Rigid Body Puns

Hello there. Before asking for advice, a little context (and maybe a little oversharing). I did 0,1g of mazatapec before a chill biking route and wasn´t expecting much more than a general feeling of wellbeing. I definitely got it, but also an unusual energy and social ease. I was so focused on the ride and the nature, being careful not to fall that I didn´t notice anything until a resting point. I couldn´t stop thinking ab how grateful I was/am for sharing that ride with my friends, for the city I´m living in right now for letting me be this close to nature. Mostly and unexpectedly, I felt sooo grateful for my body. Lately I had been feeling body conscious, looking after my diet, etc. that i had not stopped to think how much i love just being physically active (aerials, yoga, running, etc) and challenging myself not to "look better", but to get and feel better, if that makes sense.

I could use help or opinions on what comes next. I´m no longer sure of MDing on a regular basis as I had planned, since the setting and activity meant a lot for me. Been on this sub for a bit and I see most people follow a schedule for some time to see improvements on mental health and wellbeing. Could occasional or non-rigid use have the same effects? Also, what dosages work for you if you enjoy sports or just physical activity and MDing? Cheers and thanks for reading

[EDIT]

Thanks for sharing your experiences. I read all of them and I'm just glad everyone has a different routine and activities working for them

Abstract

We present Brax, an open source library for rigid body simulation with a focus on performance and parallelism on accelerators, written in JAX. We present results on a suite of tasks inspired by the existing reinforcement learning literature, but remade in our engine. Additionally, we provide reimplementations of PPO, SAC, ES, and direct policy optimization in JAX that compile alongside our environments, allowing the learning algorithm and the environment processing to occur on the same device, and to scale seamlessly on accelerators. Finally, we include notebooks that facilitate training of performant policies on common OpenAI Gym MuJoCo-like tasks in minutes.

https://arxiv.org/abs/2106.13281

GitHub repo description

Brax is a differentiable physics engine that simulates environments made up of rigid bodies, joints, and actuators. It's also a suite of learning algorithms to train agents to operate in these environments (PPO, SAC, evolutionary strategy, and direct trajectory optimization are implemented).

Brax is written in JAX and is designed for use on acceleration hardware. It is both efficient for single-core training, and scalable to massively parallel simulation, without the need for pesky datacenters.

https://github.com/google/brax

I'm interested in performing high fidelity modeling & simulation of an aircraft about an oblate, rotating earth. Does anyone have a good reference that derives the equations of motion for oblate earth that is easy to understand? I find that Stevens & Lewis, which is a common reference, is fairly disjointed in its explanation and lacks derivations. I know many references that do a good job of deriving the equations for flat earth, but I'd like to simulate at a high fidelity.

Dear all,

I hope this is the correct subreddit to post this.

I have been self-learning rigid-body kinematics and dynamics over the past year since I am building a motion simulator for land vehicles, but there is something I cannot quite figure out after consulting several textbooks (because it was never explicitly explained, and I am a bit of a dullard).

-----

Say there is a rigid body with body coordinates centered at its center-of-mass, C, with external force vector F acting on C, and global coordinate system, G.

Let

• (u,v):= longitudinal and lateral velocities of C in body coordinates
• (X,Y):= global coordinates
• r := rotation of C around the z-z axis
• (fx,fy):= external forces acting on C in body coordinates
• (Fx,Fy):= external forces acting on C in global coordinates
• T:= torque acting on rigid body
• k1d:= 1st time derivative of arbitrary variable k, e.g. X1d:= velocity in global coordinates
• m:= mass of C
• Izz:= moment of inertia of C around z-z axis
• (a_i,b_i):= x and y lever arms acted upon by sub-components fx_i, fy_i acting at a distance a_i and b_i (each, respectively).
  • fx = sum(fx_i)
  • fy = sum(fy_i)

Using force balance in the body coordinate, the equations of motion works out to:

m(u1d - r1d*v) = fx

m(v1d + r1d*u) = fy

Izz*r2d = sum(a_i*fx_i) + sum(b_i*fy_i)

(assume for Izz*r2d that a_i and b_i have the proper sign conventions)

-----

Now, I would like to compute the global displacement of C when acted upon by fx and fy, calculated in body coordinates. To do that, I would need to specify a system of ODEs and apply an ODE solver to it.

Let [K] be the state space vector with elements K(i) ∈[K] for i = 1,2,...k .

From my readings, I get the impression that one would specify the gradient space, [dK/dt] as follows:

1. X1d = K(4)*cos(K(3)) - K(5)*sin(K(3))
2. Y1d = K(4)*sin(K(3)) + K(5)*cos(K(3))
3. r1d = K(6)
4. u1d = 1/m*fx + r1d*v
5. v1d = 1/m*fy - r1d*u
6. r2d = 1/Izz*(sum(a_i*fx_i) + sum(b_i*fy_i))

Now here are my questions:

1. Does that sound right to you? Am I missing something? When I ran my simulations, something doesn't seem to add up, and I am trying to eliminate algebraic or trigonometric error as a cause.
2. I cannot seem to get my head around the significance of r1d*v and r1d*u, since they are "fictitious forces" due to the rotation of the body-centered coordinates. Does that mean u1d and v1d is the "absolute acceleration" of C? i.e. if an acceleromet

Hello, I am making a 3D Demo with a player character gun that instances bullets. I've been running into a problem where the spawned node spawns and goes in directions other than where my raycast is pointing (example)

Here is the script in the player's physics process:

if Input.is_action_just_pressed("fire"):
		if aimCast.is_colliding():
			var livebullet = bullet.instance()
			muzzle.add_child(livebullet)
			livebullet.look_at(aimCast.get_collision_point(), Vector3.UP)
			livebullet.shoot = true
#			livebullet.apply_impulse(transform.basis.z, -transform.basis.z * 10)

and here is the bullet script:

     extends RigidBody

var SPEED = 5
const DAMAGE = 50

var timer = 0

var hit_something = false
var shoot = false
var direction = Vector3()

func _ready():
	set_as_toplevel(true)

func _physics_process(delta):
	pass
	if shoot:
	   apply_impulse(transform.basis.z, -transform.basis.z)

I commented out the apply impulse in the physics process script, however, I've found that if I do that, the rigidbody will fly in the correct direction, however, it will not go any direction other than -Z (directly forward, no up or down regardless of where I'm pointing)

Any help would be appreciated, thanks!

Collision point*

https://reddit.com/link/o08m7b/video/awufe00rtd571/player

I have an idea to directly import the 3d model as gltf (with or without rigging) and then attaching the rigid body nodes where it needs ragdoll like physics (like in the hands of the model or in the body) attached together by pin joints(hinges are good but I found pin joints more appropriate).

While the only constraint is that the model has to be seperated into meshes corresponding the different parts where the physics or rigid bodies would be applied. For example, if I want to apply physics on hands then my hands should be a seperate mesh so that it could be parented to rigid body node.

Although this approach might be time consuming but IMO this can produce better ragdoll results and if we make parent nodes according to bone inheritance, it can produce active ragdoll like physics.

I am not saying that current ragdoll support is not good , but it is buggy and hard for newbies to use it.

The only problem is that I will not bend the mesh like the physical bones ( it only turns the mesh thus producing ragged joints).

Please tell your opinion guys...

In order to make the object appear at a frame in the middle, it seems I have to select Animated under Rigid Body settings but then it won't move.