LPV design for the control of heterogeneous traffic flow with autonomous vehicles

Abstract

The paper proposes a strategy to control heterogeneous traffic flow which contains both autonomous and human-driven vehicles. The purpose of the control strategy is to consider differences in the longitudinal driving characteristics of autonomous and humandriven vehicles. In the paper the modeling of the heterogeneous traffic flow based on the results of the VISSIM traffic simulator is presented. The traffic model yielded is in a Linear Parameter-Varying (LPV) form. The control design is based on the Takagi-Sugeno methodology, in which the performances, the constraints of the ramp-controlled interventions and the uncertainties are incorporated. The design task leads to an optimization with Linear Matrix Inequality (LMI) constraints. The result of the method is the optimal intervention of the freeway ramps with which traffic inflow can be controlled.