip; whereas dry tire in rain will be under its optimal > temperature, again resulting in reduced grip. > > Tire degradation: > - I would base this also on the work done by the friction force, > simply multiply the work with a constant to keep in a suitable range, > and sum up these to get total degradation number. > - As tire wear increases with tire temperature, I would increase this > multiplier with temperature. I have no idea how this is in real life, > so first I would use a linear function, then look into the paper you > sent me to find a better formula :). > - With increased total degradation number grip decreases slowly and > linearly. At a certain value there is a sudden drop, when basically > all rubber worn off, and the wheel runs on the canvas. Maybe another > drop, when the tire gets a puncture. > - This grip - degradation curve can be described by another 2-3 pairs > of parameters, describing at which degradation how much grip we have, > and the curve is linear within a segment (like the torque curve). > > The coding of this is much easier then to find good parameters. > > What are your aims with SD? I mean you want to use it for some > research, or want to make it better? I ask because for a research > project you can hack in anything, but if the aim is to improve SD we > need to respect the development rules. > > Anyway, if you are interested, I can put in such a simple model as an > optional thing, then you have like 10 parameters to play with and find > realistic values. > > Cheers, > Kristof ```
 Re: [Speed-dreams-devel] blocked wheel From: Kristóf Kály-Kullai - 2013-05-23 19:07:45 Attachments: Message as HTML ```Hello Max, So now my question = have you any ideas how tire wearing should affect on > mfB mfC mfE ? > In my plans these parameters will remain the same, regardless tire wear or temperature. See wheel.cpp how they are calculated from parameters given in the cars xml file. Cheers, Kristof ```
 Re: [Speed-dreams-devel] blocked wheel From: Max Rozanoff - 2013-05-23 19:11:33 Attachments: Message as HTML ```I see but what you will change for wearing simulation ? some mu factor ? Max On 23/5/13 23:07 , Kristóf Kály-Kullai wrote: > Hello Max, > > So now my question = have you any ideas how tire wearing should > affect on mfB mfC mfE ? > > > > In my plans these parameters will remain the same, regardless tire > wear or temperature. See wheel.cpp how they are calculated from > parameters given in the cars xml file. > > Cheers, > Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-04-29 20:58:09 ```Hello Max, > Me and my friend has hobby project, were you playing as driver's race > engeneer Nice idea. Dreaming about future: would be great to couple it one day with SDs online race mode - while someone drives, the other guy plans the strategy, maybe on the other end of Earth. > We spend enough time for searching good physic engine and finally > stopped on speed dreams Thanks, that feels good. > Tire physics is very important for us and we would help to make your car > physics model better I agree, tire physics are the most important for a good simulator. > So if you will be so kind to provide few equations about tires we can > try to program it as part of the simu library > Or you can do it by yourself, and we will take care of finding good > values for parameters I will have a look this week. If the equations are done, it is 1-2 hours to code it, soI will do it. > We understand that speed dreams has own rules for development process so > we are ready to follow them Great, then hopefully we will have a good collaboration. Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-04-29 21:07:03 ```Hi Kristof On 30/4/13 00:58 , Kristóf Kály-Kullai wrote: > Hello Max, > >> Me and my friend has hobby project, were you playing as driver's race >> engeneer > Nice idea. > Dreaming about future: would be great to couple it one day with SDs > online race mode - while someone drives, the other guy plans the > strategy, maybe on the other end of Earth. Actually we planned to use robots but using real human as drivers sound interesting >> So if you will be so kind to provide few equations about tires we can >> try to program it as part of the simu library >> Or you can do it by yourself, and we will take care of finding good >> values for parameters > I will have a look this week. If the equations are done, it is 1-2 > hours to code it, soI will do it. Excellent. Good luck. Cant wait for result :) > >> We understand that speed dreams has own rules for development process so >> we are ready to follow them > Great, then hopefully we will have a good collaboration. > > So is there any instructions that we need to follow ? Or we can try whatever we want, but at the end we need to make a special report (i.e. some docs with patch files ) ? Regards Max ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-04-29 21:18:11 ```Hello Max, > So is there any instructions that we need to follow ? Or we can try > whatever we want, but at the end we need to make a special report (i.e. > some docs with patch files ) ? Well, to start you can read on our wiki - Roles and responsibilities: http://sourceforge.net/apps/trac/speed-dreams/wiki/RolesResponsibilities - Developers Charter: http://sourceforge.net/apps/trac/speed-dreams/wiki/DevelopersCharter - and our Release Method (it is quite detailed and long): http://sourceforge.net/apps/trac/speed-dreams/wiki/ReleaseMethod At first you would be newcomers, and then yes, basically you create patches with some docs when needed. After some time if you still want to become a developer, the team will vote, and if the majority agrees you will become a developer with write access to our repo. Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-06 21:11:39 ```Hi Kristof Any luck with equations ? If you need more articles about tyres - let me know. Lot of them are closed by paid subscription but internet is huge and with patient you can find anything you need ) Regards Max > > I will have a look this week. If the equations are done, it is 1-2 > hours to code it, soI will do it. > >> We understand that speed dreams has own rules for development process so >> we are ready to follow them > > Great, then hopefully we will have a good collaboration. > > Cheers, > Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-05-07 08:20:21 ```Hello Max, well, temperature is almost done, I just need to fix a bug in my implementation. Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-10 16:59:19 ```Hi Kristof Forget to say - sometimes ago i bought "Motor vehicle dynamic. Modeling and Simulation" by G.Genta But looks like its complicated for me. If you need it - i can send it to you Or maybe i can pass it to you, since at begining of June i will be at the nord of France Regards Max On 7/5/13 12:20 , Kristóf Kály-Kullai wrote: > Hello Max, > > well, temperature is almost done, I just need to fix a bug in my implementation. > > Cheers, > Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-05-13 14:19:53 ```Hello Max, > Forget to say - sometimes ago i bought "Motor vehicle dynamic. Modeling > and Simulation" by G.Genta > But looks like its complicated for me. If you need it - i can send it to > you If you are sure it is too complicated for you, and noone in your neighborhood can help, than yes, I think it could give me a few ideas. > Or maybe i can pass it to you, since at begining of June i will be at > the nord of France Well, I live and work in Hungary, not in France. Though if one of our French developers wants to have alook at the book, you can hand it him. Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-13 14:24:58 ```Hi Kristof So give me a post address to send this book Regards Max On 13/5/13 18:19 , Kristóf Kály-Kullai wrote: > Hello Max, > >> Forget to say - sometimes ago i bought "Motor vehicle dynamic. Modeling >> and Simulation" by G.Genta >> But looks like its complicated for me. If you need it - i can send it to >> you > If you are sure it is too complicated for you, and noone in your > neighborhood can help, than yes, I think it could give me a few ideas. > >> Or maybe i can pass it to you, since at begining of June i will be at >> the nord of France > Well, I live and work in Hungary, not in France. Though if one of our > French developers wants to have alook at the book, you can hand it > him. > > Cheers, > Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-05-16 22:20:15 ```Hello Max, the tire temperature model is now committed. See the ticket #819, and the wiki page mentioned in that. Please use that ticket to track issues with it. Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: - 2013-05-16 22:39:10 ```> Hello Max, > > the tire temperature model is now committed. See the ticket #819, and > the wiki page mentioned in that. Please use that ticket to track > issues with it. Awesome work, look forward to playing with it..... however my excuse for being 'lazy' is that my PC died a few weeks ago and I've be swapped/busy with other stuff. I now have another PC (albeit it limited) up but am waiting for more hard-disks to arrive (from Ebay'ers). Simon ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-17 05:50:27 ```Hi Kristof Grear job, thank you ! Now what kind of test you want me to perform ? Regards Max On 17/5/13 02:20 , Kristóf Kály-Kullai wrote: > Hello Max, > > the tire temperature model is now committed. See the ticket #819, and > the wiki page mentioned in that. Please use that ticket to track > issues with it. > > Cheers, > Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-17 06:16:02 Attachments: Message as HTML ```Kristof I have a question - your equations doesn't include track temperature, this parameter will be used in the future ? Also please correct wiki page "http://sourceforge.net/apps/trac/speed-dreams/wiki/TireTempDeg"; in " mu(T) dependence" section you need to replace Tair on Topt so it should looks like this: mu(T) = mu0 * MAX(1 - muTmult * (T - Topt)^2 , 0.1) Regards Max On 17/5/13 02:20 , Kristóf Kály-Kullai wrote: > Hello Max, > > the tire temperature model is now committed. See the ticket #819, and > the wiki page mentioned in that. Please use that ticket to track > issues with it. > > Cheers, > Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-05-17 12:43:55 ```Hello Max, > I have a question - your equations doesn't include track temperature, > this parameter will be used in the future ? Not now, probably for simuv5. I think track temperature directly has a minor effect on tire temperature, as only a small part of the tire is in contact with it. But when the track is hotter than air, it means that air temperature is increased close to the track, it gradually decreases from track temperature to air temperature measured far away from the asphalt. And the tire is close enough to the track that this temperature rise should be taken into account. But for