Automatic recognition of standing long jumps regularity using inertial sensors

Standing long jump is a complex motor test that is typically used for the functional assessment of lower extremities. This study introduces a new approach in the analysis of rotational movement of body parts during performance of standing long jump. An experiment was conducted for collecting the jump database from 10 healthy trained students using inertial sensors. Segment and joint angle signals were calculated, and their shape was analysed and described with 10 features that were used for the distinction of jump correctness. A rule-based algorithm was developed for the recognition of SLJ regularity and achieved 91.67% accuracy. Compared to a decision tree trained using cross-validation, the developed algorithm showed better performance; however, additional testing on larger datasets is needed.