Robot Virtual Commissioning

A core technology that pre-validates and optimizes real robot systems in virtual environments using Isaac Sim.
This dramatically saves time and costs while minimizing errors and risks that could occur in real environments.

PRINCIPLES

GPU-Based Digital Twin

Unlike traditional CPU-based simulations, Isaac Sim operates on a GPU-accelerated physics engine (NVIDIA PhysX).
This enables high-precision dynamics simulation that closely resembles reality.

  1. Ray Tracing

    Utilizing ray tracing, the latest graphics card technology, to realistically depict light reflections, shadows, and material textures. Such vivid virtual environments maximize the accuracy of Synthetic Data Generation (SDG) for AI robot training and sensor simulation.

  2. Digital Twin

    Virtually replicates real robot system movements, sensor data, and environmental factors, serving as a platform supporting the entire 'Design-Tuning-Deployment' cycle of robots.

FEATURES

Key Features of Simvis No-Code Automation Solution

  1. Diverse Asset Integration

    Integrates all assets including robots, sensors, and environments based on USD (Universal Scene Description). Supports various robot design formats such as URDF, MJCF, and CAD to maximize utilization of existing assets.

    Format Support :
    URDF, MJCF, CAD, etc.
    Integration Method :
    Pixar's USD-Based
    Advantage :
    Easy Asset Reuse & Collaboration
  2. ROS/ROS 2 Integration Bridge

    Supports seamless integration with ROS (Robot Operating System) and ROS 2, the most widely used by robot developers. This enables testing and validating real robot control code in virtual environments without modification.

    Supported Versions :
    ROS (including 2 Humble), ROS 2
    Features :
    Custom Message Generation, Real-Time Communication
    Advantage :
    Development Convenience & Real Robot Deployment Ready
  3. Synthetic Data Generation (SDG)

    Generate large-scale, high-quality synthetic datasets for AI robot training in simulation environments without the time and cost of data collection. Ensure diversity by randomly assigning various conditions such as lighting, background, and position.

    Key Tool :
    Isaac Replicator
    Data Type :
    Images, Depth Maps, Segmentation, etc.
    Advantage :
    Solving Data Scarcity Issues & Cost Reduction
  4. Advanced Sensor Simulation

    Physically accurate simulation of various sensors including cameras, LiDAR, IMU, and contact sensors using RTX technology. This enables verification and fine-tuning of sensor integration errors before actual sensor deployment.

    Sensor Types :
    Camera, LiDAR, IMU, Contact Sensors, etc.
    Rendering Technology :
    RTX (Ray Tracing)
    Advantage :
    Sensor Data Validation Identical to Reality

PROS & CONS

Advantages & Considerations

Accuracy
Advantages
Provides high-precision environments and sensor data resembling reality through RTX-based ray tracing
Considerations
Advanced tuning required to perfectly reflect all physical characteristics of real environments
Efficiency
Advantages
Dramatically reduce time and costs by virtually validating the entire development cycle without real robots (Accelerate Build-Test-Train cycle)
Considerations
High-performance NVIDIA RTX GPU hardware is essential (High initial investment cost)
Scalability
Advantages
Easy integration with various 3D tools and assets based on the Omniverse ecosystem
Considerations
Initial learning curve is somewhat steep for users without robot development and simulation experience
Operations
Advantages
Supports headless mode (CLI/Web) in addition to GUI mode for cloud or remote distributed simulation environments
Considerations
Sufficient disk space required for Omniverse Cache, etc.

APPLICATIONS

Key Application Areas

  1. Robot Cell & Factory Layout Validation

    Simulate optimal placement and robot paths before building new manufacturing lines or robot cells to pre-eliminate interference and inefficiencies.

  2. Logistics Warehouse AMR Path Optimization

    Virtually validate dynamic path planning, fleet management, and task allocation engines of Autonomous Mobile Robots (AMR) to maximize logistics efficiency.

  3. AI-Based Robot System Training

    Train AI-based vision and recognition models through Synthetic Data Generation (SDG), and build reinforcement learning environments to train robot control policies.

  4. Complex Task Safety Testing

    Ensure stability by repeatedly testing scenarios that are dangerous or impossible in real environments (e.g., collision tests, extreme environmental conditions) in virtual environments.