Neural-network based modeling for stop&go behavior in real traffic flow

The first step towards an autonomous vehicle is adaptive cruise control (ACC) and stop&go maneuver systems since these kinds of systems adapt the speed of a vehicle to that of the preceding one (ACC) and get the vehicle to stop if the lead vehicle stops. There have been attempts to model stop&go waves via microscopic and macroscopic traffic models. But modeling the maneuver itself is presented only in a few studies. The purpose of this study is to design two neural network models for stop&go maneuver. These models are designed based on the real traffic data and model the velocity and longitudinal distance (spacing) with the front vehicle for the vehicle which performs a stop&go maneuver. Using the field data, the performance of the presented models is validated and compared with the real traffic datasets. The results show very close compatibility between the model outputs and maneuvers in real traffic flow.