Side-slip Angle Estimation of Autonomous Road Vehicles Based on Big Data Analysis

The paper proposes a side-slip angle estimation method for autonomous road vehicles using a big data approach. The core of the solution is that on-board signals of numerous autonomous vehicles are available, which can be used to generate the side-slip estimator. The estimation is based on the ordinary linear regression method with OLS subset selection using large amount of data collected from the car sensors. The advantage of the proposed solution is that the numerically complex data mining operations are performed off-line, while the side-slip angle estimation of the vehicle using only its own on-board signals requires low computation effort. The efficiency of the estimation is presented through several CarSim simulations, in which the parameters of the vehicle and the road are varied. Moreover, the method is compared to the simulation results of a sensor fusion based Kalman filtering method.