The article deals with the problem of traffic in three traffic-light intersections in the street network of Iasi city. Usually, the existing traffic lights do not fully take into account the size, structure and parameters of the queue of vehicles waiting at intersections; which in turn affects the traffic capacity of the intersection. In the study, we used a representative set of measurements taken at the three intersections to interpret the variation in the crossing speeds of the traffic light intersection space by the vehicle columns using the linear regression method. We studied several categories of vehicles queuing before the stop line at the intersection. We made a statistical analysis of the time taken to cross three intersections by vehicles in different starting positions. The main results of the research showed estimates in good agreement with observational data as well as a significant reduction in vehicle speed (to 58%) when there are different categories of vehicles queuing at the intersection. This study can be used in the urban traffic diagnosis method, being part of a traffic study together with urban traffic forecasting and therapy.