Robotics applications tend to add even more accuracy, robustness and reliability since even more autonomy is required. Research areas where is applicable are autonomous guided vehicles (AGV) and wheeled mobile robots (WMR). Navigation is a main task in their structure. Because of previously discussed and perform such task, an offline path planner is presented. Next, WMR is designed with dead-reckoning characteristics. Since data uncertainty is presented, a vision system that feedbacks the actual position (called state) of the WMR is used to eliminate dead-reckoning errors. In this paper, a proposed architecture is presented to obtain a robust navigation with a comparison between fuzzy and classical control for a trajectory path.