now, we do not have a proper modeling of the environment, so it is easier to set everything to 273 K, and change from it when having a correct model. > Also please correct wiki page > "http://sourceforge.net/apps/trac/speed-dreams/wiki/TireTempDeg"; in " > mu(T) dependence" section you need to replace Tair on Topt > so it should looks like this: > mu(T) = mu0 * MAX(1 - muTmult * (T - Topt)^2 , 0.1) Thanks, good catch. As for the tests, well, I think it is up to you to decide on the car/robot combos you would like to use for the strategy game, and find good parameters to give correct temperatures. Well, with 0 C air temperature I suggest to decrease real life optimum tire temperatures by 20-25 degrees. Note, that by default cold mu factor is set to 1, meaning that mu does not change by temperature. It is because: - IMO it is easier to set up temperature parameters when mu does not change, and mu dependence after, that is one by one; - it is another measure against unwanted activation of this feature; - and robots will probably have problems with changing mu. As for robots, I have added an effectiveMu field to tWheelState, and the macro _tyreEffMu(i) to access it quickly from tCarElt. tCarElt is sent to robots, so they can be patched to use this to find braking points and estimate cornering speeds instead of using the constant mu. I guess robot developers will not have time to do it before releasing 2.1.0, sowe have to do it by ourselves now, and commit as an optional feature, maybe only after the release. And when temperature parameters are ready and robots can handle changing mu, it is time to find a cold mu factor. I suggest to set a target time penalty for the out lap on cold tires, and find the value which results in the desired out-lap time. Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-23 11:22:12 Attachments: Message as HTML ```Hi Kristof I'm continue with your code for handling tire temperature and i spot strange thing once you calculated tire condition in tireCond variable, you store it in car->carElt->_tyreCondition(index) but on the next step you didnt use that value, tireCond initialized with 1.0 and recalculated from scratch: if (car->features & FEAT_TIRETEMPDEG) { tireCond = 1 - wheel->muTmult * (wheel->Ttire - wheel->Topt)*(wheel->Ttire - wheel->Topt); tireCond = MAX(tireCond, 0.1); mu *= tireCond; } why you didnt use current value of tireCond ? Thank you for your reply Max ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-17 14:22:47 ```On 17/5/13 16:43 , Kristóf Kály-Kullai wrote: > Hello Max, > >> I have a question - your equations doesn't include track temperature, >> this parameter will be used in the future ? > Not now, probably for simuv5. I think track temperature directly has a > minor effect on tire temperature, as only a small part of the tire is > in contact with it. But when the track is hotter than air, it means > that air temperature is increased close to the track, it gradually > decreases from track temperature to air temperature measured far away > from the asphalt. And the tire is close enough to the track that this > temperature rise should be taken into account. I see For an example for last Spain GP Ttrack were about 38C when Tair were 20C I have no idea how temperature changed from asphalt till the top of the wheel. So we can leave this, cause its gonna be too complicated But i have another question - can you add air wind that cooling the wheel We able to know Tair and wind speed. There is can be problem with wind direction, but we can determine car position and calculate what area goes under the wind Adding wind we can operate with track temperature (what you already done) and air temperature > > But for now, we do not have a proper modeling of the environment, so > it is easier to set everything to 273 K, and change from it when > having a correct model. Ok, thank you once again >> Also please correct wiki page >> "http://sourceforge.net/apps/trac/speed-dreams/wiki/TireTempDeg"; in" >> mu(T) dependence" section you need to replace Tair on Topt >> so it should looks like this: >> mu(T) = mu0 * MAX(1 - muTmult * (T - Topt)^2 , 0.1) > Thanks, good catch. > > As for the tests, well, I think it is up to you to decide on the > car/robot combos you would like to use for the strategy game, and find > good parameters to give correct temperatures. Well, with 0 C air > temperature I suggest to decrease real life optimum tire temperatures > by 20-25 degrees. Note, that by default cold mu factor is set to 1, > meaning that mu does not change by temperature. It is because: > - IMO it is easier to set up temperature parameters when mu does not > change, and mu dependence after, that is one by one; > - it is another measure against unwanted activation of this feature; > - and robots will probably have problems with changing mu. > > As for robots, I have added an effectiveMu field to tWheelState, and > the macro _tyreEffMu(i) to access it quickly from tCarElt. tCarElt is > sent to robots, so they can be patched to use this to find braking > points and estimate cornering speeds instead of using the constant mu. > I guess robot developers will not have time to do it before releasing > 2.1.0, sowe have to do it by ourselves now, and commit as an optional > feature, maybe only after the release. I will try to change Mu on effectiveMu in simplix robot source code and will see what will happens But as i know simplix robot create driving