Mixed Traffic Optimization with Platooning Technology Based on Stackelberg Game Theory

In this paper, we propose a game theoretical decision making framework for mixed traffic with platooning technology. Different from considering an interaction between multiple automated vehicles (AVs) in a same lane and an ego car (EC) on the other lane, we first use platooning technology to assemble the AVs to harvest the traffic efficacy. Therefore, the number of players shrinks to two, which could save plenty of computation time. However, the platooning technology offers multiple gap choices for the EC and the technology is reluctant to the EC's lane-changing behaviour, which are not considered in the previous literature. To cope with the above problems, a Stackelberg game is used to generate a proper decision for the platoon against the cut-in behaviour, taking into account the driving safety and ride comfort. To obtain mixed traffic optimization, we then adapt the platoon's acceleration range to discourage the EC's cut-in intention by increasing the cost of the EC side when the cut-in happens. A sufficient condition on the platoon's acceleration range which ensures the EC to stay in its initial lane as the optimal strategy based on Stackelberg game is derived. Numerical examples confirm the outcomes of our proposed method.