Real-Time Physics Simulation Forum

Could you provide some simple example? thanks.

after changing to tag 2.82(git checkout 2.82), i find the GimpactTestDemo. i don't know why it is deleted in the later version. hope it is useful for future user.

add a picture to clarify this case. any ideas about the gimpact mesh is helpful, thanks in advance.

I have also tried using btGImpactMeshShape, with no change in results. Could you post some code of how did you use the btGImpactMeshShape? i will try it later. i find the algorithm for the gimpact need to be regiestered externally like below. btCollisionDispatcher * dispatcher = static_cast<btColli...

your two collison objects all are the concave but now the engine doesn't support the concave vs. convave collision in the narrowphase. you can refer to this function btDefaultCollisionConfiguration::getCollisionAlgorithmCreateFunc() in file btdefaultcollisionconfiguration.cpp for details. i think yo...

i just write a similar test but use two boxes instead of the concave and don't find the issue you said. so next i need the data to create the concave shape. thanks. I know from other posts on this forum that a contact manifold is created in the broadphase based on AABB intersection and may reasonabl...

i would like to work with you to find the rootcause if you can provide the sample code.

Just for your reference.
1. site-packages/pybullet_envs$ python3.5 -m agents.train_ppo --logdir=./ --config=pybullet_humanoid
2. use the tf.train.NewCheckpointReader(checkpoint_path) to get the parameters that you need.

Moreover, the fact that resetBaseVelocity overrides was confirmed by erwin coumans here: https://github.com/bulletphysics/bullet3/issues/1845#issuecomment-463617025 my understanding is that we shouldn't use the resetBaseVelocity () during a running simulation, but can be used freely before stepsimu...

The way I've temporarily got around this issue is by calling resetBaseVelocity() , although I'm aware that as Erwin stated, these reset commands will override the effects of the physics simulation. hi arpastrana, i understand calling the resetBasePositionAndOrientation() will ignore the dynamic of ...

it is indeed strange, i think your case needs to debug step by step. i would like to debug with you if you can provide a simple test.

i find many parameters like weights_dense1_w in the example, these parameters should be the result trained by other modules. i want to know where can i find these modules? Any other related document is also helpful. thanks.

did you try the Cluster? maybe it can help to make your test work well.

//only self-collision for Cluster, not Vertex-Face yet from the comment, it seems that it only support self-collision for Cluster. Could you try these interfaces to enable the cluster? pSoftBody->m_cfg.collisions = btSoftBody::fCollision::CL_SS | btSoftBody::fCollision::CL_RS | btSoftBody::fCollisi...

hi PcChip,

please add this interface below which should can fix your issue.