Developing customized cargo bike routes based on individual preferences and vehicle specifics

Abstract

Background Providing adequate routing is crucial for all road users traveling varying tours. Existing routing algorithms focus on cars and conventional bikes that are used privately. These offer customized settings based on the users' needs. To date, there is no tailored solution offering optimal cargo bike routes taking into account special requirements and preferences of commercial cargo bike users.

Goal We aim to fill this gap and develop a solution determining specific cargo bike routes in urban areas, possibly improving route efficiency and driver satisfaction. In addition, we aim to answer the following research questions: What influence do factors such as vehicle type and infrastructure have on the route choice of cargo bike drivers? To what extent can personalized navigation solutions increase the efficiency of cargo bike use?

Methods Our approach combines various methods and uses data from different sources to generate routes for commercial cargo bike users. The approach builds on data from empirically collected cargo bike trips. These are 30,000 cargo bike routes that have been recorded under real world conditions by users of various economic sectors, such as logistics, construction and manufacturing (craftsmen), service industries, retail, and the public sector. Trip data includes additional information, such as the characteristics of the used type of cargo bike. In addition, we use infrastructural and spatial parameters derived mainly from OpenStreetMap, such as bike infrastructure, street type, steepness/slope, urban green, connectivity, share/repair facilities. Parameters further include route characteristics of specific relevance for cargo bikes, such as stairs or narrow passages. Using an in-house routing algorithm (Strazoon, developed by 7P), as well as common routing engines such as Google Maps, we create a set of alternative routes for each tracked route. Discrete (route) choice models are then used to derive influencing factors for cargo cyclists' route choices. Potential influencing factors in the model are travel time, street type, type of bike infrastructure, surface, motorized traffic as well as vehicular and person-related factors. Quantitative outputs are then validated qualitatively by conducting ride-alongs with cargo bikes couriers and interviews with experts. Based on the results the routing algorithm is enhanced considering personalized input.

Results and conclusion Results show routing preferences of users of various cargo bikes in commercial use. The contribution offers up to date scientific results for cargo bike route choice. The routing solution developed helps to increase the economic potential and the acceptance of the use of cargo bikes in commercial use.