update deeproboticslab changes

source/robot_lab/robot_lab/assets/deeprobotics.py

@@ -2,7 +2,7 @@
 # SPDX-License-Identifier: Apache-2.0
 
 import isaaclab.sim as sim_utils
-from isaaclab.actuators import DCMotorCfg
+from isaaclab.actuators import DCMotorCfg, DelayedPDActuatorCfg
 from isaaclab.assets.articulation import ArticulationCfg
 
 from robot_lab.assets import ISAACLAB_ASSETS_DATA_DIR
@@ -11,7 +11,7 @@
     spawn=sim_utils.UrdfFileCfg(
         fix_base=False,
         merge_fixed_joints=True,
-        replace_cylinders_with_capsules=False,
+        replace_cylinders_with_capsules=True,
         asset_path=f"{ISAACLAB_ASSETS_DATA_DIR}/Robots/deeprobotics/lite3_description/urdf/lite3.urdf",
         activate_contact_sensors=True,
         rigid_props=sim_utils.RigidBodyPropertiesCfg(
@@ -24,7 +24,7 @@
             max_depenetration_velocity=1.0,
         ),
         articulation_props=sim_utils.ArticulationRootPropertiesCfg(
-            enabled_self_collisions=False, solver_position_iteration_count=4, solver_velocity_iteration_count=0
+            enabled_self_collisions=False, solver_position_iteration_count=4, solver_velocity_iteration_count=1
         ),
         joint_drive=sim_utils.UrdfConverterCfg.JointDriveCfg(
             gains=sim_utils.UrdfConverterCfg.JointDriveCfg.PDGainsCfg(stiffness=0, damping=0)
@@ -41,23 +41,25 @@
     ),
     soft_joint_pos_limit_factor=0.9,
     actuators={
-        "Hip": DCMotorCfg(
+        "Hip": DelayedPDActuatorCfg(
             joint_names_expr=[".*_Hip[X,Y]_joint"],
             effort_limit=24.0,
-            saturation_effort=24.0,
             velocity_limit=26.2,
             stiffness=30.0,
-            damping=0.5,
+            damping=1.0,
             friction=0.0,
+            min_delay=0,  # physics time steps (min: 5.0*0=00.0ms)
+            max_delay=5,  # physics time steps (max: 5.0*5=25.0ms)
         ),
-        "Knee": DCMotorCfg(
+        "Knee": DelayedPDActuatorCfg(
             joint_names_expr=[".*_Knee_joint"],
             effort_limit=36.0,
-            saturation_effort=36.0,
             velocity_limit=17.3,
             stiffness=30.0,
-            damping=0.5,
+            damping=1.0,
             friction=0.0,
+            min_delay=0,  # physics time steps (min: 5.0*0=00.0ms)
+            max_delay=5,  # physics time steps (max: 5.0*5=25.0ms)
         ),
     },
 )
@@ -79,7 +81,7 @@
             max_depenetration_velocity=1.0,
         ),
         articulation_props=sim_utils.ArticulationRootPropertiesCfg(
-            enabled_self_collisions=False, solver_position_iteration_count=4, solver_velocity_iteration_count=0
+            enabled_self_collisions=False, solver_position_iteration_count=4, solver_velocity_iteration_count=1
         ),
         joint_drive=sim_utils.UrdfConverterCfg.JointDriveCfg(
             gains=sim_utils.UrdfConverterCfg.JointDriveCfg.PDGainsCfg(stiffness=0, damping=0)

source/robot_lab/robot_lab/tasks/manager_based/locomotion/velocity/config/quadruped/deeprobotics_lite3/rough_env_cfg.py

@@ -82,8 +82,8 @@ def __post_init__(self):
         # Root penalties
         self.rewards.lin_vel_z_l2.weight = -2.0
         self.rewards.ang_vel_xy_l2.weight = -0.05
-        self.rewards.flat_orientation_l2.weight = 0
-        self.rewards.base_height_l2.weight = 0
+        self.rewards.flat_orientation_l2.weight = -5.0
+        self.rewards.base_height_l2.weight = -10.0
         self.rewards.base_height_l2.params["target_height"] = 0.35
         self.rewards.base_height_l2.params["asset_cfg"].body_names = [self.base_link_name]
         self.rewards.body_lin_acc_l2.weight = 0
@@ -92,50 +92,52 @@ def __post_init__(self):
         # Joint penalties
         self.rewards.joint_torques_l2.weight = -2.5e-5
         self.rewards.joint_vel_l2.weight = 0
-        self.rewards.joint_acc_l2.weight = -2.5e-7
-        # self.rewards.create_joint_deviation_l1_rewterm("joint_deviation_hip_l1", -0.2, [".*_hip_joint"])
+        self.rewards.joint_acc_l2.weight = -1e-8
+        self.rewards.create_joint_deviation_l1_rewterm("joint_deviation_hip_l1", -0.5, [".*HipX.*"])
         self.rewards.joint_pos_limits.weight = -5.0
         self.rewards.joint_vel_limits.weight = 0
         self.rewards.joint_power.weight = -2e-5
-        self.rewards.stand_still.weight = -2.0
-        self.rewards.joint_pos_penalty.weight = -1.0
-        self.rewards.joint_mirror.weight = -0.05
+        self.rewards.stand_still.weight = -0.3
+        self.rewards.joint_pos_penalty.weight = 0
+        self.rewards.joint_mirror.weight = 0
         self.rewards.joint_mirror.params["mirror_joints"] = [
             ["FL_(HipX|HipY|Knee).*", "HR_(HipX|HipY|Knee).*"],
             ["FR_(HipX|HipY|Knee).*", "HL_(HipX|HipY|Knee).*"],
         ]
 
