mobility-gen

name

mobility-gen

description

Generate synthetic sensor datasets for mobile robots using Isaac Sim's MobilityGen extension. Two-phase pipeline: headless trajectory recording → replay-and-render for RGB/depth/seg/normals/pose data. Covers `RandomPathFollowingScenario`, `RandomAccelerationScenario`, and custom robot subclassing (`WheeledMobilityGenRobot`, `PolicyMobilityGenRobot`, holonomic). Use when generating mobile-robot training data, recording sim trajectories, replaying with sensors, or implementing a custom MobilityGenRobot. Distinct from `data-collection-sim` (static-scene SDG) — this skill is robot-trajectory-driven. See `navigation-primitives` for shared occupancy-map / A* / kinematics substrate; `isaac-sim-robot-navigation` for non-SDG runtime nav.

MobilityGen Synthetic Data Generation

Two-phase pipeline: record trajectories (physics, no rendering) → replay & render (sensors added).

Read These Skills First

• navigation-primitives — OccupancyMap, A* planner, robot footprints (Spot Z=0.69), differential/holonomic kinematics, look-at chase cameras, shared gotchas. MobilityGen consumes this substrate; this skill assumes you know it.
• occupancy-map — produces the map.yaml consumed by OccupancyMap.from_ros_yaml
• data-collection-sim — sibling SDG path for static scenes with randomized object/camera poses (no robot trajectory)

When To Use This Skill (vs siblings)

Goal	Use
Record trajectories then re-render with sensors for SDG (training data)	this skill
Drive a robot through a scene in real time, see it move	isaac-sim-robot-navigation
Annotated frames with no robot motion (object pose randomization)	data-collection-sim

Related Skills

• navigation-primitives — shared navigation substrate (read first)
• data-collection-sim — static-scene SDG sibling
• isaac-sim-sensor — sensor primitives (camera, LiDAR, IMU, contact)
• isaac-sim-robot-navigation — runtime navigation sibling
• isaac-sim-headless-deployment — --no-window headless launch and SimulationApp batch pattern

Environment

• Isaac Sim source tree: $ISAAC_SIM_DIR/source/ for source builds. The variable $ISAAC_SIM_SRC is a convenience alias for that path; declare it once at the top of your launcher (e.g. ISAAC_SIM_SRC="$ISAAC_SIM_DIR/source").
• Python launcher: $ISAAC_SIM_DIR/python.sh
• Replay script: $ISAAC_SIM_SRC/standalone_examples/replicator/mobility_gen/replay_directory.py
• Extension examples: $ISAAC_SIM_SRC/extensions/isaacsim.replicator.mobility_gen.examples/
• Data dir: $MOBILITY_GEN_DATA (env var). Default to a workspace-local path such as $WORKSPACE_DIR/MobilityGenData or $HOME/MobilityGenData.
  • recordings/ — timestamped trajectory dirs
  • replays/ — rendered output
  • maps/ — occupancy map YAML + PNG files

Extension Loading (Critical)

Extensions are not auto-loaded. Always pass --enable flags when running python.sh:

"$ISAAC/python.sh" my_script.py --enable isaacsim.replicator.mobility_gen.examples

"$ISAAC/python.sh" my_script.py \
  --enable isaacsim.asset.gen.omap \
  --enable isaacsim.replicator.mobility_gen.examples

Enabling isaacsim.replicator.mobility_gen.examples auto-loads isaacsim.replicator.mobility_gen as a dependency. All extension-dependent imports must come AFTER SimulationApp(...) is initialized.

Phase 1: Automated Trajectory Recording (Headless)

KeyboardTeleoperationScenario and GamepadTeleoperationScenario require an interactive UI. For headless batch recording use RandomPathFollowingScenario or RandomAccelerationScenario.

API note (Kit 110): the snippet below uses isaacsim.core.api.objects, isaacsim.core.utils.stage, and the MobilityGen get_world / new_world helpers. These wrap the legacy isaacsim.core.api.World flow that MobilityGen's scenario/writer plumbing expects. For new code outside MobilityGen, prefer the experimental APIs (isaacsim.core.experimental.utils.stage, the isaacsim.core.simulation_manager.SimulationManager lifecycle, and isaacsim.core.experimental.objects for primitives) — see the cheatsheet. Do not mix the two within a single MobilityGen session.

