Automatic Lane Merge based on Model Predictive Control

Autonomous driving has been regarded as the most promising industry since last decade. Among a variety of functionalities an autonomous vehicle has, the automatic merging maneuver is one of the most challenging ones because the maneuver has to be finished in a dynamic traffic environment within limited distance. This paper proposes an integrated path planning and trajectory tracking algorithm based on Model Predictive Control to achieve automatic lane merge in a mixed traffic environment with traditional vehicles (controlled purely by human drivers), semi-autonomous vehicles and fully autonomous vehicles. A bicycle model of vehicle dynamics is used as the prediction model in the algorithm design, while a high-fidelity model with non-linear tyre dynamics is employed in simulation. Moreover, a lane selection function with an add-on threshold function has been used to ensure the safety of the maneuver. The comparison of the simulation results between the proposed algorithm and a bench-marked two-layer control strategy has been given to demonstrate the effectiveness of the proposed controller.