A Computer Vision Application for the Kinematic Measurement of a Cylindrical Mobile Robot

In the last few years, the interest in Computer Vision is growing more and more. Thanks to the reduction of the hardware price and to faster Personal Computers, Computer Vision is used in simple applications in place of traditional measurement systems. In this paper, we present a Computer Vision application in the field of robotic kinematic. In the first part, we shortly describe a new Cylindrical shaped Mobile Robot, in the second part a measurement system based on computer vision is presented to estimate the position, the linear speed and the pitch angle of the vehicle running on a testbed. The Vision system is based on a Pattern Matching algorithm, after a calibration of the whole vision system and scenario. This work starts from the necessity to measure the kinematic behavior of the vehicle without using the traditional encoders that influence somehow the vehicle motion. We carried out some experimental tests in order to compare measurements with optical encoders and with computer vision system. As conclusion of our experiences, we could use this innovative system to quickly measure the linear motion data of a known shape mobile robot