WIP: RL environment for the surface gripper

source/isaaclab_assets/isaaclab_assets/robots/universal_robots.py

@@ -113,7 +113,7 @@
     "shoulder_lift_joint": -1.5707,
     "elbow_joint": 1.5707,
     "wrist_1_joint": -1.5707,
-    "wrist_2_joint": 1.5707,
+    "wrist_2_joint": -1.5707,
     "wrist_3_joint": 0.0,
 }
 

source/isaaclab_tasks/isaaclab_tasks/manager_based/manipulation/lift/config/ur10_SG/__init__.py

@@ -0,0 +1,42 @@
+# Copyright (c) 2022-2025, The Isaac Lab Project Developers (https://github.com/isaac-sim/IsaacLab/blob/main/CONTRIBUTORS.md).
+# All rights reserved.
+#
+# SPDX-License-Identifier: BSD-3-Clause
+import gymnasium as gym
+import os
+
+from . import agents
+
+##
+# Register Gym environments.
+##
+
+##
+# Joint Position Control
+##
+
+gym.register(
+    id="Isaac-Lift-Cube-UR10SG-Short-v0",
+    entry_point="isaaclab.envs:ManagerBasedRLEnv",
+    kwargs={
+        "env_cfg_entry_point": f"{__name__}.joint_pos_env_cfg:UR10SGCubeLiftEnvCfg",
+        "rsl_rl_cfg_entry_point": f"{agents.__name__}.rsl_rl_ppo_cfg:UR10LiftCubePPORunnerCfg",
+        "skrl_cfg_entry_point": f"{agents.__name__}:skrl_ppo_cfg.yaml",
+        "rl_games_cfg_entry_point": f"{agents.__name__}:rl_games_ppo_cfg.yaml",
+        "sb3_cfg_entry_point": f"{agents.__name__}:sb3_ppo_cfg.yaml",
+    },
+    disable_env_checker=True,
+)
+
+gym.register(
+    id="Isaac-Lift-Cube-UR10SG-Long-v0",
+    entry_point="isaaclab.envs:ManagerBasedRLEnv",
+    kwargs={
+        "env_cfg_entry_point": f"{__name__}.joint_pos_env_cfg:UR10SGLongCubeLiftEnvCfg",
+        "rsl_rl_cfg_entry_point": f"{agents.__name__}.rsl_rl_ppo_cfg:UR10LiftCubePPORunnerCfg",
+        "skrl_cfg_entry_point": f"{agents.__name__}:skrl_ppo_cfg.yaml",
+        "rl_games_cfg_entry_point": f"{agents.__name__}:rl_games_ppo_cfg.yaml",
+        "sb3_cfg_entry_point": f"{agents.__name__}:sb3_ppo_cfg.yaml",
+    },
+    disable_env_checker=True,
+)

source/isaaclab_tasks/isaaclab_tasks/manager_based/manipulation/lift/config/ur10_SG/agents/__init__.py

@@ -0,0 +1,4 @@
+# Copyright (c) 2022-2025, The Isaac Lab Project Developers (https://github.com/isaac-sim/IsaacLab/blob/main/CONTRIBUTORS.md).
+# All rights reserved.
+#
+# SPDX-License-Identifier: BSD-3-Clause

source/isaaclab_tasks/isaaclab_tasks/manager_based/manipulation/lift/config/ur10_SG/agents/rl_games_ppo_cfg.yaml

