Abstract: A car like
robot, path tracking is a problem of practical importance in the
field of robotics and autonomous vehicles. The aim is to have a mobile robot
follow a given reference path autonomously. Introduction of modern technologies
such as power steering, antilock braking and traction control has helped to
reduce the accentuation of the drivers. Present scenario in the world of
automobile is complete automation of vehicle. To gain trust on fully automated
vehicles we need promising technology which can guarantee fast and a safe
journey. The same similar situation arises in the case of car like robot. While
taking the system for an application the parameters such as velocity, time are
of prime concerns. The increasing velocity increases the vulnerability to slip
and roll over depending on the dynamics and condition of the environment. In
this scenario it is necessary to design a controller that continuously takes
the action on the system.
Model Predictive Controller (MPC) is a controller which anticipates control
input depending on system current states. The Prediction and Control Horizon
depends on the states of the system. So, for high speed applications more
storage space of the computed states may arise, which intern results to large
computational time of Model Predictive Controller. To resolve this problem an
Explicit Model Predictive Controller is implemented. Multi parametric model predictive control (mp-MPC) enables application of
MPC controllers with low on-line computational demand and fast sampling by
shifting the burden of optimization to the off-line controller design stage.
Since the lateral dynamic model of the system is considered for the ease of
computation the bicycle model of the system is also done. To achieve the
fast and secure traction of automated vehicle, prediction is imminent.
Computational complexity of conventional MPC can be overcome by using Multi
Parametric Model Predictive Control successfully.