Migration: for the full omni.isaac.* → isaacsim.* mapping when porting scripts off the legacy World flow, see Renaming Extensions.

from isaacsim import SimulationApp
simulation_app = SimulationApp(launch_config={"headless": True})

import datetime, os, tempfile
import isaacsim.core.api.objects as objects
import isaacsim.replicator.mobility_gen.examples
from isaacsim.core.utils.stage import open_stage, save_stage
from isaacsim.replicator.mobility_gen.impl.config import Config
from isaacsim.replicator.mobility_gen.impl.occupancy_map import OccupancyMap
from isaacsim.replicator.mobility_gen.impl.robot import ROBOTS
from isaacsim.replicator.mobility_gen.impl.scenario import SCENARIOS
from isaacsim.replicator.mobility_gen.impl.utils.global_utils import get_world, new_world
from isaacsim.replicator.mobility_gen.impl.writer import MobilityGenWriter

DATA_DIR = os.environ.get(
    "MOBILITY_GEN_DATA",
    os.path.join(os.environ.get("WORKSPACE_DIR", os.path.expanduser("~")), "MobilityGenData"),
)

SCENE_USD    = "/path/to/warehouse.usd"
OMAP_YAML    = f"{DATA_DIR}/maps/warehouse/map.yaml"
ROBOT_TYPE   = "CarterRobot"               # JetbotRobot | CarterRobot | H1Robot | SpotRobot
SCENARIO     = "RandomPathFollowingScenario"  # or RandomAccelerationScenario
NUM_EPISODES = 5
MAX_STEPS    = 2000

robot_cls    = ROBOTS.get(ROBOT_TYPE)
scenario_cls = SCENARIOS.get(SCENARIO)

config = Config(scenario_type=SCENARIO, robot_type=ROBOT_TYPE, scene_usd=SCENE_USD)
occupancy_map = OccupancyMap.from_ros_yaml(OMAP_YAML)

open_stage(SCENE_USD)
cached_stage = os.path.join(tempfile.mkdtemp(), "stage.usd")
save_stage(cached_stage, save_and_reload_in_place=)

world = new_world(physics_dt=robot_cls.physics_dt)
world.initialize_simulation_context()
objects.GroundPlane(, visible=)

robot    = robot_cls.build()
scenario = scenario_cls.from_robot_occupancy_map(robot, occupancy_map)

os.makedirs(os.path.join(DATA_DIR, ), exist_ok=)

 episode  (NUM_EPISODES):
    world.reset()
    scenario.reset()

    name   = datetime.datetime.now().strftime()
    path   = os.path.join(DATA_DIR, , name)
    writer = MobilityGenWriter(path)
    writer.write_config(config)
    writer.write_occupancy_map(occupancy_map)
    writer.copy_stage(cached_stage)

    step = 
     :
        world.step(render=)
        is_alive = scenario.step(step_size=robot_cls.physics_dt)
        writer.write_state_dict_common(scenario.state_dict_common(), step)
        step += 
          is_alive  step >= MAX_STEPS:
            

    ()

simulation_app.close()

Phase 2: Replay & Render

Replay all recordings in $MOBILITY_GEN_DATA/recordings/ and write sensor data to replays/.

: "${MOBILITY_GEN_DATA:=${WORKSPACE_DIR:-$HOME}/MobilityGenData}"
ISAAC="$ISAAC_SIM_DIR"
SRC="$ISAAC_SIM_DIR/source"

CUDA_VISIBLE_DEVICES=0 DISPLAY=:99 nohup \
  "$ISAAC/python.sh" \
  "$SRC/standalone_examples/replicator/mobility_gen/replay_directory.py" \
  --input  "$MOBILITY_GEN_DATA/recordings" \
  --output "$MOBILITY_GEN_DATA/replays" \
  --render_interval 40 \
  --rgb_enabled True \
  --depth_enabled True \
  --segmentation_enabled True \
  --normals_enabled False \
  --render_rt_subframes 1 \
  --enable isaacsim.replicator.mobility_gen.examples \
  > /tmp/mobility_gen_replay.log 2>&1 &

--render_interval 40 = 1 frame per 40 physics steps (~5 Hz at 200 Hz physics). Increase --render_rt_subframes for better quality at the cost of speed.