@@ -0,0 +1,84 @@
+# Copyright (c) 2022-2025, The Isaac Lab Project Developers (https://github.com/isaac-sim/IsaacLab/blob/main/CONTRIBUTORS.md).
+# All rights reserved.
+#
+# SPDX-License-Identifier: BSD-3-Clause
+
+params:
+  seed: 42
+
+  # environment wrapper clipping
+  env:
+    clip_observations: 100.0
+    clip_actions: 100.0
+
+  algo:
+    name: a2c_continuous
+
+  model:
+    name: continuous_a2c_logstd
+
+  network:
+    name: actor_critic
+    separate: False
+    space:
+      continuous:
+        mu_activation: None
+        sigma_activation: None
+
+        mu_init:
+          name: default
+        sigma_init:
+          name: const_initializer
+          val: 0
+        fixed_sigma: True
+    mlp:
+      units: [256, 128, 64]
+      activation: elu
+      d2rl: False
+
+      initializer:
+        name: default
+      regularizer:
+        name: None
+
+  load_checkpoint: False # flag which sets whether to load the checkpoint
+  load_path: '' # path to the checkpoint to load
+
+  config:
+    name: UR10SG_lift
+    env_name: rlgpu
+    device: 'cuda:0'
+    device_name: 'cuda:0'
+    multi_gpu: False
+    ppo: True
+    mixed_precision: False
+    normalize_input: True
+    normalize_value: True
+    value_bootstrap: False
+    num_actors: -1
+    reward_shaper:
+      scale_value: 0.01
+    normalize_advantage: True
+    gamma: 0.99
+    tau: 0.95
+    learning_rate: 1e-4
+    lr_schedule: adaptive
+    schedule_type: legacy
+    kl_threshold: 0.01
+    score_to_win: 100000000
+    max_epochs: 1500
+    save_best_after: 100
+    save_frequency: 50
+    print_stats: True
+    grad_norm: 1.0
+    entropy_coef: 0.001
+    truncate_grads: True
+    e_clip: 0.2
+    horizon_length: 24
+    minibatch_size: 24576
+    mini_epochs: 8
+    critic_coef: 4
+    clip_value: True
+    clip_actions: False
+    seq_len: 4
+    bounds_loss_coef: 0.0001

source/isaaclab_tasks/isaaclab_tasks/manager_based/manipulation/lift/config/ur10_SG/agents/rsl_rl_ppo_cfg.py

@@ -0,0 +1,38 @@
+# Copyright (c) 2022-2025, The Isaac Lab Project Developers (https://github.com/isaac-sim/IsaacLab/blob/main/CONTRIBUTORS.md).
+# All rights reserved.
+#
+# SPDX-License-Identifier: BSD-3-Clause
+
+from isaaclab.utils import configclass
+
+from isaaclab_rl.rsl_rl import RslRlOnPolicyRunnerCfg, RslRlPpoActorCriticCfg, RslRlPpoAlgorithmCfg
+
+
+@configclass
+class UR10LiftCubePPORunnerCfg(RslRlOnPolicyRunnerCfg):
+    num_steps_per_env = 24
+    max_iterations = 1500
+    save_interval = 50
+    experiment_name = "franka_lift"
+    policy = RslRlPpoActorCriticCfg(
+        init_noise_std=1.0,
+        actor_obs_normalization=False,
+        critic_obs_normalization=False,
+        actor_hidden_dims=[256, 128, 64],
+        critic_hidden_dims=[256, 128, 64],
+        activation="elu",
+    )
+    algorithm = RslRlPpoAlgorithmCfg(
+        value_loss_coef=1.0,
+        use_clipped_value_loss=True,
+        clip_param=0.2,
+        entropy_coef=0.006,
+        num_learning_epochs=5,
+        num_mini_batches=4,
+        learning_rate=1.0e-4,
+        schedule="adaptive",
+        gamma=0.98,
+        lam=0.95,
+        desired_kl=0.01,
+        max_grad_norm=1.0,
+    )

source/isaaclab_tasks/isaaclab_tasks/manager_based/manipulation/lift/config/ur10_SG/agents/sb3_ppo_cfg.yaml

