Lane change trajectory prediction using artificial neural network

Abstract

In this paper, the effectiveness of lane change (LC) trajectory prediction on the basis of past motion parameters of LC vehicle is studied. A vehicle’s LC trajectory is modelled as a time series and back propagation neural network is used for short-range and long-range prediction. Results using field data indicate that future LC trajectory cannot be predicted with sufficient accuracy using past motion parameters of the vehicle only. The results also show variation in the change of motion parameters during LC. This suggests external neighbourhood influence and need for incorporating this to increase the accuracy of forecasting.