生成多个资产

生成多个资产#

典型的生成配置（在 生成基本物体到场景中 教程中介绍）复制相同的资产（或 USD 基元）到不同的已解析的基元路径表达式中。例如，如果用户指定要在 “/World/Table_.*/Object” 生成资产，则相同的资产将在路径 “/World/Table_0/Object” 、 “/World/Table_1/Object” 等处创建。

但有时，可能希望在基元路径下生成不同的资产，以确保模拟的多样性。本指南介绍了如何使用生成功能在每个基元路径下创建不同的资产。

示例脚本 multi_asset.py 用作参考，位于 IsaacLab/source/standalone/demos 目录中。

multi_asset.py 的代码
  1# Copyright (c) 2022-2024, The Isaac Lab Project Developers.
  2# All rights reserved.
  3#
  4# SPDX-License-Identifier: BSD-3-Clause
  5
  6"""This script demonstrates how to spawn multiple objects in multiple environments.
  7
  8.. code-block:: bash
  9
 10    # Usage
 11    ./isaaclab.sh -p source/standalone/demos/multi_asset.py --num_envs 2048
 12
 13"""
 14
 15from __future__ import annotations
 16
 17"""Launch Isaac Sim Simulator first."""
 18
 19
 20import argparse
 21
 22from omni.isaac.lab.app import AppLauncher
 23
 24# add argparse arguments
 25parser = argparse.ArgumentParser(description="Demo on spawning different objects in multiple environments.")
 26parser.add_argument("--num_envs", type=int, default=1024, help="Number of environments to spawn.")
 27# append AppLauncher cli args
 28AppLauncher.add_app_launcher_args(parser)
 29# parse the arguments
 30args_cli = parser.parse_args()
 31
 32# launch omniverse app
 33app_launcher = AppLauncher(args_cli)
 34simulation_app = app_launcher.app
 35
 36"""Rest everything follows."""
 37
 38import random
 39
 40import omni.usd
 41from pxr import Gf, Sdf
 42
 43import omni.isaac.lab.sim as sim_utils
 44from omni.isaac.lab.assets import ArticulationCfg, AssetBaseCfg, RigidObjectCfg
 45from omni.isaac.lab.scene import InteractiveScene, InteractiveSceneCfg
 46from omni.isaac.lab.sim import SimulationContext
 47from omni.isaac.lab.utils import Timer, configclass
 48from omni.isaac.lab.utils.assets import ISAACLAB_NUCLEUS_DIR
 49
 50##
 51# Pre-defined Configuration
 52##
 53
 54from omni.isaac.lab_assets.anymal import ANYDRIVE_3_LSTM_ACTUATOR_CFG  # isort: skip
 55
 56
 57##
 58# Randomization events.
 59##
 60
 61
 62def randomize_shape_color(prim_path_expr: str):
 63    """Randomize the color of the geometry."""
 64    # acquire stage
 65    stage = omni.usd.get_context().get_stage()
 66    # resolve prim paths for spawning and cloning
 67    prim_paths = sim_utils.find_matching_prim_paths(prim_path_expr)
 68    # manually clone prims if the source prim path is a regex expression
 69    with Sdf.ChangeBlock():
 70        for prim_path in prim_paths:
 71            # spawn single instance
 72            prim_spec = Sdf.CreatePrimInLayer(stage.GetRootLayer(), prim_path)
 73
 74            # DO YOUR OWN OTHER KIND OF RANDOMIZATION HERE!
 75            # Note: Just need to acquire the right attribute about the property you want to set
 76            # Here is an example on setting color randomly
 77            color_spec = prim_spec.GetAttributeAtPath(prim_path + "/geometry/material/Shader.inputs:diffuseColor")
 78            color_spec.default = Gf.Vec3f(random.random(), random.random(), random.random())
 79
 80
 81##
 82# Scene Configuration
 83##
 84
 85
 86@configclass
 87class MultiObjectSceneCfg(InteractiveSceneCfg):
 88    """Configuration for a multi-object scene."""
 89
 90    # ground plane
 91    ground = AssetBaseCfg(prim_path="/World/defaultGroundPlane", spawn=sim_utils.GroundPlaneCfg())
 92
 93    # lights
 94    dome_light = AssetBaseCfg(
 95        prim_path="/World/Light", spawn=sim_utils.DomeLightCfg(intensity=3000.0, color=(0.75, 0.75, 0.75))
 96    )
 97
 98    # rigid object
 99    object: RigidObjectCfg = RigidObjectCfg(
100        prim_path="/World/envs/env_.*/Object",
101        spawn=sim_utils.MultiAssetSpawnerCfg(
102            assets_cfg=[
103                sim_utils.ConeCfg(
104                    radius=0.3,
105                    height=0.6,
106                    visual_material=sim_utils.PreviewSurfaceCfg(diffuse_color=(0.0, 1.0, 0.0), metallic=0.2),
107                ),
108                sim_utils.CuboidCfg(
109                    size=(0.3, 0.3, 0.3),
110                    visual_material=sim_utils.PreviewSurfaceCfg(diffuse_color=(1.0, 0.0, 0.0), metallic=0.2),
111                ),
112                sim_utils.SphereCfg(
113                    radius=0.3,
114                    visual_material=sim_utils.PreviewSurfaceCfg(diffuse_color=(0.0, 0.0, 1.0), metallic=0.2),
115                ),
116            ],
117            random_choice=True,
118            rigid_props=sim_utils.RigidBodyPropertiesCfg(
119                solver_position_iteration_count=4, solver_velocity_iteration_count=0
120            ),
121            mass_props=sim_utils.MassPropertiesCfg(mass=1.0),
