Linear Parameter Varying (LPV) based Observer Design for an Autonomous Vehicle

The autonomous driving is the most emerging trend in the current research era. The critical issues in this research area are the state estimation of vehicle trajectories and the nonlinearity of the vehicle model. In the last two decades, there are various research papers that have addressed this issue by using a different state estimation approach like Kalman filter, observer, etc. In addition to this, there is a huge development in the vehicle model to consider its nonlinear characteristics. The author in this paper has proposed a state observer in which the existing vehicle model is reformulated using a linear parameter varying (LPV) approach. In this LPV approach, the nonlinear vehicle model is restructured in such a way that this model is based on the parameter which is online accessible time-varying. Along with this, the state estimation of the vehicle in this method uses the global position system (GPS), by considering the course angle as a time-varying parameter of the LPV system. Further, in this method, to avoid heavy computation cost and complexity of the system, instead of the extended Kalman filter approach, an H 2 filter-based observer approach is used for state estimation purposes. This proposed technique is validated using the MATLAB/Simulink model and the results obtained in this are discussed in this paper. The results are shown in this paper having very minute errors and this model is robust to noise present in the measurement of sensors.