Walking Parameters Estimation Based on a Wrist-Mounted Inertial Sensor for a Walker User

This paper presents an algorithm for estimating walking parameters (step length, step speed, and walking distance) using a wrist-mounted inertial measurement unit for those who use a front-wheeled walker. Two walking styles are considered: step-by-step walking and continuous walking. A new zero-velocity detection algorithm is proposed for when the user walks step by step. Then, a standard inertial navigation algorithm is applied using a Kalman filter and zero-velocity update. For continuous walking, a step detection algorithm is proposed based on the bio-mechanical relationship between arm and foot movements. A step length estimation model was derived by testing 128 different features. The accuracy of walking parameters estimation was verified through 20-m walking experiments with five subjects.