The variation in stride length is an important index of gait stability. The evaluation requires a device that can accurately measure the stride length during ground walking. However, there are currently no practical and reliable measurement systems for clinical use. The purpose of this study was to improve the stride estimation method using a single inertial sensor and to verify the accuracy of stride length estimation. Stride length was estimated by integrating the horizontal acceleration at the ankle joint. To reduce the impact acceleration at ground contact, which is the primary factor causing estimation errors, we mounted an inertial sensor on the lower leg. In the present method, the following four factors contribute to the stride length estimation error: the sensitivity of the inertial sensor, consideration of the directional angle, detection of integral intervals, and estimation of the relative position vector from the inertial sensor to the ankle joint. This study presents solutions to these problems and verifies the accuracy of stride length estimation by comparing it with optical motion capture measurements. The results indicated that the sensitivity of the inertial sensor improved the accuracy of the stride length estimation. Furthermore, we found that the standard deviation of stride estimation per step was approximately 32 mm, regardless of cadence and stride length. We conclude that the stride length estimation by the acceleration integral method is applicable to the estimation of the average stride length; however, it is difficult to apply it to the evaluation of stride variation from step to step.
