jaxsim 0.4.1.dev11__tar.gz → 0.4.1.dev15__tar.gz

{jaxsim-0.4.1.dev11 → jaxsim-0.4.1.dev15}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: jaxsim
-Version: 0.4.1.dev11
+Version: 0.4.1.dev15
 Summary: A differentiable physics engine and multibody dynamics library for control and robot learning.
 Author-email: Diego Ferigo <dgferigo@gmail.com>
 Maintainer-email: Diego Ferigo <dgferigo@gmail.com>, Filippo Luca Ferretti <filippo.ferretti@iit.it>

{jaxsim-0.4.1.dev11 → jaxsim-0.4.1.dev15}/src/jaxsim/_version.py

@@ -12,5 +12,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '0.4.1.dev11'
-__version_tuple__ = version_tuple = (0, 4, 1, 'dev11')
+__version__ = version = '0.4.1.dev15'
+__version_tuple__ = version_tuple = (0, 4, 1, 'dev15')

{jaxsim-0.4.1.dev11 → jaxsim-0.4.1.dev15}/src/jaxsim/integrators/common.py

@@ -5,11 +5,12 @@ from typing import Any, ClassVar, Generic, Protocol, Type, TypeVar
 import jax
 import jax.numpy as jnp
 import jax_dataclasses
-import jaxlie
 from jax_dataclasses import Static
 
 import jaxsim.api as js
+import jaxsim.math
 import jaxsim.typing as jtp
+from jaxsim import exceptions
 from jaxsim.utils.jaxsim_dataclass import JaxsimDataclass, Mutability
 
 try:
@@ -539,48 +540,38 @@ class ExplicitRungeKuttaSO3Mixin:
     `PyTreeType = ODEState` to integrate the quaternion on SO(3).
     """
 
-    @classmethod
-    def integrate_rk_stage(
-        cls, x0: js.ode_data.ODEState, t0: Time, dt: TimeStep, k: js.ode_data.ODEState
-    ) -> js.ode_data.ODEState:
-
-        op = lambda x0_leaf, k_leaf: x0_leaf + dt * k_leaf
-        xf: js.ode_data.ODEState = jax.tree_util.tree_map(op, x0, k)
-
-        W_Q_B_tf = xf.physics_model.base_quaternion
-
-        return xf.replace(
-            physics_model=xf.physics_model.replace(
-                base_quaternion=W_Q_B_tf / jnp.linalg.norm(W_Q_B_tf)
-            )
-        )
-
     @classmethod
     def post_process_state(
         cls, x0: js.ode_data.ODEState, t0: Time, xf: js.ode_data.ODEState, dt: TimeStep
     ) -> js.ode_data.ODEState:
 
-        # Indices to convert quaternions between serializations.
-        to_xyzw = jnp.array([1, 2, 3, 0])
+        # Extract the initial base quaternion.
+        W_Q_B_t0 = x0.physics_model.base_quaternion
 
-        # Get the initial rotation.
-        W_R_B_t0 = jaxlie.SO3.from_quaternion_xyzw(
-            xyzw=x0.physics_model.base_quaternion[to_xyzw]
+        # We assume that the initial quaternion is already unary.
+        exceptions.raise_runtime_error_if(
+            condition=jnp.logical_not(jnp.allclose(W_Q_B_t0.dot(W_Q_B_t0), 1.0)),
+            msg="The SO(3) integrator received a quaternion at t0 that is not unary.",
         )
 
-        # Get the final angular velocity.
-        # This is already computed by averaging the kᵢ in RK-based schemes.
-        # Therefore, by using the ω at tf, we obtain a RK scheme operating
-        # on the SO(3) manifold.
-        W_ω_WB_tf = xf.physics_model.base_angular_velocity
-
-        # Integrate the orientation on SO(3).
-        # Note that we left-multiply with the exponential map since the angular
-        # velocity is expressed in the inertial frame.
-        W_R_B_tf = jaxlie.SO3.exp(tangent=dt * W_ω_WB_tf) @ W_R_B_t0
+        # Get the angular velocity ω to integrate the quaternion.
+        # This velocity ω[t0] is computed in the previous timestep by averaging the kᵢ
+        # corresponding to the active RK-based scheme. Therefore, by using the ω[t0],
+        # we obtain an explicit RK scheme operating on the SO(3) manifold.
+        # Note that the current integrator is not a semi-implicit scheme, therefore
+        # using the final ω[tf] would be not correct.
+        W_ω_WB_t0 = x0.physics_model.base_angular_velocity
+
+        # Integrate the quaternion on SO(3).
+        W_Q_B_tf = jaxsim.math.Quaternion.integration(
+            quaternion=W_Q_B_t0,
+            dt=dt,
+            omega=W_ω_WB_t0,
+            omega_in_body_fixed=False,
+        )
 
