Initial Contact and Toe-Off Event Detection Method for In-Shoe Motion Sensor

Initial contact (IC) and toe-off (TO) events are tools for measuring and analyzing human gait. As a simpler method for detecting gait events that depend only on inertial measurement unit (IMU) signals is needed, in this study, we propose a simpler signal feature-based method for detecting gait events that is feasible for use in in-shoe motion sensor (IMS) systems, and these exact features are used to determine the timings of IC and TO according to biomechanical knowledge. We then evaluate the precision of the method. Twenty-six healthy subjects were recruited to participate in the experiments, during which an IMS along with a Vicon 3D motion analyzer was applied to measure the trajectory of the foot and to judge the IC and TO timings. Temporal features of the foot to ground kinematic waveform at the time of IC and TO are newly discovered by synchronizing the two systems. The temporal precision of an algorithm for automatic IC and TO detection is evaluated on the basis of root mean square error (RMSE) and intraclass correlation coefficient (ICC). The RMSE of the TO detection was 1.22%, and that of the IC was 1.40%. The ICC of the TO detection was 0.7011, and that of the IC was 0.7721. The results demonstrate the high detection accuracy and reliability of this simpler IC and TO automatic detection algorithm for IMSs.