Add HiwinRA6201621Reacher task

Reference: https://github.com/j3soon/OmniIsaacGymEnvs-DofbotReacher/tree/b29d31bda4843218dd8689f89d15f5b1ab2762e4

.dockerignore

@@ -0,0 +1,2 @@
+# Ignore everything
+*

Dockerfile

@@ -0,0 +1,29 @@
+# Ref: https://catalog.ngc.nvidia.com/orgs/nvidia/containers/isaac-sim
+FROM nvcr.io/nvidia/isaac-sim:2023.1.0-hotfix.1
+
+ENV ISAAC_SIM=/isaac-sim
+WORKDIR /root
+# Install common tools
+RUN apt-get update && apt-get install -y git wget vim \
+    && rm -rf /var/lib/apt/lists/*
+# Download and Install Anaconda
+# Ref: https://www.anaconda.com/products/distribution#Downloads
+RUN wget -q https://repo.anaconda.com/archive/Anaconda3-2022.10-Linux-x86_64.sh \
+    && bash Anaconda3-2022.10-Linux-x86_64.sh -b -p $HOME/anaconda3
+# Patch Isaac Sim 2023.1.0
+# Ref: https://github.com/j3soon/isaac-extended
+RUN git clone https://github.com/j3soon/isaac-extended.git \
+    && cp $ISAAC_SIM/setup_python_env.sh $ISAAC_SIM/setup_python_env.sh.bak \
+    && cp ~/isaac-extended/isaac_sim-2023.1.0-patch/linux/setup_python_env.sh $ISAAC_SIM/setup_python_env.sh
+# Set up conda environment for Isaac Sim
+# Ref: https://docs.omniverse.nvidia.com/isaacsim/latest/installation/install_python.html#advanced-running-with-anaconda
+RUN . ~/anaconda3/etc/profile.d/conda.sh \
+    # conda remove --name isaac-sim --all
+    && cd $ISAAC_SIM \
+    && conda env create -f environment.yml
+# Activation commands for the conda environment
+RUN echo ". ~/anaconda3/etc/profile.d/conda.sh" >> ~/.bashrc \
+    && echo "conda activate isaac-sim" >> ~/.bashrc \
+    && echo ". ${ISAAC_SIM}/setup_conda_env.sh" >> ~/.bashrc
+
+WORKDIR /isaac-sim

LICENSE.txt

@@ -3,6 +3,7 @@ Omniverse Isaac Gym Reinforcement Learning Environments for Isaac Sim
 BSD 3-Clause License
 
 Copyright (c) 2018-2022, NVIDIA Corporation
+Copyright (c) 2022-2023, Johnson Sun
 All rights reserved.
 
 Redistribution and use in source and binary forms, with or without

omniisaacgymenvs/cfg/config.yaml

@@ -72,3 +72,4 @@ hydra:
   run:
     dir: .
 
