Application of nonlinear complementary filters to human motion analysis

Analyzing human motion has involved the use of the Kalman filter for calculating the orientation of the inertial measurement unit (IMU), based on many current motion analysis systems. However, its pitfalls, including high computation costs and difficult implementation renders itself sub-optimal in the clinical environment. Meanwhile, a filter developed by Mahony et al., which was developed for vehicular studies, possessed features that overcame the former´s shortcomings. In this study, we compare the efficacies of these two filters to determine the applicability of the Mahony filter in the clinic. Approximations from both systems, on the upper and lower body, prove the Mahony filter to perform as consistently and accurately as Kalman filter. Thus, the cheaper alternative presented by Mahony filter may be applicable in sports and medicine.