Hi Dan and Harry, Thank you very much for looking into this for me. I confirm I can view the videos/files in your zipped file.
I find it interesting that you did not observe any weird behavior. Harry, can you please confirm which control log file you used when making that video? Did you also test both the trial2 and trial3 files? Would it also be possible for you to show me the lines that my code prints to the console for both of those runs? The current version of Chrono I am using is from late October or early November 2024 (no idea which version that would be). I've thought about trying the latest version (but someone else did the install for me and it was a pain to go through), but I couldn't tell from the github repo if anything related to the components to run this simulation have changed since then. Thoughts? One other question, how did you get the camera to track the vehicle like that in the "demo_rigid" video? Thanks, Ted On Mon, Jan 20, 2025 at 9:56 PM Dan Negrut <[email protected]> wrote: > Hi Ted, > > One of our lab students, Harry Zhang (thanks, Harry), too a look at the > material you sent. > > For some reason he cannot post on the forum at this point, so I’ll convey > the message from him. > > > > Please take a look here: > https://drive.google.com/file/d/1dEzGlTbs_n4dlS2QtM6juDvpV-aOtFTL/view?usp=sharing) > > > 1) Harry ran the same script as you did, for rigid terrain, and generated > a movie out of it on his Linux machine. From the demo, we didn't spot > abnormal behavior. Can you double check the video provided at the link > above? > > > > 2) For SCM, one issue is that you are using a "TMEasy" tire while running > the demo. That won’t work on SCM deformable terrain. SCM terrain typically > (but not always) requires a rigid tire, e.g., "Polaris_Rigid.json" or > "Polaris_RigidMeshTire.json" as json input. Harry attached a demo file that > he wrote to reflect this. Maybe you want to modify your SCM part and use > rigid tire during simulation. Or use Harry’s model and build off it. Note > that you can also use a deformable tire via finite element, but that is an > overkill. It’s not simple to set up, and it takes very long to finish a > simulation. Given that your project aims at solving a controls problem, > using rigid tires on SCM is adequate. > > > > Good luck with your project. > > > > Dan > > --------------------------------------------- > > Bernard A. and Frances M. Weideman Professor > > NVIDIA CUDA Fellow > > Department of Mechanical Engineering > > Department of Computer Science > > University of Wisconsin - Madison > > 4150ME, 1513 University Avenue > > Madison, WI 53706-1572 > > 608 772 0914 > > http://sbel.wisc.edu/ > > http://projectchrono.org/ > > --------------------------------------------- > > > > *From:* Ted Sender <[email protected]> > *Sent:* Sunday, January 19, 2025 11:41 PM > *To:* Dan Negrut <[email protected]> > *Cc:* ProjectChrono <[email protected]> > *Subject:* Re: [chrono] Tires Rotating Beyond Mechanical Limits > > > > Hi Dan, > > > > I created a minimally working example that will allow you and your team to > see the exact problem I am experiencing. I am attaching a zipped file that > contains two folders, a "src" folder for code and a "data" folder with some > necessary data. You will need to create a build directory associated with > the "src" folder. > > > > The "data" folder provides a few vehicle control sequences and the > associated terrain heightmap. "control_log_trial1.txt" contains a sequence > of commands that pushes chrono right near its simulation capabilities in > this scenario, but " control_log_trial2.txt" and "control_log_trial3.txt " > push it over the edge completely as the front right tire rotates well past > its mechanical limit while the wheel is not in contact with the ground. The > code is preconfigured to run the "trial2" commands. > > > > *Usage Instructions:* > > 1. At the top of the main() function you will see a code block that > configures a few settings. Follow the instructions provided to configure > the location of a few folders and select which control log file to use. > > 2. Build the code (you will need to create the build directory and > configure it using cmake). > > 3. Run the build/Release/my_demo.exe file from a powershell. > > > > *Notes:* > > 1. During part of this simulation, I print information regarding the right > steering wheel so you can see the tire and spindle forces/torques and the > actual steering angle. The timestamp on these print lines is off by 5 > seconds from the actual simulation time, as the vehicle is "disabled" for > the first 5 seconds of the simulation. > > 2. This code has only been tested on Windows 10 > > > > *Other Instructions:* > > Use the "find" command and search for "INSTRUCTIONS". If you want to try > this setup with SCM terrain instead of RigidTerrain, then follow the > instructions at the other two INSTRUCTIONS code blocks. > > > > *Other things I have tried:* > > 1. I tried all of the tire collision types: SINGLE_POINT, FOUR_POINT, and > ENVELOPE, but they all produce the same error. > > 2. I still get the same result with SCM terrain. > > > > Please let me know if you are able to find someone who can look into this > issue and if they are able to observe this problem using the code/data > provided. > > > > Thanks, > > Ted > > > > > > On Tue, Jan 14, 2025 at 2:41 PM Ted Sender <[email protected]> wrote: > > I wanted to give a brief update. I tried getting SCM terrain to work with > the default settings. I think I set things up correctly by looking through > the source code to see what I have to do. And I am STILL seeing this weird > problem. The problem seems to be that when the tire force/torque is > reported as zero, the wheel spindle torque can sometimes be non-negligible > (on the order of hundreds in magnitude) and causes the wheel to rotate > beyond its limit. It's like the applied torque to the spindle is as if the > tire is in collision, but the collision hasn't occurred yet. That's my best > guess. > > > > I tried changing some of the collision parameters that I am aware of, but > this had no effect whatsoever. Any help would be greatly appreciated. > > > > Ted > > > > > > On Sun, Jan 12, 2025 at 1:35 PM Ted Sender <[email protected]> wrote: > > Hi Dan, > > > > I realized there was no video attached. The system wouldn't let me attach > it for some reason and it said someone had to approve the post before I > could edit it. Here is the video. Let me know if you have trouble viewing > it. > > > > I'd like to use SCM terrain, but I'm using someone else's setup that they > provided to me and I have no idea how to get SCM to work. > > > > Ted > > > > > > On Sun, Jan 12, 2025 at 10:19 AM Dan Negrut <[email protected]> wrote: > > Ted – there was no video attached… > > > > One thing that might be worth trying is to use SCM deformable terrain. If > you have a couple of hours, give it a try – folks in our lab have had good > success with that. > > > > Dan > > --------------------------------------------- > > Bernard A. and Frances M. Weideman Professor > > NVIDIA CUDA Fellow > > Department of Mechanical Engineering > > Department of Computer Science > > University of Wisconsin - Madison > > 4150ME, 1513 University Avenue > > Madison, WI 53706-1572 > > 608 772 0914 > > http://sbel.wisc.edu/ > > http://projectchrono.org/ > <https://urldefense.com/v3/__http:/projectchrono.org/__;!!Mak6IKo!NU3uC2Icpilhf0GjZfHYjvyr6D50Ysb-UYoG53uZBAOd6g15I0cEHV6VwpRSo9bD-toG_vkG0mdKTb1H$> > > --------------------------------------------- > > > > *From:* [email protected] <[email protected]> *On > Behalf Of *Ted Sender > *Sent:* Saturday, January 11, 2025 10:32 PM > *To:* ProjectChrono <[email protected]> > *Subject:* [chrono] Tires Rotating Beyond Mechanical Limits > > > > Hello, > > > > I am currently using Chrono for vehicle simulations with a polaris > vehicle. I am using rigid tires on rigid terrain. I am noticing that with > certain terrain profiles one or both of the front tires rotate very quickly > beyond its mechanical limit when the vehicle drives over a harsh bump/dip. > > > > These are the current settings I am using: > > Solver: Barzilai-Borwein, tolerance 1e-6, max iters 400, time step 1 ms > > Collision: Bullet, default suggested envelope 1 mm, default suggested > margin 1 mm > > Terrain: Rigid, coeff friction 0.6 > > > > I have tried using different solvers, but they did not seem to work (some > just failed to perform the calculations or crashed the program). I tried > lowering the time step from 1 ms to 0.2 ms, but that did not work either. > > > > Attached is a video showing an example of the problem I am having. Right > around t = 6 seconds, the front right wheel drives over a dip and then the > wheel suddenly rotates clockwise well past its mechanical limit (the limit > is about 0.48 radians, but the wheel rotates like 2 radians). > > > > I also logged some force/moment values to the terminal to see what was > going on, thinking that it could have been a problem with the tire forces. > Attached is a .txt file showing the tire's steering angle, tire force and > torque values, and the applied force and torque to that wheel's spindle (I > believe all forces/moments are expressed in the global frame). > > > > From this data, it seems that when the tire has no reported force/torque, > the spindle still has some applied force/torque. At t = 5.1 seconds, this > occurs, but is not a problem as the spindle torque is rather small. > However, at t = 5.90 seconds, the tire force is zero but the spindle torque > starts to increase and then the wheel rotates beyond its limit. > > > > Does this seem like a problem with the collision model/settings or is it > caused by something else? Any help is greatly appreciated. I hope the data > provided is helpful, but if you need some more values, let me know. > > > > Thanks, > > Ted > > -- > You received this message because you are subscribed to the Google Groups > "ProjectChrono" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion visit > https://groups.google.com/d/msgid/projectchrono/38e5d8a2-5ccc-47bf-93cb-bb0f6ff6e327n%40googlegroups.com > <https://urldefense.com/v3/__https:/groups.google.com/d/msgid/projectchrono/38e5d8a2-5ccc-47bf-93cb-bb0f6ff6e327n*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!IWB4NaZMiARYuUywzXWMFmkLZzGIyuN5qnFl3Q6IJPyX-E8W4TJaXJwRH3RrQod2PHaq9IZU0O17gy62$> > . > > -- You received this message because you are subscribed to the Google Groups "ProjectChrono" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/projectchrono/CA%2Byw%3DfrxK71%2BKvWxre9_qWfsQG2KzM%2BOc%2Biitu6%3DwBs_N6xodw%40mail.gmail.com.
