Modeling and design of smart mobility systems

Abstract Smart mobility entails the coordinated use of technology to increase the quality and efficiency of mobility provision, while minimizing or reducing the space consumed and externalities generated by transportation supply. In order to plan and design for such smart mobility systems, it becomes necessary to understand how their dynamics play out in response to urban design elements. Smart city urban design, therefore, cannot be conceived of without considering details about the mobility system that will serve it. This chapter starts with an introduction to transportation demand planning and modeling approaches. It considers traditional four-step travel demand modeling and compares it to later approaches that consider the effects of individual activity–travel behavior and decision-making. It also provides an overview of integrated transportation and land-use planning, where the interaction of the transportation system and its resulting accessibility affects location choice behavior of firms and households. Agent-based modeling approaches capable of capturing the high degree of complexity and coordination in emerging smart mobility systems are introduced. In what follows, two contrasting approaches to the question of integrated urban systems design with smart mobility systems are set forth. In the first approach, the agent-based traffic simulation, MATSim, is integrated into an urban design workflow such that an agent-based travel demand this rapidly produced from typical urban design digital drawings such as a road network layer, massing of buildings by use, and other design elements such as the location of transit stops, parking space, active mobility network, etc. In this sketch planning approach, future mobility systems are made available to serve the transportation demand that would be plausibly produced from the urban design. In this way, the urban designer can then evaluate and compare designs and mobility systems, tinker with various design elements and system parameters, and rerun the simulation to have the design and mobility flows that it will produce coevolve. This approach is tested in an application of a prospective neighborhood design in Singapore. A hypothetical new transit system, served by autonomous vehicles and capable of moving the passenger between any combination of stops in the study area without predetermined routes, is tested for three different network designs. The experiment reveals that, in order to provide a predefined level of service, each of the three designs will require a different fleet size, while producing radically different experiences to the individuals who will be living and working in the neighborhood. The second experiment explores how the spacing of transit stops in a neighborhood affect the performance of such a dynamically routed transit service, resulting in different levels of use of the road network and associated externalities that residents will be exposed to. Insights from how this smart mobility system will perform can inform trade-offs about where to invest in infrastructure expenditure: should one build more transit stops or should one rather invest in maximizing walkability to a smaller set of transit stops and so be able to bundle rides together?