Personal identification and visualization of relationships by using human trajectories

Good communication among workers helps make comfortable atmosphere in an office and contributes to the increase company productivity. Therefore, relationship between workers is an important information the manager needs to know by examining the worker´s comfort level in an office environment. Moreover, the quantified information of the human relationship is useful for robots to provide services to working people in an office. There are many non-contact systems that can measure the positions of multiple persons in a room by using cameras and laser range sensors. However, who-is-who information is difficult to obtain in these systems without special devices, such as identification tags. This study proposes a personal identification method by using movement trajectories that are measured using cameras. Hidden Markov model is used to obtain who-is-who information. Moreover, visualization method of the relationship between workers is proposed based on the physical distance established among workers in an office throughout a day. The validity of the visualized relationships obtained by the proposed system is confirmed by comparing its results with that from self-evaluation questionnaires about a worker´s strength of relationship with their co-workers.