Human following behavior for an autonomousmobile robot

Abstract: Robots are used more and more to change people's life. In recent times, some robots have found their place in houses. Those robots are specic ones as for example robotic vacuum cleaners. The next step in robotics' evolution is to make people live with robots in everyday life. This project attempts to make a robot follow a person. To achieve that, a people detector using a RGB-D sensor is implemented. The result of this detector is sent to a tracking algorithm using depth segmentation and a region growing algorithm. At the end, a navigation algorithm containing path planning and obstacle avoidance is used to make the mobile robot move. This method can make the robot follow a person that walks slowly in front of it. Suggestions are presented to improve the method and adapt it to everyday life. Thanks to this study, mobile robots can be used in personal home and assist people. An application could be for example to follow a person that needs medical assistance and help this person when necessary.