Human whole body motion characterization from embedded Kinect

Non-verbal communications such as kinesthetics, or body language and posture are important codes used to establish and maintain interpersonal relationships. They can also be utilized for safe and efficient human robot interactions. A correct interpretation of the human activity through the analysis of certain spatio-temporal and dynamic parameters represent an outstanding benefit for the quality of human machine communication in general. This paper presents an effective markerless motion capture system provided by a mobile robot for sensing human activity, in non-invasive fashion. We present a physical model based method exploiting the embedded Kinect. Its performances are evaluated first comparing the results to those obtained with a precise 3D motion capture marker based system and to data obtained from a dynamic posturography platform. Then an experiment in real life conditions is performed to assess the system sensitivity to some gait disturbances.