Event-based networked control scheme for two-wheeled systems

Abstract

This paper provides an eficient and reliable networked optimal control design of a two-wheeled inverted pendulum system to achieve wide-range system stabilization. Linear quadratic regulator (LQR) theory and its mod-iied version with exponential term are initially cast into the LMI framework to guarantee stabilization as a selfbalancing robot. Next, an integral action is inserted into the LQR performance index to decrease the error substantially. Finally, the event-based policy is applied alongside LQR to decrease communication congestion over the network. An observer-based control is used; once the states are estimated, an optimal control law is generated to regulate the torque of the motors. Simulation results demonstrate that the system performance still optimal while the event-based regime is used.