Microsoft Kinect Sensor for Real-time Color Tracking Robot

This paper presents the application of the Microsoft Kinect sensor for low-cost color-based real-time target tracking for robotic application. The image color and depth measurements from the Kinect sensor are utilized in identifying and localizing the target object and tracking it. Closed loop proportional-integral (PI) control algorithm is implemented to control a mobile robot for a smooth tracking of the target object. The iRobot Create is used as the development platform for the experiments. The voice recognition feature of the Kinect sensor is also accessible for simplifying the robot-human interaction when commanding the robot to track specific colored objects. Preliminary results of successful implementation of the real-time color tracking technique is demonstrated by letting the robot identify and track target objects with specified colors.