         # Action penalties
-        self.rewards.action_rate_l2.weight = -0.01
+        self.rewards.action_rate_l2.weight = -0.02
 
         # Contact sensor
-        self.rewards.undesired_contacts.weight = -1.0
+        self.rewards.undesired_contacts.weight = -0.5
         self.rewards.undesired_contacts.params["sensor_cfg"].body_names = [f"^(?!.*{self.foot_link_name}).*"]
-        self.rewards.contact_forces.weight = -1.5e-4
+        self.rewards.contact_forces.weight = -1e-2
         self.rewards.contact_forces.params["sensor_cfg"].body_names = [self.foot_link_name]
 
         # Velocity-tracking rewards
-        self.rewards.track_lin_vel_xy_exp.weight = 3.0
-        self.rewards.track_ang_vel_z_exp.weight = 1.5
+        self.rewards.track_lin_vel_xy_exp.weight = 1.2
+        self.rewards.track_ang_vel_z_exp.weight = 0.6
 
         # Others
-        self.rewards.feet_air_time.weight = 0
+        self.rewards.feet_air_time.weight = 1.0
         self.rewards.feet_air_time.params["threshold"] = 0.5
         self.rewards.feet_air_time.params["sensor_cfg"].body_names = [self.foot_link_name]
+        self.rewards.feet_air_time_variance.weight = -4.0
+        self.rewards.feet_air_time_variance.params["sensor_cfg"].body_names = [self.foot_link_name]
         self.rewards.feet_contact.weight = 0
         self.rewards.feet_contact.params["sensor_cfg"].body_names = [self.foot_link_name]
-        self.rewards.feet_contact_without_cmd.weight = 0.1
+        self.rewards.feet_contact_without_cmd.weight = 0
         self.rewards.feet_contact_without_cmd.params["sensor_cfg"].body_names = [self.foot_link_name]
         self.rewards.feet_stumble.weight = 0
         self.rewards.feet_stumble.params["sensor_cfg"].body_names = [self.foot_link_name]
-        self.rewards.feet_slide.weight = 0
+        self.rewards.feet_slide.weight = -0.05
         self.rewards.feet_slide.params["sensor_cfg"].body_names = [self.foot_link_name]
         self.rewards.feet_slide.params["asset_cfg"].body_names = [self.foot_link_name]
-        self.rewards.feet_height.weight = 0
+        self.rewards.feet_height.weight = -0.2
         self.rewards.feet_height.params["target_height"] = 0.05
         self.rewards.feet_height.params["asset_cfg"].body_names = [self.foot_link_name]
-        self.rewards.feet_height_body.weight = 0
-        self.rewards.feet_height_body.params["target_height"] = -0.25
+        self.rewards.feet_height_body.weight = -2.5
+        self.rewards.feet_height_body.params["target_height"] = -0.35
         self.rewards.feet_height_body.params["asset_cfg"].body_names = [self.foot_link_name]
         self.rewards.feet_gait.weight = 0
         self.rewards.feet_gait.params["synced_feet_pair_names"] = (("FL_FOOT", "HR_FOOT"), ("FR_FOOT", "HL_FOOT"))
@@ -154,6 +156,6 @@ def __post_init__(self):
         self.curriculum.command_levels = None
 
         # ------------------------------Commands------------------------------
-        # self.commands.base_velocity.ranges.lin_vel_x = (-1.5, 1.5)
-        # self.commands.base_velocity.ranges.lin_vel_y = (-0.8, 0.8)
-        # self.commands.base_velocity.ranges.ang_vel_z = (-1.5, 1.5)
+        self.commands.base_velocity.ranges.lin_vel_x = (-1.5, 1.5)
+        self.commands.base_velocity.ranges.lin_vel_y = (-0.8, 0.8)
+        self.commands.base_velocity.ranges.ang_vel_z = (-1.5, 1.5)

source/robot_lab/robot_lab/tasks/manager_based/locomotion/velocity/mdp/rewards.py

