Neural Network Control for the Balancing Performance of a Single-Wheel Transportation Vehicle: Gyrocycle

Abstract

A single-wheel mobile robot has been developed for carrying a driver. Since a single-wheel mobile robot carries a human driver, the size and weight are larger compared with other sing-wheel mobile robots. To maximize the gyroscopic effects, Gyrocycle is designed to have two flywheels that need to be synchronized. In addition to the synchronization of two flywheels, Gyrocycle is tested for the robust balancing perfromance by unknown payloads. A neural network control method is used to control the balance. Experimental studies are conducted to verify the perfromance by the neural network controller