Design and real-time control of a new structure two-wheeled robot

• Main Author: Sayidmarie, O.
• Other Authors: Tokhi, M. O.
• Published: University of Sheffield 2016
• Subjects: 629.8
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.686515

The work presented in this thesis deals with the design and real-time implementation of a control system for a new configuration of two-wheeled robot. In real life applications, two wheeled robots are considered to carry payloads of different sizes at different positions and different motion speeds along the vertical axis. Such parameters will have an impact on the robot stability and the control mechanism, hence their detailed study is essential for robust performance of the system. This work investigates the impact of the dynamic change of the payload position on the system damping characteristics whilst the robot is in its balancing state. A mathematical model is developed to describe and study the dynamics of the two-wheeled robot system. Accurate tilt angle measurement is achieved through applying a sensor fusion realized by complementary filter. A PID control strategy is considered to balance and position the robot assuming a fixed location of the payload. Then the controller is extended to provide the robot with the ability of self-standing from its rest position without human interaction. The developed controller and the sensor fusion filter are implemented on a microcontroller development board. The results show that the implementation of the controller fulfils the requirement to balance the robot. The two-wheeled robot test rig was modified to accommodate the payload actuation unit and improve the overall performance. Consequentially, the controller of the robot is also modified with the sensor fusion algorithm enhanced by implementing Kalman filter. The controller comprise a PID feedback with a feedforward approach. Furthermore, gain scheduling approach is utilized to ensure smooth and fast braking of the two-wheeled robot. The control approach is extended to an intelligent controller, where a PD-like fuzzy controller is design. A binary coding technique is developed for real-time implementation of the fuzzy controller. Such coding and implementation eliminates the need to store a big lookup table for the controller rules. The controller is tested in laboratory experiments and its robustness is demonstrated with application of various disturbance forces on the system. In order to evaluate the performance of the new configuration two-wheeled robot, various experiments are conducted under different conditions. The results demonstrate that the proposed two-wheeled robot configuration and the proposed control approaches form a solid foundation and a framework for assisted mobility of disabled and elderly people.