122            collision_props=sim_utils.CollisionPropertiesCfg(),
123        ),
124        init_state=RigidObjectCfg.InitialStateCfg(pos=(0.0, 0.0, 2.0)),
125    )
126
127    # articulation
128    robot: ArticulationCfg = ArticulationCfg(
129        prim_path="/World/envs/env_.*/Robot",
130        spawn=sim_utils.MultiUsdFileCfg(
131            usd_path=[
132                f"{ISAACLAB_NUCLEUS_DIR}/Robots/ANYbotics/ANYmal-C/anymal_c.usd",
133                f"{ISAACLAB_NUCLEUS_DIR}/Robots/ANYbotics/ANYmal-D/anymal_d.usd",
134            ],
135            random_choice=True,
136            rigid_props=sim_utils.RigidBodyPropertiesCfg(
137                disable_gravity=False,
138                retain_accelerations=False,
139                linear_damping=0.0,
140                angular_damping=0.0,
141                max_linear_velocity=1000.0,
142                max_angular_velocity=1000.0,
143                max_depenetration_velocity=1.0,
144            ),
145            articulation_props=sim_utils.ArticulationRootPropertiesCfg(
146                enabled_self_collisions=True, solver_position_iteration_count=4, solver_velocity_iteration_count=0
147            ),
148            activate_contact_sensors=True,
149        ),
150        init_state=ArticulationCfg.InitialStateCfg(
151            pos=(0.0, 0.0, 0.6),
152            joint_pos={
153                ".*HAA": 0.0,  # all HAA
154                ".*F_HFE": 0.4,  # both front HFE
155                ".*H_HFE": -0.4,  # both hind HFE
156                ".*F_KFE": -0.8,  # both front KFE
157                ".*H_KFE": 0.8,  # both hind KFE
158            },
159        ),
160        actuators={"legs": ANYDRIVE_3_LSTM_ACTUATOR_CFG},
161    )
162
163
164##
165# Simulation Loop
166##
167
168
169def run_simulator(sim: SimulationContext, scene: InteractiveScene):
170    """Runs the simulation loop."""
171    # Extract scene entities
172    # note: we only do this here for readability.
173    rigid_object = scene["object"]
174    robot = scene["robot"]
175    # Define simulation stepping
176    sim_dt = sim.get_physics_dt()
177    count = 0
178    # Simulation loop
179    while simulation_app.is_running():
180        # Reset
181        if count % 500 == 0:
182            # reset counter
183            count = 0
184            # reset the scene entities
185            # object
186            root_state = rigid_object.data.default_root_state.clone()
187            root_state[:, :3] += scene.env_origins
188            rigid_object.write_root_state_to_sim(root_state)
189            # robot
190            # -- root state
191            root_state = robot.data.default_root_state.clone()
192            root_state[:, :3] += scene.env_origins
193            robot.write_root_state_to_sim(root_state)
194            # -- joint state
195            joint_pos, joint_vel = robot.data.default_joint_pos.clone(), robot.data.default_joint_vel.clone()
196            robot.write_joint_state_to_sim(joint_pos, joint_vel)
197            # clear internal buffers
198            scene.reset()
199            print("[INFO]: Resetting scene state...")
200
201        # Apply action to robot
202        robot.set_joint_position_target(robot.data.default_joint_pos)
203        # Write data to sim
204        scene.write_data_to_sim()
205        # Perform step
206        sim.step()
207        # Increment counter
208        count += 1
209        # Update buffers
210        scene.update(sim_dt)
211
212
213def main():
214    """Main function."""
215    # Load kit helper
216    sim_cfg = sim_utils.SimulationCfg(dt=0.005, device=args_cli.device)
217    sim = SimulationContext(sim_cfg)
218    # Set main camera
219    sim.set_camera_view([2.5, 0.0, 4.0], [0.0, 0.0, 2.0])
220
221    # Design scene
222    scene_cfg = MultiObjectSceneCfg(num_envs=args_cli.num_envs, env_spacing=2.0, replicate_physics=False)
223    with Timer("[INFO] Time to create scene: "):
224        scene = InteractiveScene(scene_cfg)
225
226    with Timer("[INFO] Time to randomize scene: "):
227        # DO YOUR OWN OTHER KIND OF RANDOMIZATION HERE!
228        # Note: Just need to acquire the right attribute about the property you want to set
229        # Here is an example on setting color randomly
230        randomize_shape_color(scene_cfg.object.prim_path)
231
232    # Play the simulator
233    sim.reset()
234    # Now we are ready!
235    print("[INFO]: Setup complete...")
236    # Run the simulator
237    run_simulator(sim, scene)
238
239
240if __name__ == "__main__":
241    # run the main execution
242    main()
243    # close sim app
244    simulation_app.close()