Replay Output Structure

replays/<recording_name>/
  config.json
  stage.usd
  occupancy_map/map.yaml, map.png
  state/
    common/<step>.npy          # robot pose, joint positions, velocities
    rgb/<camera_name>/<step>.jpg
    segmentation/<camera_name>/<step>.png
    depth/<camera_name>/<step>.png   # 16-bit inverse depth
    normals/<camera_name>/<step>.npy

Available Robots

Name	Type	Notes
JetbotRobot	Wheeled (differential)	Small, physics_dt=0.005, Jetbot USD
CarterRobot	Wheeled (differential)	Nova Carter, physics_dt=0.005
H1Robot	Humanoid (policy)	Unitree H1, flat-terrain RL policy
SpotRobot	Quadruped (policy)	Boston Dynamics Spot, flat-terrain RL policy

Available Scenarios

Name	Mode	Headless?
KeyboardTeleoperationScenario	Manual (WASD)	No — needs UI
GamepadTeleoperationScenario	Manual (gamepad)	No — needs UI
RandomAccelerationScenario	Automated (brownian)	Yes
RandomPathFollowingScenario	Automated (A* path following)	Yes

RandomPathFollowingScenario plans an A* path from the robot's current position to a random free-space goal and follows it with proportional steering. Episode ends when goal is reached or robot collides.

Add a Custom Robot

Two base classes exist depending on robot type. Both handle build() and write_action() — set class-level attributes only.

Wheeled (differential drive)

Subclass WheeledMobilityGenRobot. No need to override build() or write_action():

from isaacsim.replicator.mobility_gen.examples.robots import WheeledMobilityGenRobot
from isaacsim.replicator.mobility_gen.examples.misc import HawkCamera
from isaacsim.replicator.mobility_gen.impl.robot import ROBOTS

@ROBOTS.register()
class MyRobot(WheeledMobilityGenRobot):
    physics_dt: float = 0.005
    z_offset: float = 0.25

    chase_camera_base_path = "chassis"
    chase_camera_x_offset: float = -1.5
    chase_camera_z_offset: float = 0.8
    chase_camera_tilt_angle: float = 60.0

    front_camera_base_path = "chassis/front_hawk"
    front_camera_rotation = (0.0, 0.0, 0.0)
    front_camera_translation = (0.2, 0.0, 0.1)
    front_camera_type = HawkCamera

    occupancy_map_radius: float = 0.5
    occupancy_map_z_min: float = 0.1
    occupancy_map_z_max: float = 0.5
    occupancy_map_cell_size: float = 0.05
    occupancy_map_collision_radius: float = 0.5

    keyboard_linear_velocity_gain: float = 1.0
    keyboard_angular_velocity_gain: float = 1.0
    gamepad_linear_velocity_gain: float = 1.0
    gamepad_angular_velocity_gain:  = 

    random_action_linear_velocity_range = (-, )
    random_action_angular_velocity_range = (-, )
    random_action_linear_acceleration_std:  = 
    random_action_angular_acceleration_std:  = 
    random_action_grid_pose_sampler_grid_size:  = 
    path_following_speed:  = 
    path_following_angular_gain:  = 
    path_following_stop_distance_threshold:  = 
    path_following_forward_angle_threshold = 
    path_following_target_point_offset_meters:  = 

    wheel_dof_names = [, ]
    usd_url:  = 
    chassis_subpath:  = 
    wheel_base:  = 
    wheel_radius:  = 

Reference implementations in isaacsim.replicator.mobility_gen.examples.robots:

• JetbotRobot: NVIDIA Jetbot, wheel_base=0.1125, wheel_radius=0.03, chassis_subpath="chassis"
• CarterRobot: Nova Carter, wheel_base=0.413, wheel_radius=0.14, chassis_subpath="chassis_link"