@@ -0,0 +1,34 @@
+# Copyright (c) 2022-2025, The Isaac Lab Project Developers (https://github.com/isaac-sim/IsaacLab/blob/main/CONTRIBUTORS.md).
+# All rights reserved.
+#
+# SPDX-License-Identifier: BSD-3-Clause
+
+# Reference: https://github.com/DLR-RM/rl-baselines3-zoo/blob/master/hyperparams/ppo.yml#L32
+seed: 42
+
+# epoch * n_steps * nenvs: 500×512*8*8
+n_timesteps: 16384000
+policy: 'MlpPolicy'
+n_steps: 64
+# mini batch size: num_envs * nsteps / nminibatches 2048×512÷2048
+batch_size: 192
+gae_lambda: 0.95
+gamma: 0.99
+n_epochs: 8
+ent_coef: 0.00
+vf_coef: 0.0001
+learning_rate: !!float 3e-4
+clip_range: 0.2
+policy_kwargs:
+  activation_fn: 'nn.ELU'
+  net_arch:
+    pi: [256, 128, 64]
+    vf: [256, 128, 64]
+target_kl: 0.01
+max_grad_norm: 1.0
+
+# # Uses VecNormalize class to normalize obs
+# normalize_input: True
+# # Uses VecNormalize class to normalize rew
+# normalize_value: True
+# clip_obs: 5

source/isaaclab_tasks/isaaclab_tasks/manager_based/manipulation/lift/config/ur10_SG/agents/skrl_ppo_cfg.yaml

@@ -0,0 +1,85 @@
+# Copyright (c) 2022-2025, The Isaac Lab Project Developers (https://github.com/isaac-sim/IsaacLab/blob/main/CONTRIBUTORS.md).
+# All rights reserved.
+#
+# SPDX-License-Identifier: BSD-3-Clause
+
+seed: 42
+
+
+# Models are instantiated using skrl's model instantiator utility
+# https://skrl.readthedocs.io/en/latest/api/utils/model_instantiators.html
+models:
+  separate: False
+  policy:  # see gaussian_model parameters
+    class: GaussianMixin
+    clip_actions: False
+    clip_log_std: True
+    min_log_std: -20.0
+    max_log_std: 2.0
+    initial_log_std: 0.0
+    network:
+      - name: net
+        input: OBSERVATIONS
+        layers: [256, 128, 64]
+        activations: elu
+    output: ACTIONS
+  value:  # see deterministic_model parameters
+    class: DeterministicMixin
+    clip_actions: False
+    network:
+      - name: net
+        input: OBSERVATIONS
+        layers: [256, 128, 64]
+        activations: elu
+    output: ONE
+
+
+# Rollout memory
+# https://skrl.readthedocs.io/en/latest/api/memories/random.html
+memory:
+  class: RandomMemory
+  memory_size: -1  # automatically determined (same as agent:rollouts)
+
+
+# PPO agent configuration (field names are from PPO_DEFAULT_CONFIG)
+# https://skrl.readthedocs.io/en/latest/api/agents/ppo.html
+agent:
+  class: PPO
+  rollouts: 24
+  learning_epochs: 8
+  mini_batches: 4
+  discount_factor: 0.99
+  lambda: 0.95
+  learning_rate: 1.0e-04
+  learning_rate_scheduler: KLAdaptiveLR
+  learning_rate_scheduler_kwargs:
+    kl_threshold: 0.01
+  state_preprocessor: RunningStandardScaler
+  state_preprocessor_kwargs: null
+  value_preprocessor: RunningStandardScaler
+  value_preprocessor_kwargs: null
+  random_timesteps: 0
+  learning_starts: 0
+  grad_norm_clip: 1.0
+  ratio_clip: 0.2
+  value_clip: 0.2
+  clip_predicted_values: True
+  entropy_loss_scale: 0.001
+  value_loss_scale: 2.0
+  kl_threshold: 0.0
+  rewards_shaper_scale: 0.01
+  time_limit_bootstrap: False
+  # logging and checkpoint
+  experiment:
+    directory: "ur10sg_lift"
+    experiment_name: ""
+    write_interval: auto
+    checkpoint_interval: auto
+
+
+# Sequential trainer
+# https://skrl.readthedocs.io/en/latest/api/trainers/sequential.html
+trainer:
+  class: SequentialTrainer
+  timesteps: 36000
+  environment_info: log