The increasing complexity of industrial automation demands sophisticated solutions for managing and optimizing robot fleets. At InOrbit, we're focused on empowering companies to maximize the value of their robotic deployments through advancements in Physical AI and simulation.
Today we're pleased to highlight our ongoing collaboration with NVIDIA, which is instrumental in driving these advancements. Together with partners like Globant, we're enabling organizations like Genentech, a member of the Roche Group, to achieve complex operational goals and unlock unprecedented levels of efficiency.
At the core of this collaboration is InOrbit Space Intelligence™, a platform designed for AI-powered robot orchestration. With Space Intelligence, InOrbit leverages real-world robot data to create accurate behavioral models, enabling intelligent task allocation and seamless operation within shared, multi-system workspaces.
To provide a clearer understanding of this technology in action, we've created a brief overview video showcasing the key elements of our latest collaboration:
As illustrated in the video, InOrbit utilizes real-world data to model robot behaviors and create semantically annotated maps for heterogeneous fleet orchestration. Time Capsule captures detailed robot data, which is then used to create robust models of actual robot behavior within a real operational environment.
Leveraging these data-driven models, InOrbit facilitates the simulation of robot operations and the training of custom AI agents. These agents are designed to optimally allocate robots to fulfill business objectives while maximizing overall fleet throughput, even across diverse robot models and brands.
To ensure the accuracy and adaptability of these simulations, InOrbit employs multi-sensory data from robots to build photorealistic digital twins using neural 3D mapping techniques, that can be used within the advanced simulation environment of NVIDIA Omniverse. This integration allows for rigorous testing of learned agents under a wide range of conditions in accurate virtual environments.
Simulation data can then easily be brought back into InOrbit and the exact same, optimized missions can be run in the real world. This establishes a continuous feedback loop, where data from real-world missions informs and improves the system. This is central to InOrbit Space Intelligence™, enabling the ongoing optimization of AI-driven workflows.
InOrbit is excited to announce this news in conjunction with NVIDIA’s GTC 2025 AI conference. For a deeper dive into this announcement and the details of our technology, we encourage you to read today’s official press release.
Through strategic collaborations and a focus on leveraging Physical AI and advanced simulation, InOrbit is committed to empowering the robotics industry to achieve new levels of operational excellence.
