Abstract: A customised adaptive cruise control (ACC) system based on model predictive control (MPC) and driving style identification to accommodate various driving styles while ensuring car-following, comfort, and fuel-efficiency performances. A series of real-world vehicle experiments are carried out to gather the driving data of 66 randomly selected drivers in order to determine the controller parameters that correspond to various driving styles. The experimental data is then clustered using an unsupervised machine learning technique. A driving style classifier is created using supervised machine learning on the basis of this information, and it can be used to identify drivers' driving styles online. The control issue with the customised ACC system is thus defined as a multi-objective optimisation issue that may be resolved using the MPC approach. The simulation findings demonstrate that the suggested personalised ACC system can provide varying performances and cater to the needs of various driving styles.
|
DOI:
10.17148/IJARCCE.2023.125113
