Updating Travel Demand O/D Matrices from Old Travel Surveys through More Recent Traffic Counts: a case study in Turin

In order to effectively implement any transport policy, it is essential to study the traffic demand and consequently to have the appropriate movements or trips information, expressed as origin-destination matrices. It is possible to say then, that the generation of origin-destination matrices is crucial for transport planning. Such matrices represent the amount of people travelling or the amount of goods being transported between different zones of an area. There are several ways to obtain origin-destination matrices. Traditional methods like surveys are generally expensive and time-consuming and that is why they are generally implemented only every 5-10 years. As a consequence, decision makers and analysts are often relying on old O/D matrices that need to be updated. This is a difficult and challenging task, especially when unforeseen conditions (e.g. pandemics or economic shocks) trigger an unforeseen and abrupt change in travel demand levels and patterns, compared to the latest available information. This thesis analyses to which extent the matrix estimation model based on the OmniTRANS software can help in achieving such goal. The tool is originally intended to estimate a matrix, based on a prior matrix and a set of constraints represented for example by traffic counts. We use it to update origin-destination matrices for car trips where the prior matrix is from an old survey whereas traffic counts are more recent. The performance of the developed method is tested in the city of Turin (Italy) and its surroundings. The matrices used in the estimation are obtained from the IMQ 2013 raw data, IMQ being the series of travel surveys performed in the region of Piedmont whose latest dataset dates back 2013. The traffic counts included in the estimation are located all over the city of Turin and belong to year 2019. The results obtained in the thesis allowed estimating an update of origin-destination matrices of car trips from year 2013 to year 2019 in the study area.