Cooperative Highway Lane Merge of Connected Vehicles Using Nonlinear Model Predictive Optimal Controller

Abstract

Of all driving functions, one of the critical maneuvers is the lane merge. A cooperative Nonlinear Model Predictive Control (NMPC)-based optimization method for implementing a highway lane merge of two connected autonomous vehicles is presented using solutions obtained by the direct multiple shooting method. A performance criteria cost function, which is a function of the states and inputs of the system, was optimized subject to nonlinear model and maneuver constraints. An optimal formulation was developed and then solved on a receding horizon using direct multiple shooting solutions; this is implemented using an open-source ACADO code. Numerical simulation results were performed in a real-case scenario. The results indicate that the implementation of such a controller is possible in real time, in different highway merge situations.