About the project
This work investigates how dynamics can be identified from data and independently evaluated. Different methods are first compared on synthetic reference systems with known properties. Measurement noise, data separation and multi-step prediction are part of the investigation. These benchmarks do not yet constitute validation on human gait data or a clinical analysis product.
Focus areas
- Comparison on known dynamical reference systems
- Investigation of measurement noise and prediction error
- Reproducible evaluation with separate datasets
Our role and project context
Research and software work by Dr Daniel Friemert; not a commercially released product