Investigating the Effect of Look-Ahead Distance on Automated Steering Performance

Abstract

In this study, effect of look ahead distance on autonomous vehicle control is investigated. A simple output feedback control law with feedback of lateral and heading error at preview distance, is employed to analyze closed loop behavior against various look ahead distance, l s values. Firstly, analysis are performed with linearized vehicle model in complex domain with closed-loop pole properties such as settling time and damping ratio. Then, automated lane change performance is investigated in time domain. Throughout the simulation studies, a nonlinear vehicle model with 3 degrees-of-freedom yaw plane dynamics having nonlinear magic formula tire model is used. Allowable range of look ahead distance l s is investigated in terms of stability, safety/comfort and lane change performance. Stability analysis are performed with closed loop pole locations in complex plane for ls values in the range of 0-25 [m]. Safety/comfort analysis are performed with automated lane change maneuver by comparing peak values of lateral acceleration and side slip angle for ls values in the range of 2-25 [m]. Lane change performance is analyzed with the peak values of lateral distance overshoot and steering command magnitudes for ls values in the range of 2 - 25 [m].