Analysis of Deep Learning Model for Trajectory Prediction of Vehicle

In this research, Long Short-Term Memory (LSTM) model and Gated Recurrent Unit Model (GRU) have been applied to the open-source HighD dataset to check the trajectory prediction of surrounding vehicles. The LSTM model has given better results as compared to the GRU model with an accuracy of 86%. The Mean Square Error (MSE) method is used to understand the learning effect of the models. Prediction errors were calculated and measured in the longitudinal direction to get the accuracy of the learning models in more detail. If the vehicle will change its path or lane in the longitudinal direction then will get the trajectory of that vehicle.