Gait Parameter Acquisition While Chasing Resident Using Home Robot

Abstract Gait parameters are key information to know the physical health of elderly people and if necessary, the parameters are used to avoid unnecessary injury due to unexpected falling resulting from weakened muscle. Many methods to estimate gait parameters have been studied such as the one using motion capture cameras. However, most existing methods require expensive sensors and markers attached to a person. This type of method is only feasible when it is used in a testing room. Thus it is difficult to obtain the gait parameters during normal life. In this paper, an algorithm for a home robot to acquire gait parameters while chasing a resident in a house is proposed. The home robot has a Kinect v2 sensor on its top to measure the distance. As this is a home robot always standing by the resident, the gait parameters will be always acquired and stored in the database. This robot does not need a special testing room. The home robot chases the resident from behind to estimate the physical health of the resident. The gait parameters such as knee joint angle, stride time, stride length and walking speed are precisely obtained by this algorithm. The algorithm of acquiring the gait parameters and chasing has been tuned to be used in a house with an elderly person. Several tests have been conducted considering scenarios of using this robot in a daily life.