Holonomic (e.g. Kaya 3-wheel)

Override build() to use a different controller and write_action() to remap the 2D action:

from isaacsim.replicator.mobility_gen.examples.robots import WheeledMobilityGenRobot
from isaacsim.replicator.mobility_gen.examples.misc import HawkCamera
from isaacsim.replicator.mobility_gen.impl.robot import ROBOTS
from isaacsim.replicator.mobility_gen.impl.utils.global_utils import get_world, join_sdf_paths
from isaacsim.core.prims import Articulation as _ArticulationView
from isaacsim.robot.wheeled_robots.robots import WheeledRobot as _WheeledRobot
from isaacsim.robot.wheeled_robots.controllers.holonomic_controller import HolonomicController
from isaacsim.robot.wheeled_robots.robots.holonomic_robot_usd_setup import HolonomicRobotUsdSetup
from isaacsim.storage.native import get_assets_root_path

@ROBOTS.register()
class KayaRobot(WheeledMobilityGenRobot):
    physics_dt: float = 0.005
    z_offset: float = 0.02
    chase_camera_base_path = "base_link"
    chase_camera_x_offset: float = -0.5
    chase_camera_z_offset: float = 0.3
    chase_camera_tilt_angle: float = 60.0
    front_camera_base_path = "base_link/front_hawk"
    front_camera_rotation = (0.0, 0.0, 0.0)
    front_camera_translation = (0.1, 0.0, 0.05)
    front_camera_type = HawkCamera

    occupancy_map_radius:  = 
    occupancy_map_z_min:  = 
    occupancy_map_z_max:  = 
    occupancy_map_cell_size:  = 
    occupancy_map_collision_radius:  = 
    random_action_linear_velocity_range = (-, )
    random_action_angular_velocity_range = (-, )
    random_action_linear_acceleration_std:  = 
    random_action_angular_acceleration_std:  = 
    random_action_grid_pose_sampler_grid_size:  = 
    path_following_speed:  = 
    path_following_angular_gain:  = 
    path_following_stop_distance_threshold:  = 
    path_following_forward_angle_threshold = 
    path_following_target_point_offset_meters:  = 
    keyboard_linear_velocity_gain:  = 
    keyboard_angular_velocity_gain:  = 
    gamepad_linear_velocity_gain:  = 
    gamepad_angular_velocity_gain:  = 

    wheel_dof_names = [, , ]
    usd_url:  = get_assets_root_path() + 
    chassis_subpath:  = 
    wheel_radius:  = 
    wheel_base:  = 
    com_prim_subpath:  = 


     ():
        world = get_world()
        robot = world.scene.add(
            _WheeledRobot(prim_path, wheel_dof_names=cls.wheel_dof_names, create_robot=, usd_path=cls.usd_url)
        )
        view = _ArticulationView(join_sdf_paths(prim_path, cls.chassis_subpath))
        world.scene.add(view)
        kaya_setup = HolonomicRobotUsdSetup(
            robot_prim_path=prim_path,
            com_prim_path=join_sdf_paths(prim_path, cls.com_prim_subpath),
        )
        (wheel_radius, wheel_positions, wheel_orientations, mecanum_angles, wheel_axis, up_axis) = \
            kaya_setup.get_holonomic_controller_params()
        controller = HolonomicController(
            name=, wheel_radius=wheel_radius, wheel_positions=wheel_positions,
            wheel_orientations=wheel_orientations, mecanum_angles=mecanum_angles,
            wheel_axis=wheel_axis, up_axis=up_axis,
        )
        camera = cls.build_front_camera(prim_path)
         cls(prim_path=prim_path, robot=robot, articulation_view=view, controller=controller, front_camera=camera)

     ():
        action = .action.get_value()
        
        .robot.apply_wheel_actions(.controller.forward(command=[action[], , action[]]))

Policy-based (legged robots)

Subclass PolicyMobilityGenRobot and implement build_policy(). write_action() converts the 2D action [lin_vel, ang_vel] into the 3D command [x, 0, yaw] automatically.

from isaacsim.replicator.mobility_gen.examples.robots import PolicyMobilityGenRobot
from isaacsim.replicator.mobility_gen.impl.robot import ROBOTS
from isaacsim.robot.policy.examples.robots import H1FlatTerrainPolicy
import numpy as np

@ROBOTS.register()
class MyLegedRobot(PolicyMobilityGenRobot):
    physics_dt: float = 0.005
    z_offset: float = 1.05
    usd_url = "/path/to/robot.usd"
    articulation_path = "pelvis"
    controller_z_offset: float = 1.05
    # ... same occupancy_map_*, random_action_*, path_following_* attrs as Wheeled ...

