© COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only. The needs for monitoring systems to be used in houses are getting stronger because of the increase of the single household population due to the low birth rate and longevity. Among others, gait parameters are under the spotlight to be examined as the relations with several diseases have been reported. It is known that the gait parameters obtained at a walk test are different from those obtained under the daily life. Thus, the system which can measure the gait parameters in the real living environment is needed. Generally, gait abilities are evaluated by a measurement test, such as Timed Up and Go test and 6-minute walking test. However, these methods need measurers, so the accuracy depends on them and the lack of objectivity is pointed out. Although, a precise motion capture system is used for more objective measurement, it is hard to be used in daily measurement, because the subjects have to put the markers on their body. To solve this problem, marker less sensors, such as Kinect, are developed and used for gait information acquisition. When they are attached to a mobile robot, there is no limitation of distance. However, they still have challenges of calibration for gait parameters, and the important gait parameters to be acquired are not well examined. Therefore, in this study, we extract the important parameters for gait analysis, which have correlations with diseases and age differences, and suggest the gait parameters extraction from depth data by Kinect v2 which is mounted on a mobile robot aiming at applying to the living environment.