+use_urdf: False

omniisaacgymenvs/cfg/task/HiwinRA6201621Reacher.yaml

@@ -0,0 +1,126 @@
+# Ref: /omniisaacgymenvs/cfg/task/ShadowHand.yaml
+# used to create the object
+name: HiwinRA6201621Reacher
+
+physics_engine: ${..physics_engine}
+
+# if given, will override the device setting in gym.
+env:
+  numEnvs: ${resolve_default:2048,${...num_envs}}
+  envSpacing: 4
+  episodeLength: 600
+
+  clipObservations: 5.0
+  clipActions: 1.0
+
+  useRelativeControl: False
+  dofSpeedScale: 20.0
+  actionsMovingAverage: 1.0
+  controlFrequencyInv: 2 # 60 Hz
+
+  startPositionNoise: 0.01
+  startRotationNoise: 0.0
+
+  resetPositionNoise: 0.01
+  resetRotationNoise: 0.0
+  resetDofPosRandomInterval: 0.2
+  resetDofVelRandomInterval: 0.0
+
+ # Random forces applied to the object
+  forceScale: 0.0
+  forceProbRange: [0.001, 0.1]
+  forceDecay: 0.99
+  forceDecayInterval: 0.08
+
+  # reward -> dictionary
+  distRewardScale: -2.0
+  rotRewardScale: 1.0
+  rotEps: 0.1
+  actionPenaltyScale: -0.0002
+  reachGoalBonus: 250
+  velObsScale: 0.2
+
+  observationType: "full" # can only be "full"
+  successTolerance: 0.1
+  printNumSuccesses: False
+  maxConsecutiveSuccesses: 0
+
+  useURDF: ${resolve_default:True,${...use_urdf}}
+
+sim:
+  dt: 0.0083 # 1/120 s
+  add_ground_plane: True
+  add_distant_light: False
+  use_gpu_pipeline: ${eq:${...pipeline},"gpu"}
+  use_fabric: True
+  enable_scene_query_support: False
+  disable_contact_processing: False
+
+  # set to True if you use camera sensors in the environment
+  enable_cameras: False
+
+  default_material:
+    static_friction: 1.0
+    dynamic_friction: 1.0
+    restitution: 0.0
+
+  physx:
+    # per-scene
+    use_gpu: ${eq:${....sim_device},"gpu"} # set to False to run on CPU
+    worker_thread_count: ${....num_threads}
+    solver_type: ${....solver_type} # 0: PGS, 1: TGS
+    bounce_threshold_velocity: 0.2
+    friction_offset_threshold: 0.04
+    friction_correlation_distance: 0.025
+    enable_sleeping: True
+    enable_stabilization: True
+    # GPU buffers
+    gpu_max_rigid_contact_count: 1048576
+    gpu_max_rigid_patch_count: 33554432
+    gpu_found_lost_pairs_capacity: 20971520
+    gpu_found_lost_aggregate_pairs_capacity: 20971520
+    gpu_total_aggregate_pairs_capacity: 20971520
+    gpu_max_soft_body_contacts: 1048576
+    gpu_max_particle_contacts: 1048576
+    gpu_heap_capacity: 33554432
+    gpu_temp_buffer_capacity: 16777216
+    gpu_max_num_partitions: 8
+
+  hiwin:
+    # -1 to use default values
+    override_usd_defaults: False
+    enable_self_collisions: False
+  object:
+    # -1 to use default values
+    override_usd_defaults: False
+    make_kinematic: True
+    enable_self_collisions: False
+    enable_gyroscopic_forces: True
+    # also in stage params
+    # per-actor
+    solver_position_iteration_count: 8
+    solver_velocity_iteration_count: 0
+    sleep_threshold: 0.000
+    stabilization_threshold: 0.0025
+    # per-body
+    density: -1
+    max_depenetration_velocity: 1000.0
+  goal_object:
+    # -1 to use default values
+    override_usd_defaults: False
+    make_kinematic: True
+    enable_self_collisions: False
+    enable_gyroscopic_forces: True
+    # also in stage params
+    # per-actor
+    solver_position_iteration_count: 8
+    solver_velocity_iteration_count: 0
+    sleep_threshold: 0.000
+    stabilization_threshold: 0.0025
+    # per-body
+    density: -1
+    max_depenetration_velocity: 1000.0
+sim2real:
+  enabled: False
+  fail_quietely: False
+  verbose: False

omniisaacgymenvs/cfg/train/HiwinRA6201621ReacherPPO.yaml

@@ -0,0 +1,79 @@
+# Ref: /omniisaacgymenvs/cfg/train/ShadowHandPPO.yaml
+params:
+  seed: ${...seed}
+  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: ${if:${...checkpoint},True,False}
+  load_path: ${...checkpoint}
+
+  config:
+    name: ${resolve_default:HiwinRA6201621Reacher,${....experiment}}
+    full_experiment_name: ${.name}
+    device: ${....rl_device}
+    device_name: ${....rl_device}
+    env_name: rlgpu
+    multi_gpu: ${....multi_gpu}
+    ppo: True
+    mixed_precision: False
+    normalize_input: True
+    normalize_value: True
+    value_bootstrap: True
+    num_actors: ${....task.env.numEnvs}
+    reward_shaper:
+      scale_value: 0.01
+    normalize_advantage: True
+    gamma: 0.99
+    tau: 0.95
+    learning_rate: 5e-3
+    lr_schedule: adaptive
+    schedule_type: standard
+    kl_threshold: 0.02
+    score_to_win: 100000
+    max_epochs: ${resolve_default:5000,${....max_iterations}}
+    save_best_after: 100
+    save_frequency: 200
+    print_stats: True
+    grad_norm: 1.0
+    entropy_coef: 0.0
+    truncate_grads: True
+    e_clip: 0.2
+    horizon_length: 64
+    minibatch_size: 32768
+    mini_epochs: 5
+    critic_coef: 4
+    clip_value: True
+    seq_length: 4
+    bounds_loss_coef: 0.0001
+
+    player:
+      deterministic: True
+      games_num: 100000
+      print_stats: True