Dynamic Recognition Ability Model of Drivers at Entrance of Freeway Tunnel

摘要

In order to study the risk recognition ability of drivers at the entrance of a freeway tunnel, a driving test was carried out at the tunnel section. Based on measurement data from the field, the visibility level for obstacles at the entrance section of the tunnel was calculated through Adrian model. Twenty drivers were then chosen for an obstacle recognition test at the entrance section of the tunnel, and the dynamic recognition law for obstacles of drivers under different velocities was analyzed. Finally, RBF neural network was adopted to build a dynamic recognition model, and the recommended simulation values of visibility for obstacles under distinct velocities were 10.22 and 16.45. The results showed that RBF neural network model fit the recognition results of obstacle visibility and distance under dynamic conditions therefore the recommended values of visibility level can provide theoretical and practical bases for visual environmental improvement and illumination design at the entrance section of the tunnel.