An Efficientnet Based Method for Autonomous Vehicles Trajectory Prediction

Abstract

Autonomous driving refers to the use of computer, network, control, communication and other technologies to achieve real-time control of the vehicle. This technology needs to predict the future trajectory of surrounding moving targets as accurately as possible, and adjust its own driving path as needed. The key to this technology is how to predict the trajectory of the vehicle as accurately as possible based on the data of the vehicle itself, nearby moving objects, and traffic lights. Based on a large amount of processed road data, this paper uses the EfficientNet model to predict the trajectory of autonomous driving. We mainly use the EfficientNet network model to train and test the data set, and use the loss value to adjust the model to improve the accuracy of model prediction. Through experiments, we finally found that the EfficientNet model has better prediction performance than the two models VGG16 and ResNet34.