Dynamic Digital Twin

We are pioneering research where dynamic digital twins synchronize the digital world with any changes taking place in the physical world. Our solution is built on a Neural Radiance Fields (NeRF) which is a trained neural network that learns how to represent a space from any viewpoint. Standard NeRF represent a static scene. We are pushing our research efforts beyond static to introduce dynamic updates into a NeRF. Our Dynamic Digital Twin solution takes camera feeds that are provided by mobile agents, such as drones and robots or fixed cameras, and integrates the newly gathered information into the NeRF representation.

The Dynamic Digital Twin solution provides industrial operators with up-to-date augmented digital views of physical facilities, allowing the operator to monitor, analyze and control the physical facility based on the digital representation.

We successfully tested our proof-of-concept solution in an industrial warehouse environment where we conducted real-time inventory monitoring. Drones flew throughout the facility and took nightly inventory scans that provided real-time photorealistic representations of the warehouse shelves as items were added to and removed from inventory. The photorealism enables the warehouse workers to easily match what they see with the warehouse records, without having to be physically present. The digital aspect of our solution allows for overlays of product information related to each item. Warehouse managers can monitor all information directly from a tablet rather than spending the time to make a physical check. The Dynamic Digital Twin solution improved the accuracy of inventory management, cutting costs and minimizing risk, while increasing operational efficiencies and safety.

We have also tested the Dynamic Digital Twin solution at sporting events. The NeRF provided a view of the physical playing space using high-resolution cameras to capture all the action and render the players directly in the NeRF. Spectators could view any play from a self-selected angle, enabling real-time interactions and adding to the overall spectator experience. Coaches could see real-time statistics annotated on player representations in the digital twin, allowing them to make quick decisions on strategy and substitutions. After the games, athletes can use our solution to review their movements from any angle to determine how to improve their performance the next time out.