@@ -112,7 +112,7 @@ def joint_pos_penalty(
     """Penalize joint position error from default on the articulation."""
     # extract the used quantities (to enable type-hinting)
     asset: Articulation = env.scene[asset_cfg.name]
-    cmd = torch.linalg.norm(env.command_manager.get_command(command_name), dim=1)
+    cmd = torch.norm(env.command_manager.get_command(command_name)[:, :2], dim=1)
     body_vel = torch.linalg.norm(asset.data.root_lin_vel_b[:, :2], dim=1)
     running_reward = torch.linalg.norm(
         (asset.data.joint_pos[:, asset_cfg.joint_ids] - asset.data.default_joint_pos[:, asset_cfg.joint_ids]), dim=1
@@ -135,7 +135,7 @@ def wheel_vel_penalty(
     asset_cfg: SceneEntityCfg = SceneEntityCfg("robot"),
 ) -> torch.Tensor:
     asset: Articulation = env.scene[asset_cfg.name]
-    cmd = torch.linalg.norm(env.command_manager.get_command(command_name), dim=1)
+    cmd = torch.norm(env.command_manager.get_command(command_name)[:, :2], dim=1)
     body_vel = torch.linalg.norm(asset.data.root_lin_vel_b[:, :2], dim=1)
     joint_vel = torch.abs(asset.data.joint_vel[:, asset_cfg.joint_ids])
     contact_sensor: ContactSensor = env.scene.sensors[sensor_cfg.name]
@@ -404,7 +404,7 @@ def feet_contact(
     contact_num = torch.sum(contact, dim=1)
     reward = (contact_num != expect_contact_num).float()
     # no reward for zero command
-    reward *= torch.linalg.norm(env.command_manager.get_command(command_name), dim=1) > 0.1
+    reward *= torch.norm(env.command_manager.get_command(command_name)[:, :2], dim=1) > 0.1
     reward *= torch.clamp(-env.scene["robot"].data.projected_gravity_b[:, 2], 0, 0.7) / 0.7
     return reward
 
@@ -416,7 +416,7 @@ def feet_contact_without_cmd(env: ManagerBasedRLEnv, command_name: str, sensor_c
     # compute the reward
     contact = contact_sensor.compute_first_contact(env.step_dt)[:, sensor_cfg.body_ids]
     reward = torch.sum(contact, dim=-1).float()
-    reward *= torch.linalg.norm(env.command_manager.get_command(command_name), dim=1) < 0.1
+    reward *= torch.norm(env.command_manager.get_command(command_name)[:, :2], dim=1) < 0.1
     reward *= torch.clamp(-env.scene["robot"].data.projected_gravity_b[:, 2], 0, 0.7) / 0.7
     return reward
 
@@ -515,7 +515,7 @@ def feet_height(
     )
     reward = torch.sum(foot_z_target_error * foot_velocity_tanh, dim=1)
     # no reward for zero command
-    reward *= torch.linalg.norm(env.command_manager.get_command(command_name), dim=1) > 0.1
+    reward *= torch.norm(env.command_manager.get_command(command_name)[:, :2], dim=1) > 0.1
     reward *= torch.clamp(-env.scene["robot"].data.projected_gravity_b[:, 2], 0, 0.7) / 0.7
     return reward
 
@@ -545,7 +545,7 @@ def feet_height_body(
     foot_z_target_error = torch.square(footpos_in_body_frame[:, :, 2] - target_height).view(env.num_envs, -1)
     foot_velocity_tanh = torch.tanh(tanh_mult * torch.norm(footvel_in_body_frame[:, :, :2], dim=2))
     reward = torch.sum(foot_z_target_error * foot_velocity_tanh, dim=1)
-    reward *= torch.linalg.norm(env.command_manager.get_command(command_name), dim=1) > 0.1
+    reward *= torch.norm(env.command_manager.get_command(command_name)[:, :2], dim=1) > 0.1
     reward *= torch.clamp(-env.scene["robot"].data.projected_gravity_b[:, 2], 0, 0.7) / 0.7
     return reward
 

source/robot_lab/robot_lab/tasks/manager_based/locomotion/velocity/velocity_env_cfg.py

@@ -311,7 +311,7 @@ class EventCfg:
         mode="startup",
         params={
             "asset_cfg": SceneEntityCfg("robot", body_names=".*"),
-            "com_range": {"x": (-0.05, 0.05), "y": (-0.05, 0.05), "z": (-0.05, 0.05)},
+            "com_range": {"x": (-0.05, 0.05), "y": (-0.05, 0.05), "z": (-0.05, 0.05)},  # TODO: move to config
         },
     )
 
@@ -344,7 +344,7 @@ class EventCfg:
             "stiffness_distribution_params": (0.5, 2.0),
             "damping_distribution_params": (0.5, 2.0),
             "operation": "scale",
-            "distribution": "log_uniform",
+            "distribution": "uniform",
         },
     )