Over the past 20 years, drive-by monitoring, using data from sensors in passing vehicles, has become increasingly popular due to its low cost and high efficiency at a network level. Despite significant advances, the dynamics of vehicles remains one of the main challenges to overcome before accurately identifying critical information about the infrastructure being crossed, bridges or road pavements. This paper introduces a novel approach to estimating bridge profiles (i.e. road surface profiles on the bridges) and vehicle properties using a fleet of passing vehicles. In this method, bridge profiles are first calculated using the innovative Inverse Newmark-beta integration method, and the cross-entropy optimization algorithm is employed to solve the problem. Numerical results demonstrate that the proposed approach is highly effective in extracting bridge profiles and predicting vehicle properties, even in scenarios with significant levels of measurement noise.