Modeling and analysis of a holonomic and nonholonomic novel differential drive wheeled robotic system -

Abstract

Modeling and control of wheeled mobile robots over rough terrain has been an essential task for robotic researchers for applications such as search and rescue and unmanned exploratory missions. Modeling the locomotion of a wheeled robot on rough terrain yields a highly constrained system of equations of motion. This work presents the modeling and control of a variable-diameter differentially- drive robot with a single actuator. The forward motion of this model can be dynami- cally related to the rotation of a central disk whereas the steering motion of the model can be related to the translation of this disk. The motion of this novel robotic platform design is captured via a set of differential algebraic equations (DAE). In this thesis, a stabilization technique is developed and used to reduce the DAEs of motion to a set ordinary differential equations. This stabilization method proved to be adequate to model the motion of the platform, not only on flat terrain but also on rugged uneven terrain. The developed model was used to simulate various inputs in an open-loop architecture to analyze the systems motion and its interactions with bumps in the terrain. A final task was to utilize the developed model to design controllers for the system inputs to perform two tasks, trajectory following and bump or disturbance mitigation. The stabilization of the equations of motion developed in this thesis, proved to be helpful in linearizing the dynamics the system. Accordingly, the linear system was used to analyze the systems controllability as well as to develop a Linear Quadratic Regulator which was tuned to provide robust control for the system.