Researches On The Intelligent Vehicle Path Following Based On The Driver’s Intention

Human factors are the most important factor for traffic accidents. So it is necessary to help the driver perform the driving task by developing the intelligent vehicle combined with many kinds of advanced technology. In that case, to ensure the safety of vehicle driving have became one of the hot research topic at present. However, many advanced driving assistance systems centering on the intelligent vehicle ignore the driver’s intention when evaluating the situation on the road, so it often leads to a wrong result because of the assistance systems. According to this phenomenon, this paper puts forward the man-machine coordinated control strategy based on the driver’s intention, guides the intelligent vehicle to follow the target path, and the main research work as follows:(1) Set up the Driver-in-Loop simulation experiments. Under the virtual road scenarios, the driver realizes his own driving intention by handling actuators such as the steering wheel, the accelerator pedal, the brake pedal and so on. The driver’s operating data can be measured by angle sensor and accelerator pedal sensor. Then the real-time data gathered by Data Acquisition Card PCI-6251 is transferred from the Real-Time terminal to the host by data sharing and saved in the host. A database can be established by these data, and lane change models can be trained offline by the database.(2) Establish a hybrid model based on hidden markov model(HMM) and support vector machine(SVM). Many kinds of the driver’s lane change intention recognition method have been summarized, also the advantages and weakness between HMM and SVM have been analyzed, then this paper proposed the hybrid model based on HMM and SVM which combined the temporal and classification characteristics. Five kinds of lane change models’ parameters can be getted by offline training, then the driver’s lane change intention can be recognized based on HMM-SVM on-line identification in the Driver-in-Loop simulation experiments. Simulation results show that this proposed hybrid model can recognize the driver’s lane change intention more accurately when compared with the classified approach only with the HMM or SVM. Recognition rate is reached as high as 98%, and takes only 0.006 second, which shows it has an excellent performance in real time.(3) Put forward a man-machine coordinated control strategy. According to the experiment’s requirements, the two degrees of freedom vehicle model with yawing motion and lateral motion is built, the prediction model equation is deduced by the preview-follower theory, and the optimizing index is based on lateral position deviation and the heading angle deviation between the real path and the target path. At last, the optimal MPC steering angle can be calculated. On the other hand, cost function and the value of intention comparison is considered to determine the coordination coefficient, then the final steering wheel angle is confirmed by the driver’s input and the MPC input. Simulation results show that the man-machine cooperative control strategy based on the driver’s intention can guide the intelligent vehicle to follow the target path, also this strategy achieved an excellent performance.