Support Vector Machine Based Lane-Changing Behavior Recognition and Lateral Trajectory Prediction

Abstract

With the development of technology, vehicle trajectory prediction and safety decision technology has become an important part of active safety technology. Among them, the vehicle trajectory prediction technology can predict the vehicle position, speed, and other motion states in the predicted period according to the current and historical vehicle running state, and the prediction results can provide support for judging the vehicle safety in the predicted period. In order to analyze the above problems, this study fully extracted the main feature information from the vehicle lane change track data with the help of the powerful nonlinear learning and high pattern recognition ability of support vector machine, and conducted identification modeling for the actual lane change process of the vehicle and predictive analysis of the vehicle lateral movement track. First, the lane-changing behavior of vehicles was analyzed, and the vehicle lane-changing execution stage and 10 influencing factors that could characterize lane-changing behavior were determined based on NGSIM to extract the data of lane-change-related variables. Then, a lane-changing recognition model based on gridsearch-PSO is proposed. In the Matlab environment, the model has a test accuracy of 97.68%, while the SVM model without optimization parameters has a recognition accuracy of only 80.87%. The results show that the model has strong classification ability and robustness. Finally, by using the polynomial model for lateral movement trajectory fitting, K-fold cross-validation method is used for fitting polynomial model fitting test.

Risk Factors and Predictive Models for Peripherally Inserted Central Catheter Unplanned Extubation in Patients With Cancer: Prospective, Machine Learning Study

“…We used AUC and TOPSIS methods for a comprehensive and rigorous screening of the optimal predictive model. The SVM algorithm in the optimal model robustly encompasses the data and reduces the model's complexity through linear regression with insensitive loss functions in a high-dimensional feature space [45]. Importantly, external validation of the model using independent data demonstrated significant predictive superiority.…”

Literature Review on Maneuver-Based Scenario Description for Automated Driving Simulations

“…For this, it is important for the system to predict maneuvers of the ego vehicle as well as surrounding vehicles. By this, an AD function is enabled to align its reactions with the intended behavior of the driver [44], [79], to remind the driver of actions that need to be taken [77], to warn him or her in case the intended maneuver cannot be executed safely [43] or just keep the driver informed about the environment [84]. Examples for maneuvers whose detection and classification is desired apart from the road following maneuver are turning maneuvers [43], [45], [46], [77], [79], lane change maneuvers [43], [44], [46]- [53], [55], [64], [75], [77], [79], [85], braking [50], and drifts into another lane [54].…”

Checkpoint Data-Driven GCN-GRU Vehicle Trajectory and Traffic Flow Prediction

“…require the manual definition of suitable features, by extracting useful information from individual historical trajectories as model inputs, to achieve the next trajectory point prediction. Feng et al 21 used support vector machine regression to achieve vehicle lane changing behavior and trajectory prediction.…”