    @classmethod
    def build_policy(cls, prim_path: str):
        return H1FlatTerrainPolicy(prim_path=prim_path, position=np.array([0.0, 0.0, cls.controller_z_offset]))

Reference implementations: H1Robot (articulation_path="pelvis") and SpotRobot (articulation_path="/") in the same module.

Replay with a custom robot

replay_directory.py calls load_scenario() which does ROBOTS.get(config.robot_type). If the robot isn't in the built-in extension, this raises KeyError. You cannot pass --enable to load an ad-hoc Python file — either create a proper Isaac extension, or copy the replay loop into your own script and register the robot class before calling load_scenario():

from isaacsim import SimulationApp
simulation_app = SimulationApp(launch_config={"headless": True})

import glob, os
import isaacsim.replicator.mobility_gen.examples  # built-in robots
from isaacsim.replicator.mobility_gen.impl.build import load_scenario
from isaacsim.replicator.mobility_gen.impl.robot import ROBOTS
from isaacsim.replicator.mobility_gen.impl.utils.global_utils import get_world

@ROBOTS.register()
class KayaRobot(WheeledMobilityGenRobot):
    ...  # full class definition

for recording_path in sorted(glob.glob(os.path.join(INPUT_DIR, "*"))):
    scenario = load_scenario(recording_path)   # now finds KayaRobot
    world = get_world()
    world.reset()
    scenario.enable_rgb_rendering()
    # ... rest of render loop (same as replay_directory.py)

See code/mobility-gen/05_replay_kaya.py for the complete working example.

Config / Data Format

config.json per recording:

{
  "scenario_type": "RandomPathFollowingScenario",
  "robot_type": "CarterRobot",
  "scene_usd": "/path/to/warehouse.usd"
}

state/common/<step>.npy is a numpy dict: position, orientation, joint_positions, joint_velocities, linear_velocity, angular_velocity.

Common Pitfalls

• ModuleNotFoundError: No module named 'isaacsim.replicator.mobility_gen': Extensions aren't auto-loaded. Pass --enable isaacsim.replicator.mobility_gen.examples to python.sh. All extension imports must come AFTER SimulationApp(...).
• AttributeError: 'World' object has no attribute 'initialize_simulation_context': Only the async version exists (initialize_simulation_context_async). For standalone scripts, call world.reset() instead.
• Replay KeyError: 'MyRobot': replay_directory.py only knows built-in robots. Write a wrapper script that registers your robot class before calling load_scenario().
• Custom robot produces no images during replay: Missing front_camera_* attributes, or build() passes front_camera=None. Add the attributes and call cls.build_front_camera(prim_path) in build().
• AttributeError: 'MyRobot' has no attribute 'chase_camera_base_path': chase_camera_base_path, chase_camera_x_offset, chase_camera_z_offset, chase_camera_tilt_angle are required by load_scenario() even for headless recording.
• Built-in replay fails to find robot class: Pass --enable isaacsim.replicator.mobility_gen.examples so the examples extension registers its robots/scenarios before load_scenario() runs.
• physics_dt mismatch: Recording stores the physics timestep in config.json; replay uses the same robot_type.physics_dt. Do not change robot params between record and replay.
• Occupancy map scale: MobilityGen consumes the OccupancyMap produced by occupancy-map. Ensure map.yaml origin and resolution match the USD world coordinates.
• Headless GPU: Set CUDA_VISIBLE_DEVICES=0 DISPLAY=:99 to avoid GPU contention with vLLM on GPUs 1-3.