Predicting the Time Until a Vehicle Changes the Lane Using LSTM-Based Recurrent Neural Networks

To plan safe and comfortable trajectories for automated vehicles on highways, accurate predictions of traffic situations are needed. So far, a lot research effort has been spent detecting lane change maneuvers rather than estimating the point in time actually happens. In practice, however, this temporal information might be even more useful. This paper deals with development system that accurately predicts to next surrounding highways using long short-term memory-based recurrent neural networks. An extensive evaluation based large real-world data set shows our approach is able make reliable predictions, most challenging situations, root mean squared error around 0.7 seconds. Already 3.5 seconds prior changes become highly accurate, showing median less 0.25 summary, article forms fundamental step towards downstreamed position predictions.