Modeling Car Following with Feed-Forward and Long-Short Term Memory Neural Networks

摘要: Abstract The paper investigates the capability of modeling car following behavior by training shallow and deep recurrent neural networks to reproduce observed driving profiles, collected in several experiments with pairs GPS-equipped vehicles running typical urban traffic conditions. input variables are relative speed, spacing, vehicle speed. In model, we assume that reaction is not instantaneous. However, it may occur during a time interval order few tenth seconds because both psychophysical driver’s process mechanical activation braking or dispensing traction power wheels. Experimental results confirm reliability this assumption highlight network outperforms simpler feed-forward network.