此脚本创建多个环境，每个环境都有一个刚性对象，可以是圆锥、立方体或球体，以及一个分别是 ANYmal-C 或 ANYmal-D 机器人的关节。

multi_asset.py 的结果

使用多资产生成功能#

可以使用生成器 MultiAssetSpawnerCfg 和 MultiUsdFileCfg 在每个环境中生成不同的资产和 USDs：

  • 我们将 RigidObjectCfg 中的生成配置设置为 MultiAssetSpawnerCfg ：

    object: RigidObjectCfg = RigidObjectCfg(
    prim_path="/World/envs/env_.*/Object",
    spawn=sim_utils.MultiAssetSpawnerCfg(
        assets_cfg=[
            sim_utils.ConeCfg(
                radius=0.3,
                height=0.6,
                visual_material=sim_utils.PreviewSurfaceCfg(diffuse_color=(0.0, 1.0, 0.0), metallic=0.2),
            ),
            sim_utils.CuboidCfg(
                size=(0.3, 0.3, 0.3),
                visual_material=sim_utils.PreviewSurfaceCfg(diffuse_color=(1.0, 0.0, 0.0), metallic=0.2),
            ),
            sim_utils.SphereCfg(
                radius=0.3,
                visual_material=sim_utils.PreviewSurfaceCfg(diffuse_color=(0.0, 0.0, 1.0), metallic=0.2),
            ),
        ],
        random_choice=True,
        rigid_props=sim_utils.RigidBodyPropertiesCfg(
            solver_position_iteration_count=4, solver_velocity_iteration_count=0
        ),
        mass_props=sim_utils.MassPropertiesCfg(mass=1.0),
        collision_props=sim_utils.CollisionPropertiesCfg(),
    ),
    init_state=RigidObjectCfg.InitialStateCfg(pos=(0.0, 0.0, 2.0)),
)

