Evaluation of motorway reliability based on the identification of spot speed stochastic processes. The case of A22 motorway, Italy

Abstract

The evaluation of motorway reliability is a subject of great interest for researchers of traffic engineering. For easy recognizable reasons, this subject is very important in theoretical studies and for applications, e.g. evaluation of capacity and traffic control on motorway. In this paper the different conceptions of reliability in technical and scientific literature are compared with the following definition: reliability of traffic stream, for a time interval T, is “the conditioned probability that level of speed spot process does not decrease during T as far as the section in which limit density is reached, given that the level at instant zero is equal to the average speed corresponding to the flow rate in condition of stable flow”. Then, a criterion to calculate the motorway reliability is presented. The speed spot time-series are divided into sequences of events of random and homogeneous traffic processes. For every process one can calculate two parameters: flow rate and density; then, one can obtain the relationships between the parameters of spot speed processes and vehicular density. Using these relationships and a simulation procedure of spot speed process, one can obtain a formulation of the reliability of a traffic flow running along the offside lane of a motorway carriageway. Thus, through this formulation and the measurements of flow rate and speeds, one can evaluate the probability of instability on the carriageway under investigation. On the other hand, with an inverse procedure one can calculate the capacity of one lane in this way: putting close to 1 the reliability value and considering a 15 minutes time interval, the obtained flow rate represents capacity. In the paper, the procedure explained is used to provide, in real time, the approaching of the instability on the carriageway and to calculate the carriageway capacity for the A22 Brenner Motorway (Italy).