The detailed end-to-end digital twin is based on NVIDIA Omniverse simulation platform and includes AGILOX’s fully integrated software stack. This provides a solution that will enable users to emulate and test application scenarios in a virtual representation of reality. This approach reduces costs and risks in the planning and implementation of AMR projects, while digital twins offer high reliability and flexibility for automation applications in the future.
A New Dimension in Robotic Simulation: The Power of Digital Twins
AGILOX’ innovation in robotic simulation starts with its use of detailed digital twins for each AMR model. Through this approach, every function and feature of a robot can be virtually tested, simulated, and optimized long before it operates in the physical world. The creation of these digital twins begins with CAD modeling software where AMR models are designed. These 3D models are then converted into the Universal Scene Description (USD) format, and made compatible with the NVIDIA Omniverse platform.
Omniverse’s USD framework provides a shared 3D environment to enhance the visual fidelity of these models, making them appear more realistic while keeping them static initially. Then, AGILOX applies a series of plug-ins, such as sensor behavior, to add further realism and interactivity. These modules simulate essential components of AMRs, such as cameras, LiDAR, and other sensor functionalities. Once these sensors are integrated, the digital twin becomes dynamic, closely emulating the behaviors expected in a real-world scenario.
Through the NVIDIA Isaac Sim, AGILOX can simulate the digital twin in a virtual environment that mirrors actual warehouse or production workflows. Different parameters can be tested, allowing the digital twin to assess how the robot responds to various stimuli. This process ensures a high level of accuracy in detecting potential issues and optimizing the robot’s capabilities.
From Static Models to a Dynamic Simulation Environment
AGILOX’ digital twin journey doesn’t stop with a visually realistic model; the true breakthrough lies in achieving a fully functional, dynamic digital twin. This transition is made possible by connecting the static model to an emulator – a custom-built, in-house development that integrates AGILOX’ proprietary robotics software stack directly into the digital twin. By emulating the behaviors of the entire system, including AGILOX’ AMR-specific mapping, navigation, safety protocols, and localization, AGILOX allows for a seamless transition from virtual to physical application.
Unlike conventional testing methods that might focus on individual components in isolation, AGILOX’ approach simulates every critical aspect of the AMR. This includes collision avoidance, object recognition, and sensor-based navigation. By integrating all software components into the simulation process, AGILOX ensures that the robots are ready for deployment, thereby minimizing the likelihood of unexpected issues after implementation.
The Core Innovations Behind AGILOX’ Approach
AGILOX’ approach stands out due to two core innovations that set it apart from other robotics companies:
- Digital Twin Precision: By employing detailed digital twins for each AMR model, AGILOX can conduct exhaustive virtual testing, identifying any potential issues and optimizing performance without needing a single physical prototype in the future. This will reduce development costs and accelerate timelines, as costly and time-consuming physical FAT (Factory Acceptance Testing) will no longer be necessary.
- Comprehensive Software Stack Integration: AGILOX integrates its entire software stack into the digital twin, encompassing key functionalities like mapping, localization, navigation, safety, and collision avoidance. This holistic approach enables a high degree of realism in simulations, ensuring that the AMRs behave as they would in actual environments and allowing for the testing and refinement of the robots’ full functionality in one go.
By combining these innovations, AGILOX has created a development ecosystem that is faster, more reliable, and ultimately more cost-effective than traditional robotic development processes, helping customers implement solutions even faster in the future.
