From simulation to real-world deployment: Unity Simulation Pro early access

Sep 22, 2026|5 Min
Isaac Seah
Isaac Seah - Unity Technologies
Principal Technical Product Manager
Unity Robotics Simulation Pro Early Access

Robotics teams have been building on Unity Industry for years, working across a wide variety of challenges, from autonomous driving validation and manufacturing digital twins to surgical device development. Along the way, each of those teams solved the same set of problems independently: The robot’s physical description had to be rebuilt by hand, sensors had to be written from scratch, and integrations with open standards like ROS had to be built and maintained. By bringing these capabilities into one supported package, Unity Simulation Pro can help teams spend less time building simulation infrastructure and more time validating robotics systems.

Today, we're launching Unity Simulation Pro in Early Access, Unity's robotics simulation package built for these workflows. Unity Simulation Pro is compatible with Unity 6.3 and newer and requires a Unity Industry license.

What Simulation Pro Includes

Simulation Pro is a fully supported robotics toolkit that helps developers import, simulate, and iterate on robotics systems directly in Unity. This package brings a greater level of support to a wide range of use cases, from sensor simulation and robot training to virtual commissioning of robotics systems, making it easier than ever before to build robotics simulation and applications with Unity.

At a high level, it includes:

  • URDF (Unified Robot Description Format) importer, a drag-and-drop experience that brings a robot’s existing physical description into Unity with its kinematic chain intact
  • Sensor simulation out of the box, covering LiDAR, image, and inertial measurement unit (IMU) configurations
  • ROS 2 compatibility, connecting Unity to systems operating on Robot Operating System 2
  • Headless Linux build target, so teams can run simulations without a graphical interface and scale them in parallel on existing hardware

Robotics teams came to Unity because they needed real-time rendering, accurate physics, and sensor simulation they could trust. At the core of it, they wanted a foundation to seamlessly transition robots from the lab to the real world. Simulation Pro is that foundation, and Unity Industry customers can access this today.

Alex Blum
Alex Blum - Unity
Chief Operating Officer

How teams leverage Unity for robotics

Validating Autonomous Driving Without the Test Vehicle

TIER IV is working to make autonomous driving available to all by opening up development to improve accessibility and scalability. To support that vision, TIER IV built AWSIM, an open-source simulator developed with Unity that helps teams test, validate, and optimize Autoware, open-source autonomous driving software, in a more realistic and efficient environment.

Autonomous driving simulation places demands on rendering that typical interactive 3D applications do not, because every frame has to feed multiple high-resolution sensors in real time. Some AWSIM scenarios call for eight or more cameras running simultaneously while preserving sensor fidelity, which standard optimization techniques could not deliver on their own. Working in Unity, the team built a set of specialized optimizations to carry that load: staggered rendering across cameras to flatten frame spikes, time-scale controls to allocate more resources per frame, asset optimization, and the flexibility to use different rendering pipelines within a single project to balance accuracy against frame rate.

TIER IV also used the Unity job system and built-in Physics to hold real-time performance across vehicle, traffic, and pedestrian simulation. For hardware-in-the-loop testing, the team analyzed raw sensor packets and reproduced equivalent output from Unity, so camera sensor setups could be tested in the loop before anything was deployed to a physical test vehicle, reducing physical test vehicle hours and on-road testing cost.

100,000 Labeled Frames From a Plant With No CAD

The Korea Institute of Industrial Technology (KITECH) needed training data for AI that understands how people and robots share a manufacturing floor. This is one of the harder datasets to collect in the physical world. A well-run plant is laid out so that close-proximity moments between a worker and a machine never happen, and staging them repeatedly in order to film them is not a reasonable option.

To measure proximity, video, depth, robot joints, and human pose all had to exist at the same moment, so that distance and safety-zone labels could be calculated consistently. To start, KITECH built a digital twin of a metal casting plant from point cloud scans, since no CAD source existed. The team then built physics, sensor, and simulation systems on top of that digital twin. The sensor rig used custom sensor components matching the real rig, ensuring every observation shared one simulation clock.

They also leveraged Unity's Articulation Body components for robot simulation, replaying recorded joint trajectories so that inertia and contact were computed together, ensuring motion was physically valid rather than merely plausible.

This physical fidelity enabled KITECH to automatically generate over 100,000 pre-labeled training frames in 14 weeks, eliminating the need to manually stage dangerous proximity scenarios on an active factory floor while ensuring reliable sim-to-real transfer.

Robotic-Assisted Surgery Digital Twin

Medtronic built a data-logging and playback pipeline in Unity around a digital twin of its Hugo robotic-assisted surgery system, capturing 195 procedures across five global sites.

“Unity is our first-stop-shop for all prototypes so we can get things in front of surgeons quickly.”

Matthew Euliano
Matthew Euliano - Medtronic
Senior Software Engineer
Medtronic built a data-logging and playback pipeline in Unity around a digital twin of its Hugo robotic-assisted surgery system

The recordings became a reusable dataset for engineers and researchers. Instead of waiting for operating room access to answer a question, teams could load a recorded procedure, scrub through it, isolate the relevant moment, and inspect it from any angle. The same data also supported deterministic replays that recorded procedure data to train new AI models and inform the design of next-generation devices, so each recording continues to produce value long after the procedure itself has ended.

Availability

We are especially interested in hearing from teams running robot validation, sensor testing, or synthetic data generation at scale across manufacturing, logistics, warehousing, and other advanced robotics sectors.

Connect with the Unity Simulation Pro development team on Unity Discussions. We look forward to learning more about what you're building.