SimReady assets for
self-driving
Autonomous vehicle stacks need realistic road environments, vehicles, cyclists, pedestrians, road debris, infrastructure. Rigyd converts 3D catalogs into physics-enabled assets with collision geometry and semantic labels for CARLA, DRIVE Sim, and Omniverse.
The problem
Why existing workflows fall short.
Long-tail scenarios need rare objects
AV stacks must handle construction barrels, dropped cargo, scattered debris, stalled vehicles. Building varied assets for rare, safety-critical scenarios is prohibitively expensive by hand.
Semantic labels and physics rarely coexist
Most AV asset libraries focus on either visual fidelity or simplified physics, not both. Full-stack testing needs perception-ready labels AND accurate collision behavior on the same object.
Scenario coverage is a data problem
Reaching 99.9% scenario coverage takes thousands of object variations. Manual asset creation caps scenario diversity well before the coverage level regulators expect.
How Rigyd helps
AI-native infrastructure that automates the hard parts.
Semantically labeled OpenUSD
Each asset ships with physics properties AND perception-ready semantic labels, enabling combined camera, lidar, and collision evaluation in CARLA, NVIDIA DRIVE Sim, or Omniverse.
Edge-case object library at scale
Upload reference scans, 3D catalogs of road debris, traffic cones, and construction objects and get physics-enabled versions in minutes, unlocking long-tail scenario coverage.
Calibrated for collision modeling
Mass, friction, and restitution values are calibrated for collision modeling at highway speeds, essential for realistic pre-crash, near-miss, and emergency-maneuver scenarios.
scenario coverage needed for safe AV deployment
cost reduction vs manual asset preparation
saved per 1,000-object scenario library
Scale your AV scenario library
Convert any 3D catalog into physically accurate, semantically labeled assets for self-driving simulation.
Starts at $29/month. 30 credits included.
Frequently asked questions
Does Rigyd output semantically labeled assets for AV stacks?
Yes. Each asset includes physics properties AND perception-ready semantic labels, enabling combined camera, lidar, and collision evaluation in CARLA, NVIDIA DRIVE Sim, or Omniverse without a separate annotation step.
Can Rigyd produce long-tail scenario objects at scale?
Yes. Enterprise API converts reference scans and 3D catalogs of debris, cones, construction objects, and stalled vehicles, the long-tail objects needed for the 99.9% coverage AV regulators expect.
Are physics values calibrated for highway-speed collision modeling?
Yes. Mass, friction, and restitution are tuned for collision modeling at AV-relevant speeds, essential for realistic pre-crash, near-miss, and emergency-maneuver scenario testing.
Does Rigyd support CARLA, NVIDIA DRIVE Sim, and Omniverse for AV workflows?
Yes. OpenUSD output is native to NVIDIA DRIVE Sim and Omniverse. For CARLA, Rigyd converts to .fbx with embedded physics metadata, which CARLA's asset importer consumes directly. Semantic labels propagate to all three platforms via SemanticsAPI (USD) or per-mesh tags (FBX), enabling perception model training across the AV simulation stack from one source.
Can Rigyd generate assets for adverse weather or low-visibility scenarios?
Yes. Material classification handles wet, snow-covered, and dirty variants of standard road objects (cones, signs, debris) with adjusted friction coefficients per condition. Rigyd doesn't generate weather effects (rain particles, fog), those are simulator runtime features, but it provides the surface-property metadata weather simulation needs to behave correctly against your scenario assets.
Related reading
In-depth guides on robotics simulation, OpenUSD, and SimReady assets.
21 MAY 2026
Synthetic data generation for computer vision in robotics
Real-world labeled robotics data is expensive and slow to capture. Synthetic data is fast and unlimited, but only useful if the underlying simulation has correct physics, semantic labels, and domain randomization. Here's the complete pipeline.
18 MAY 2026
Digital twin creation pipeline for manufacturing
A factory digital twin needs every object to behave physically, not just render. This is the end-to-end pipeline: CAD intake, BIM merge, physics layer, semantic labeling, simulation runtime, at the asset volumes (10K+ unique SKUs) real factories actually contain.
13 MAY 2026
How to set up mass, friction, and joint properties for robot training
The three pillars of robot physics setup, mass, friction, joints, determine whether your trained policy transfers to real hardware. Here's the calibration target for each, the schemas, and the common mistakes that quietly break training.
Explore more
SimReady for Isaac Sim
OpenUSD assets validated for NVIDIA Isaac Sim
Models for MuJoCo
Physically accurate assets for robot learning
Assets for Gazebo
Simulation-ready models for ROS 2
Warehouse Simulation
50,000+ SKUs need physics properties
Humanoid Robots
Whole-body interaction training assets
Sim-to-Real Transfer
Close the gap with accurate physics data