Research On Intelligent Control For The Stability Of A Two-wheeled Carrier In Wide-range

The two-wheeled vehicle is a special kind of wheeled mobile robot. Through analysis we find that the system is characteristic with multi-variables, nonlinear, strong coupling, uncertain parameters, and the environment in which the car move is complicated, so does the control task that needs to finish balancing and path-planning and path-tracking at the same time. Hence the two-wheeled vehicle is an ideal platform for testing and learning various control methods. Considering the quite limited range of leaning angle that linear controller can be stablelized, this thesis focus on how to achieve stability in a wider range by means of intelligent control methods.After summarized and has profited from the domestic and foreign self-balancing cars, our vehicle is designed by cooperating with classmates. With deep analysis in dynamics and kinematic, the non-linear model of the car is obtained. This paper concentrates in intelligent control of the self-balencing car which is expected to be stable in a wider range of leaning angle, mainly using two methods: First, apply the fuzzy-linearization method (FLM) to get the model of the system and design the linear controller. Specific jobs include obtaining a new linear model which is weight sum of local linear models (LLMs) and designing a linear controller based on the new model. Second, obtain the fuzzy control rules by utilizing the adaptive neural-fuzzy inference system (ANFIS). Specific jobs include respectively designing linear controllers for each LLM and using ANFIS to study input-output data from the linear controller to form fuzzy control rules. Finally, various comparison researches have been done to above control methods, including how the system responses in case of initial leaning angle, system external disturbance and internal parameter change. Simulation results indicate that the leaning angle range in which the system can be stabilized is notablely enlarged.