Experimental results in vision-based lane keeping for highway vehicles

In this paper we address the problem of lateral control of vehicles for highway lane keeping using vision system. First we present a suitable simplified model of the plant and show its agreement with typical highway experimental conditions. Then, we formulate the control problem and propose a SISO control strategy based on driver behaviour emulation. The controller design is carried out using classical loop-shaping techniques and taking into account the model uncertainty. The experimental tests, conducted on Italian highways using a Fiat Brava 1600 ELX provided by Centro Ricerche Fiat, showed the good performances of the designed control system.