         # Replace the quaternion in the final state.
         return xf.replace(
-            physics_model=xf.physics_model.replace(base_quaternion=W_R_B_tf.wxyz),
+            physics_model=xf.physics_model.replace(base_quaternion=W_Q_B_tf),
             validate=True,
         )

{jaxsim-0.4.1.dev11 → jaxsim-0.4.1.dev15}/src/jaxsim/rbda/utils.py

@@ -2,6 +2,7 @@ import jax.numpy as jnp
 
 import jaxsim.api as js
 import jaxsim.typing as jtp
+from jaxsim import exceptions
 from jaxsim.math import StandardGravity
 
 
@@ -131,6 +132,13 @@ def process_inputs(
     if W_Q_B.shape != (4,):
         raise ValueError(W_Q_B.shape, (4,))
 
+    # Check that the quaternion is unary since our RBDAs make this assumption in order
+    # to prevent introducing additional normalizations that would affect AD.
+    exceptions.raise_value_error_if(
+        condition=jnp.logical_not(jnp.allclose(W_Q_B.dot(W_Q_B), 1.0)),
+        msg="A RBDA received a quaternion that is not normalized.",
+    )
+
     # Pack the 6D base velocity and acceleration.
     W_v_WB = jnp.hstack([W_vl_WB, W_ω_WB])
     W_v̇_WB = jnp.hstack([W_v̇l_WB, W_ω̇_WB])

{jaxsim-0.4.1.dev11 → jaxsim-0.4.1.dev15}/src/jaxsim.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: jaxsim
-Version: 0.4.1.dev11
+Version: 0.4.1.dev15
 Summary: A differentiable physics engine and multibody dynamics library for control and robot learning.
 Author-email: Diego Ferigo <dgferigo@gmail.com>
 Maintainer-email: Diego Ferigo <dgferigo@gmail.com>, Filippo Luca Ferretti <filippo.ferretti@iit.it>

{jaxsim-0.4.1.dev11 → jaxsim-0.4.1.dev15}/tests/test_automatic_differentiation.py

@@ -93,7 +93,7 @@ def test_ad_aba(
     aba = lambda W_p_B, W_Q_B, s, W_v_WB, ṡ, τ, W_f_L, g: jaxsim.rbda.aba(
         model=model,
         base_position=W_p_B,
-        base_quaternion=W_Q_B,
+        base_quaternion=W_Q_B / jnp.linalg.norm(W_Q_B),
         joint_positions=s,
         base_linear_velocity=W_v_WB[0:3],
         base_angular_velocity=W_v_WB[3:6],
@@ -150,7 +150,7 @@ def test_ad_rnea(
     rnea = lambda W_p_B, W_Q_B, s, W_v_WB, ṡ, W_v̇_WB, s̈, W_f_L, g: jaxsim.rbda.rnea(
         model=model,
         base_position=W_p_B,
-        base_quaternion=W_Q_B,
+        base_quaternion=W_Q_B / jnp.linalg.norm(W_Q_B),
         joint_positions=s,
         base_linear_velocity=W_v_WB[0:3],
         base_angular_velocity=W_v_WB[3:6],
@@ -229,7 +229,7 @@ def test_ad_fk(
     fk = lambda W_p_B, W_Q_B, s: jaxsim.rbda.forward_kinematics_model(
         model=model,
         base_position=W_p_B,
-        base_quaternion=W_Q_B,
+        base_quaternion=W_Q_B / jnp.linalg.norm(W_Q_B),
         joint_positions=s,
     )