line before the race so it can be a surprise for him that mu is changed Anyway lets see what we will have ... > And when temperature parameters are ready and robots can handle > changing mu, it is time to find a cold mu factor. I suggest to set a > target time penalty for the out lap on cold tires, and find the value > which results in the desired out-lap time. Not sure that understand your idea. Having low value of mu at the first lap (when tire is cold) not allow you to have the same speed in cornering as when your tire is warm So in this case time penalty will be calculated automatically Also i wan to ask - can i use compilied simuv4 library with SD 2 ? or i need to to compile SD executable from special branch ? Thank you for your answers and great work. And BTW - any plans for tire degradations ? :) Regards Max ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-05-18 12:34:24 Attachments: Message as HTML ```Hello Max, But i have another question - can you add air wind that cooling the wheel > We able to know Tair and wind speed. > There is can be problem with wind direction, but we can determine car > position and calculate what area goes under the wind > Well, I think it would not have a big effect. Here the easy thing to do is simply subtracting wind speed from the wheel speed (as vector), and use the results magnitude as v in the formula for tire temperature. This would not count with the car shielding some of the wheels. And is only good for open wheelers, while most of our cars have chassis around the tire, which would make it even harder to guess what happens. And when the wind blows from the back, it would actually decrease the cooling, so through a whole lap the effect becomes small (exactly zero on a circle with constant speed). And wind has a much bigger effect on aerodynamics, which is not modelled, and as it requires robots to handle the situation most likely will only happen in simuv5. But yes, it will be built to have a weather model. Another thing I have neglected is the sunshine heating up black rubber, again mostly important for open wheelers. To be constructive: for the winds effect you might try to change aircoolm by a few percent. If it has the desired result, we will find a way to change it from wind data. > I will try to change Mu on effectiveMu in simplix robot source code > and will see what will happens > But as i know simplix robot create driving line before the race so > it can be a surprise for him that mu is changed > Anyway lets see what we will have .. As mu should not change more than, say, 10% during normal degradation/heating/cooling, it does not really affect the driving line. Sure, a world class driver would do slight changes, but simplix having the same line, and only changing its speed will be enough to cope with this grip change. Only pre-calculated braking points are problematic, but I think Simplix calculates its braking dynamically (that is from its current situation). > Not sure that understand your idea. Having low value of mu at the first > lap (when tire is cold) > not allow you to have the same speed in cornering as when your tire is warm > So in this case time penalty will be calculated automatically > Yes, if the parameters are given, the robot drives and you get the time penalty. What I wanted to say is that most likely you do not have data regarding grip dependence on temperature. But most series have live timing, so you can get a good estimation of the time penalty. Then you can change the parameters, let the robots run, and compare the time penalty with the desired, then again change the parameters to get closer, etc. > Also i wan to ask - can i use compilied simuv4 library with SD 2 ? > or i need to to compile SD executable from special branch ? > With the recent changes definitely not. The problem is that tCarElt is used in several other modules and libraries, so you should replace at least half of them - than it is easier to compile the current version. You can check out our svn repository, trunk is the current development branch with all the new stuff. You can do the checkout by the command: svn co https://svn.code.sf.net/p/speed-dreams/code/trunk ./trunk It is now 7.7 GB, so the initial checkout will take a lot of time. > And BTW - any plans for tire degradations ? :) > Sure. I have an idea how to calculate it from basic statistical physic principles, will do it right now. If I do not like the result, I will use degradation=a*W^b, which I have found in recent papers. Probably with b=1, as I do not have any data on it, and it is easy to calculate :). Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Xavier Bertaux - 2013-05-18 12:49:58 Attachments: Message as HTML ```Hi Kristof, Hi Max, Le 18/05/2013 14:34, Kristóf Kály-Kullai a écrit : > Well, I think it would not have a big effect. Here the easy thing to > do is simply subtracting wind speed from the wheel speed (as vector), > and use the results magnitude as v in the formula for tire > temperature. This would not count with the car shielding some of the > wheels. And is only good for open wheelers, while most of our cars > have chassis around the tire, which would make it even harder to guess > what happens. And when the wind blows from the back, it would actually > decrease the cooling, so through a whole lap the effect becomes small > (exactly zero on a circle with constant speed). > > And wind has a much bigger effect on aerodynamics, which