Creating a diffuse bicycle traffic assignment model for Switzerland

Traffic assignment models show the amount of traffic on roads and are a commonly used tool in traffic planning for motorised and public transport. The Swiss national model for passenger traffic NPVM of 2017 also includes such a model for bicycles. The short distances and lack of a hierarchical road network of bicycle traffic make this model too sparse to be useful in traffic planning. This thesis presents a methodology to take the bicycle traffic distribution model of the NPVM and create a denser traffic pattern. The traffic distribution model provides the number of bicycle trips between each pair of traffic zones in Switzerland, which will be assigned to the road network of OpenStreetMap using the bicycle routing service Valhalla. To distribute the route start and end points within each traffic zone, the jittering and disaggregation approach introduced by Lovelace et al. (2022) was used. The distribution was weighted by population and full-time equivalents. The methodology was applied to the NPVM bicycle distribution model for 2017 and for the base scenario of 2050. To validate the results the 2017 bicycle traffic distribution model was compared to observations from automatic counting stations across Switzerland using GEH-values, SQV and linear regression. The results show bicycle traffic on roads across the country, creating a dense traffic pattern within settlements and connections between them. The validation indicates that the model performed better than the bicycle traffic distribution done for the NPVM. While the results in urban areas are acceptable, the performance of the model was poor in rural areas.