This paper presents the design and implementation of a perceptual system for the mobile robot using modern sensors and multi-point communication channels. The data extracted from the perceptual system is processed by a sensor fusion model to obtain meaningful information for the robot localization and control. Due to the uncertainties of acquiring data, an extended Kalman filter was applied to get optimal states of the system. Several experiments have been conducted and the results confirmed the functioning operation of the perceptual system and the efficiency of the Kalman filter approach.