    此函数允许您定义可以作为刚性对象生成的不同资产的列表。当 random_choice 设置为 True 时，将从列表中随机选择一个资产，并在指定的基元路径上生成它。

  • 类似地，我们将 ArticulationCfg 中的生成配置设置为 MultiUsdFileCfg ：

    robot: ArticulationCfg = ArticulationCfg(
    prim_path="/World/envs/env_.*/Robot",
    spawn=sim_utils.MultiUsdFileCfg(
        usd_path=[
            f"{ISAACLAB_NUCLEUS_DIR}/Robots/ANYbotics/ANYmal-C/anymal_c.usd",
            f"{ISAACLAB_NUCLEUS_DIR}/Robots/ANYbotics/ANYmal-D/anymal_d.usd",
        ],
        random_choice=True,
        rigid_props=sim_utils.RigidBodyPropertiesCfg(
            disable_gravity=False,
            retain_accelerations=False,
            linear_damping=0.0,
            angular_damping=0.0,
            max_linear_velocity=1000.0,
            max_angular_velocity=1000.0,
            max_depenetration_velocity=1.0,
        ),
        articulation_props=sim_utils.ArticulationRootPropertiesCfg(
            enabled_self_collisions=True, solver_position_iteration_count=4, solver_velocity_iteration_count=0
        ),
        activate_contact_sensors=True,
    ),
    init_state=ArticulationCfg.InitialStateCfg(
        pos=(0.0, 0.0, 0.6),
        joint_pos={
            ".*HAA": 0.0,  # all HAA
            ".*F_HFE": 0.4,  # both front HFE
            ".*H_HFE": -0.4,  # both hind HFE
            ".*F_KFE": -0.8,  # both front KFE
            ".*H_KFE": 0.8,  # both hind KFE
        },
    ),
    actuators={"legs": ANYDRIVE_3_LSTM_ACTUATOR_CFG},
)

    与之前类似，此配置允许选择代表关节资产的不同 USD 文件。

需要注意的事项#

相似的资产结构#

在使用相同的物理接口（刚性对象或关节类）生成和处理多个资产时，必须确保所有基元路径上的资产遵循相似的结构。对于关节来说，这意味着它们都必须具有相同数量的链接和关节、相同数量的碰撞体，以及相同的名称。如果不是这种情况，基元的物理解析可能会受到影响并且失败。

此功能的主要目的是使用户能够创建相同资产的随机版本，例如具有不同链接长度的机器人，或具有不同碰撞器形状的刚性对象。

在交互式场景中禁用物理复制#

默认情况下，标志 scene.InteractiveScene.replicate_physics 被设置为 True。此标志告知物理引擎模拟环境是彼此的副本，因此只需解析第一个环境即可了解整个模拟场景。这有助于加快模拟场景的解析速度。

但是，在不同环境中生成不同的资产的情况下，这一假设不再成立。因此，必须禁用标志 scene.InteractiveScene.replicate_physics 。

# Design scene
scene_cfg = MultiObjectSceneCfg(num_envs=args_cli.num_envs, env_spacing=2.0, replicate_physics=False)
with Timer("[INFO] Time to create scene: "):
    scene = InteractiveScene(scene_cfg)

代码执行#

要执行带有多个环境和随机资产的脚本，请使用以下命令：

./isaaclab.sh -p source/standalone/demos/multi_asset.py --num_envs 2048

此命令运行具有 2048 个环境的模拟场景，每个环境都具有随机选择的资产。要停止模拟，可以关闭窗口或在终端中按下 Ctrl+C 。