Design Of The Mobile Robot Navigation Platform Towards The Polarization Sensor

The mobile robot is a cross-technology involving mechanism, electronics, sensors and control technology. Working in a complicated and unknown environment, the mobile robot needs a precise navigation system to improve its capabilities of decision and execution. In this dissertation, a wheeled mobile robot is developed, and on the basis of the polarization sensor, a navigation system for the mobile robot is built.The mobile robot is built up based on a hybrid structure. It is mainly composed of an execution level and a decision level. The execution level adopts a four-wheel, differential motion structure which is controlled by a MCU control system. In order to improve the responsivity of the execution level, a behavior-based method is adopted. The decision level is the core of the mobile robot control system. It can achieve the functions like fusing information of the sensors, implementing the decision algorithms, and generating the control instructions for the execution level. Assisted by Visual C++ tools, we program a friendly man-machine interface for the decision level. Additionally, a fuzzy logic controller is integrated in the decision level. The fuzzy logic control technology is widely applied in mobile robot systems with non-holonomic constraints because of its needless of the accurate mathematical model.The heading information is the most significant parameter of the mobile robot navigation system in terms of its influence on the cumulate dead-reckoning errors. Odometer is one of the most popular sensors in mobile robot navigation system. However, its fundamental principle is the integration of incremental motion information over time, which inevitably leads to the unbounded accumulated errors. Combining the odometer and the polarization sensor, we develop a novel navigation system for the mobile robot.Some outdoor experiments are performed to test the capability of the mobile robot navigation system. The result indicates that the navigation system built in this dissertation exhibits an excellent performance. Nevertheless, some errors can be found in these experiments if observing the result carefully. We statist and analyze these errors in the end.