摘要
Aiming at the advantages in active safety control of in-wheel motor drive vehicles, the classified decision-control-execution is proposed to the lane departure assistance system under the premise of ensuring the stability of the vehicle. The dynamic extension boundaries which vary with vehicle speed, road curvature and road adhesion coefficient are designed in the decision layer by using the neural network algorithm, which divides different dangerous degrees into different decision regions.The active differential steering controller, the electronic differential controller and their coordinated control strategy are designed in the control layer, and the corresponding control is carried out among different decision regions. In order to avoid the dangerous situation of secondary deviation caused by the existence of drift angle when the vehicle is rectified to the lane centerline, the heading angle controller is then designed in the control layer. The different torque distribution modes are divided according to the current vehicle speed, road adhesion coefficient and lateral acceleration in the execution layer, and the economic distribution, stability distribution, joint distribution are correspondingly executed among different torque distribution modes. Finally, Carsim and Matlab/Simulink are used for simulation verification.
基金
supported by the National Natural Science Foundation of China under Grant Nos.U1564201
the Science Fund of Anhui Intelligent Vehicle Engineering Laboratory under Grant No.PA2018AFGS0026