Ground reaction force estimation method based on a single rgb camera using spring walking model

We proposed a ground reaction force (GRF) estimation method from the input of RGB video capturing human walking. To convert kinematic motion data from a video into kinetic GRF data, we used a spring-loaded inverted pendulum (SLIP) model that reflects the relationship between the center of mass (CoM) and GRFs. The trajectory of the CoM was calculated from the waist tracking of the video image. The model parameters and initial conditions were tuned to best match the simulated CoM and the data. The GRF is then calculated from a SLIP model. The proposed method well estimated the GRF with RMSE 9.26 ± 2.44 % and 10.33 ± 1.39 % for the vertical and anterior-posterior directions, respectively. The results imply that the SLIP model could serve as a framework to extract kinetic information out of kinematic data, which are measured by a regular video camera.