Design and Visio control of Two-Wheeled Inverted Pendulum Mobile Robot

Abstract

This paper addresses Visio and path following control problem of a nonholonomic Two-Wheeled Inverted Pendulum Mobile Robot. We propose control architecture based on two control layers. A speed inner loop control scheme is first designed based on state feedback technique to ensure stability of the inverted structure of the robot. A second outer loop control scheme is proposed to help the robot navigate along a desired path formed by a set of way points. It is designed inspiring the model predictive control technique. The elements of the predictive control, which are the cost function, controls and constraints, must be defined and specified: the use of different trajectories group in the control can adapt the behavior of the robot to different displacement phases. The obstacle detection architecture based on the attributes of color and texture has been developed to be implemented on an Raspberry PI and is designed as a generic high-speed image processing device. The optimization criteria are based on a maximization of performance in terms of image processing per second and a minimization of consumed resources. Our obstacles detection algorithm consists of three main steps: the color transformation, the calculation of the color and texture attributes and their classification.