Virtual Joint Model for Synthetic Kinematic Data Generation

The patent creates virtual copies of patient joints and movements through simulation. Instead of collecting actual patient data, the system generates synthetic kinematic data sets by modeling joint anatomy, applying virtual inertial measuring units, and simulating movements under various conditions. This copying approach maintains data quality while eliminating the complexity of real patient data collection.

The patent replaces the mechanical process of physical data collection in clinics with a computational simulation system. The simulation unit uses virtual models and algorithms to generate kinematic data, substituting the need for physical measurement devices, patient interactions, and clinical infrastructure with a digital ecosystem that produces equivalent training data.

The system performs preliminary actions by pre-configuring virtual joint models with diverse anatomical variations, movement patterns, and environmental conditions before data generation is needed. The simulation unit can rapidly generate diverse data sets by varying parameters such as joint anatomy, movement speed, load conditions, and patient demographics, eliminating the need for time-consuming field collection.

The patent systematically varies multiple parameters including joint anatomy (varus/valgus angles, bone dimensions), movement characteristics (speed, range of motion), loading conditions, and patient demographics to generate diverse synthetic data. This parameter variation approach enables rapid creation of numerous diverse data sets without requiring actual diverse patient populations.

The patent embeds virtual inertial measuring units within the virtual joint model structure, creating a nested configuration where sensors are integrated into the bone models. This nesting allows the measuring units to be automatically positioned and oriented according to the joint anatomy, enabling efficient data generation while managing complexity through hierarchical integration rather than separate components.

The virtual inertial measuring units serve multiple functions simultaneously: they measure linear acceleration, angular velocity, and orientation of joint bones, and their data is used for both kinematic analysis and training AI/ML applications. This multi-functionality reduces the need for separate specialized measurement systems, improving efficiency while keeping the overall model manageable.