Evaluating the Potential of Simulated 2D Radar for Gait Parameter Extraction

Radar-based gait analysis provides a contactless, costeffective alternative. This study evaluates a simulated 2D radar model for estimating spatiotemporal gait parameters, using motion capture (mocap) as the reference standard. 30 participants completed treadmill walking trials, with mocap data mapped onto a 3D animated model for radar simulation. The radar demonstrated high accuracy in cadence (97.02%) and step time (99.12%) but lower agreement in stance (76.72%) and swing percentage (48.82%), with root mean square deviation (RMSD) values of 17.28% and 17.21%, respectively. The simulation framework enables algorithm refinement before hardware implementation, improving efficiency and reducing costs. Future work will focus on enhancing gait phase detection by upgrading to 4D radar for improved spatial and angular resolution and limb differentiation.