is not > modelled, and as it requires robots to handle the situation most > likely will only happen in simuv5. But yes, it will be built to have a > weather model. I think implemented a module (or lib) weather in the future - for 3.0, I think. Better for have a real weather for track - by curl lib, these lib have max visibility, precipitation (rain or snow) time start and stop precipitation in day, air temperature and track temperature, clouds type, time start and end for fog, etc ... > > Another thing I have neglected is the sunshine heating up black > rubber, again mostly important for open wheelers. > > To be constructive: for the winds effect you might try to change > aircoolm by a few percent. If it has the desired result, we will find > a way to change it from wind data. > > > I will try to change Mu on effectiveMu in simplix robot source code > and will see what will happens > But as i know simplix robot create driving line before the race so > it can be a surprise for him that mu is changed > Anyway lets see what we will have .. > > > As mu should not change more than, say, 10% during normal > degradation/heating/cooling, it does not really affect the driving > line. Sure, a world class driver would do slight changes, but simplix > having the same line, and only changing its speed will be enough to > cope with this grip change. Only pre-calculated braking points are > problematic, but I think Simplix calculates its braking dynamically > (that is from its current situation). In start on weather (two years) I had implemented a modification on mu in race but WD wrote me that it's not a good idea because simplix calculate good raceline before race start. So robot will have a code for dynamic weather in the future but not for now, and I don't know when. And for the moment just Simplix use weather, I tried to modified USR's code for this but for the moment USR has a good setup for no rain and little rain, but with medium or high rain USR lost many time by lap (simplix time / 2, I think). But is it really necessary that robot's include this (for the moment), for me we could take this for human and keep actually code for robot - it's will added difficulty for human for winning race, because degradation tire will changed strategy for human player > > > Not sure that understand your idea. Having low value of mu at the > first lap (when tire is cold) > not allow you to have the same speed in cornering as when your > tire is warm > So in this case time penalty will be calculated automatically > > > Yes, if the parameters are given, the robot drives and you get the > time penalty. What I wanted to say is that most likely you do not have > data regarding grip dependence on temperature. But most series have > live timing, so you can get a good estimation of the time penalty. > Then you can change the parameters, let the robots run, and compare > the time penalty with the desired, then again change the parameters to > get closer, etc. > > > Also i wan to ask - can i use compilied simuv4 library with SD 2 ? > or i need to to compile SD executable from special branch ? > > > With the recent changes definitely not. The problem is that tCarElt is > used in several other modules and libraries, so you should replace at > least half of them - than it is easier to compile the current version. > > You can check out our svn repository, trunk is the current development > branch with all the new stuff. You can do the checkout by the command: > svn co https://svn.code.sf.net/p/speed-dreams/code/trunk ./trunk > It is now 7.7 GB, so the initial checkout will take a lot of time. > > > And BTW - any plans for tire degradations ? :) > > > Sure. I have an idea how to calculate it from basic statistical physic > principles, will do it right now. If I do not like the result, I will > use degradation=a*W^b, which I have found in recent papers. Probably > with b=1, as I do not have any data on it, and it is easy to calculate :). > > Cheers, > Kristof Cheers Xavier ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-18 17:08:26 Attachments: Message as HTML ```Hi Kristof Short question about trunk - text only simulation works there ? Regards Max > Also i wan to ask - can i use compilied simuv4 library with SD 2 ? > or i need to to compile SD executable from special branch ? > > > With the recent changes definitely not. The problem is that tCarElt is > used in several other modules and libraries, so you should replace at > least half of them - than it is easier to compile the current version. > > You can check out our svn repository, trunk is the current development > branch with all the new stuff. You can do the checkout by the command: > svn co https://svn.code.sf.net/p/speed-dreams/code/trunk ./trunk > It is now 7.7 GB, so the initial checkout will take a lot of time. > ```
 Re: [Speed-dreams-devel] speed dreams simu From: Kristóf Kály-Kullai - 2013-05-18 17:34:53 ```Hello Max, > Short question about trunk - text only simulation works there ? Yes, it is supposed to run. If does not work, file a bug report. Cheers, Kristof ```
 Re: [Speed-dreams-devel] speed dreams simu From: Max Rozanoff - 2013-05-18 17:48:54 ```Thank you for quick answer ! Max On 18/5/13 21:34 , Kristóf Kály-Kullai wrote: > Hello Max, > >> Short question about trunk - text only simulation works there ? > Yes, it is supposed to run. If does not work, file a bug report. > > Cheers, > Kristof ```
1 2 > >> (Page 1 of 2)