Communication delay compensation for string stability of CACC system using LSTM prediction

Abstract

Compared to traditional adaptive cruise control (ACC), cooperative ACC (CACC) can improve the response sensitivity of the following vehicles by using additional information, e.g., the acceleration of preceding vehicles, that is transmitted via inter-vehicle wireless communications. Thus, a platoon with CACC mode obtains a shorter time headway, thereby increasing road throughput while guaranteeing traffic safety. However, delays are common in wireless communication due to complex traffic conditions. The degradation of wireless communication significantly influences the string stability that refers to the attenuation of disturbance in the upstream direction of a platoon. Therefore, this study proposes the use of a deep learning method, i.e., the long short-term memory (LSTM) neural network, to predict the acceleration of the preceding vehicle by using data from onboard radar sensors. It provides the CACC platoon with another option to obtain additional information if the quality of wireless communication worsens. This type of CACC is referred to the LSTM control. Simulations proved the applicability of the LSTM control by using the next generation simulation program (NGSIM) data, in which the accuracy (goodness-of-fit index) of the LSTM prediction reached 0.766, and the LSTM control can compensate the communication delay to maintain string stability when the communication delay exceeded 0.115 s.