A digital twin is a digital model of a real-world product, system, or process that streamlines project management by integrating data from each stage of the product lifecycle. Incorporating real-world data into the model supports processes like simulation, system integration, and process monitoring. Human digital twins are particularly promising for enhancing human safety and performance in various settings, including industrial, athletic, clinical, and home applications. Advances in computer vision, AI, and machine learning have made human pose estimation and motion tracking with vision cameras possible in almost any environmental setting, though most current applications focus on post-processing of data.
Inventors at the University of Cincinnati have leveraged these technologies to develop a standalone software for real-time human digital twinning capable of pose estimation, activity recognition, and motion prediction. The software uses a 3D kinematic model of the human body and advanced filtering framework to interpret data from a variety of asynchronous sensors, such as cameras, laser sensors, and wearable devices. This allows for greater accuracy and efficiency in both real-time and offline applications, such as industrial worker training and safety, human motion monitoring and analysis in healthcare and sports, and human-machine interactions.