Mohammad Jafar Jahanbazi - Robotic Research Institute Amirkabir University of Technology Tehran, Iran
Rahele Nick Amal - Robotic Research Institute Amirkabir University of Technology Tehran, Iran
Heidar Ali Talebi - Department of Electrical Engineering Amirkabir University of Technology Tehran, Iran

Following a particular person among the people is the important topics of area human interaction with robots. Because it requires too much processing for human detection and identification of a specific person among other people, it is one of the challenges for the robots. In this paper, a suitable algorithm is easy and low computational presented. In this regard, the depth image is used. The Kinect sensor addition to low prices has many features including depth of picture in output. With this sensor and Microsoft company applications for the sensor extracted human skeleton. In this study, the skeleton of the person considered and the cost is made according to it. When a human skeleton was found, the robot has found the person's expense by cost function and followed him. However, data obtained from the sensor swing frame joints is high and it is causing problems for diagnosis. It also considered the total length of some members to be removed. Practical results show if the person is in the scope sensor and it is well recognized structure, robot can find and follow him.

mobile robot; human following; depth image; Kinect sensor