Prediction of driver cut-in intention towards platoon vehicles in mixed traffic flows

• Main Author: Wang, Xinran
• Other Authors: Su Rong
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering
• Online Access: https://hdl.handle.net/10356/181274
• Institution: Nanyang Technological University

It is of great significance to use technology to assist driving behavior to improve vehicle driving safety. In the process of driving, lane change behavior will bring greater safety risks. To solve the above problems, this dissertation proposes a lane change intention prediction algorithm based on the Transformer, which can predict the lane change of vehicles in the next 2 seconds and provide early warning and auxiliary functions for vehicle driving. After the model was proposed, the NGSIM US-101 data set after data screening and preprocessing was used to train the model, and verified its feasibility. At the same time, compared with other models such as BN, SVM, MLP, and LSTM under the same circumstances, the intent prediction algorithm based on Transformer can achieve better performance in three different scenarios: left lane change, right lane change, and lane keeping.