Trajectory Tracking-Based Lane Changing in Connected Vehicles With Unreliable Communication

Abstract

We study lane changing in connected vehicles, in the presence of surrounding vehicles with non-uniform velocities, under unreliable communication. We adopt a dynamic trajectory tracking approach, wherein a subject vehicle (SV), modeled by the nonlinear dynamic bicycle model, must follow the desired dynamics governed by the dynamic lane-change trajectory planning (DLTP) model, for changing lanes efficiently. We design a robust and adaptive Coupled Multiple Sliding Mode Control (CMSMC) law that ensures trajectory tracking despite time-varying disturbances acting on the SV. The controller accepts position and velocity inputs from the neighboring vehicles, through vehicle-to-vehicle (V2V) communication. The design is initially validated for an ideal communication channel that does not drop any data packets. Subsequently, we derive a sufficient condition that ensures the boundedness of the tracking error even when the communication channel exhibits packet drops. Using an appropriate model for packet drops, we derive an upper bound on the packet-transmission interval for each vehicle communicating with the SV, and also outline the associated condition for update of the controller. The DLTP model is validated using real-world field traffic data extracted from NGSIM; and the analytical results are substantiated through numerical computations, and simulations performed on SUMO. Our study contributes towards the co-design of controller and communication algorithms for connected vehicles under unreliable communication.