Radial basis function (RBF) neural networks are applied to predict the vehicle state with a driver under guidance.

The position, velocity, acceleration of the merging vehicle, difference between velocity and desired velocity and difference between acceleration and desired acceleration of the merging vehicle for the first <em>ta</em>/2 s are used as input features to output the velocity and position of the merging vehicle at <em>tp</em> moment. <br>将合流车辆（merging vehicle）在前<em>ta</em>/2秒内的位置、速度、加速度，以及该车辆的速度与期望速度的差值、加速度与期望加速度的差值作为输入特征，以输出<em>tp</em>时刻合流车辆的速度与位置。