Modeling Longitudinal Following Control Based on Preceding Vehicle Motion Predictor

Comparing the cooperative adaptive cruise control (CACC) with adaptive cruise control (ACC), acceleration information improves the tracking performance, but that would require communication. In this paper, an autoregressive model (AR model) based motion predictor is proposed to forecast the preceding vehicle movement at the next step by using the successive historical state information that only sensed and collected by the local sensors. On the basis of it, combining the existing works about car-following models, a new longitudinal following control model towards AVs is developed under the structure of optimal velocity theory. The stability analysis of the new model is derived and the results show that the stability condition can be indeed facilitated and the stable region in the headway-sensitivity space can be accordingly enlarged by utilizing the predicted results of preceding vehicle motion. The experiment and simulation are carried out based on the NGSIM dataset and the results are in great agreement with the analytical results, which demonstrates that the proposed longitudinal control strategy can improve the following performance of automated vehicles effectively.