MIT Energy Initiative – Consortium of Corporate Members (http://energy.mit.edu/research/mobility-future-study/)
Moshe Ben-Akiva, Christopher Zegras, Carlos Lima Azevedo, Jimi Oke, Sean Hua, Michael Choi
- Create simulation tool for future urban mobility patterns and different transportation energy scenarios
- Account for emerging and future travel modes and vehicle types, fuels, and technologies
- Incorporate consumer attitudes to new mobility technologies via experimental survey design
- Gain knowledge on the potential impacts of future mobility scenarios on different world-wide prototypical cities
The goal of this project is to develop a viable framework for analyses and predictions of passenger responses to possible future de-carbonization policies at the urban level. This framework will incorporate behavioral preferences with regard to emerging and upcoming vehicle technologies and mobility services. These will be realized via extensions to an existing agent-based simulation model for urban networks. To test its applicability to many different urban scenarios that exist at the global level, we will determine prevailing city attributes and typologies that describe mobility, energy and emissions characteristics world-wide. From these typologies, we will develop prototypical cities based on existing and simulated data. We will then perform scenario analyses to predict outcomes of potential policy options.
To answer questions regarding future mobility choices, energy consumptions and fuel types in the transportation sector for different cities, and in order to realize sensitivity of these measures to potential polices at the global scale, we propose a simulation framework for prediction of travel needs, travel behaviors, and travel conditions. This framework will partially be built on the existing data analysis and travel simulation tools that are already developed or under development in various on- going research projects, and have been continuously revised and extended in the recent years at the MIT Intelligent Transportation Systems Lab and at the Singapore-MIT Alliance for Research and Technology, Singapore.
Figure 1: Research task plan
Introduction to simulation framework (SimMobility) and its capacity for scenario testing
SimMobility integrates various mobility-sensitive behavioral models within a multi-scale simulation platform that considers land-use, transportation and communication interactions. It focuses on impacts on transportation networks, intelligent transportation services and vehicular emissions, thereby enabling the agent-based simulation of a portfolio of technology, policy and investment options under alternative future scenarios. In practice, SimMobility incorporates three different sub-models (see Figure 1):
- Long-term (LT) simulator: The simulation time step is in the range of days to months to years, and agent decisions include house location choice, job location choice, land development, car ownership. It is a land-use and transport (LUT) simulator, with a market transaction bidding model.
- Mid-term (MT) simulato: The simulation time step is in the range of seconds to minutes and agent decisions include route choice, mode choice, activity pattern and its (re)scheduling, departure time choice. SimMobility mid-term is a mesoscopic simulator designed for activity- based modeling, with explicit pre-day and within-day behavior including re-routing and re- scheduling, and multiple transport modes.
- Short-term (ST) simulator: The simulation time step can be a fraction of a second and agent decisions include driving lane changing, braking, accelerating, gap acceptance, but also route choice. SimMobility short-term model is a traffic micro-simulator, extended with a communications simulator as well as pedestrians and public transport.To answer questions regarding future mobility choices, energy consumptions and fuel types in the transportation sector for different cities, and in order to realize sensitivity of these measures to potential polices at the global scale, we propose a simulation framework for prediction of travel needs, travel behaviors, and travel conditions. This framework will partially be built on the existing data analysis and travel simulation tools that are already developed or under development in various on- going research projects, and have been continuously revised and extended in the recent years at the MIT Intelligent Transportation Systems Lab and at the Singapore-MIT Alliance for Research and Technology, Singapore.
Figure 2: SimMobility Framework
Across all these scales, SimMobility implements the activity-based modeling paradigm, i.e. all choices are ultimately tied to the agents’ goal of performing activities on the corresponding time scale. Agents can be grouped in flexible fashions, from households to firms, and can have varying roles including operators, bus drivers or real-estate agents.
In terms of applications to scenario-testing contexts, SimMobility is a research laboratory via which a variety of future mobility concepts can be tested. These include network and access restrictions and pricing scenarios, parking prices, adaptive traffic control systems, real time traffic advisory, autonomous mobility on demand (AMOD), among others.
For this project, SimMobility LT and MT are of interest. The representation of generic daily mobility patterns and its effects on long-term individual choices (such as vehicle purchase) will be studied. Furthermore, extensions to the current framework are also proposed to accommodate modeling of future technologies, travel modes, alternative fuels, ownership choices, and de-carbonization policies and restrictions.
Under this project, this simulation framework will be extended to understand how much impact carbon prices, carbon control policies, innovative vehicle technology and mobility services may have on the future of the transportation sector. Predefined decarbonizing policies and their implications such as car registration fees, public transportation subsidies, congestion pricing, and parking policy are considered as given inputs in this proposed research, and reflected in the simulation scenarios. In addition, future fuels and drivetrains characteristics such as fuel price and efficiency and fueling characteristics (e.g. availability of fueling stations) are also input and reflected in the simulation.
To allow for flexibility in the simulation of mobility for efficiency assessment of future mobility scenarios, we propose certain extensions to SimMobility to incorporate future travel modes (supply and demand sides). These modes include flexible mobility on demand (FMOD), autonomous mobility on demand (AMOD), private autonomous vehicles, and other possible mobility services and vehicle technology. The proposed extensions also take into account alternative fuel types and drivetrains, and their associated range anxiety or refueling behavior. Behavioral data collected will allow for an estimation of consumer preferences towards these technologies and their integration within SimMobility.
The alternative fuel type/drivetrains would include:
- Internal Combustion Engines with Enhancements
- Hybrid Electric Vehicles (HEV)
- Plug-in Electric Vehicles (BEV and PHEV)
- Fuel Cell Vehicles (FC)
- Compressed Natural Gas Vehicles (CNG)
Prototype city generation
The creation of prototype cities will be a means of not only testing and validating our framework, but also obtaining insights into patterns in multiple cities around the world. Thus, we will develop one SimMobility city model for each urban typology that is determined from the data mining process. To keep this task feasible within the existing timeframe and the research budget, and also in order to stay reasonably loyal to the real-world city characteristics and network structures, our strategy will be to obtain easily and publicly available data (e.g. roadways, public transit services, geo-referenced population) for one representative city from each cluster. Subsequently, we will synthesize the missing data components required for the SimMobility model of these [semi-]virtual cities.
When choosing a representative city from each cluster, we will take into consideration the availability of public datasets, as this will improve the chance of accessing essential network data for the city synthesis task. However, it is reasonable to assume that we may end up with one or more city clusters for which we may not access any publicly available data. For such possible clusters, we will have to synthesize the entire virtual city model from the scratch, using methods we will develop during the course of this project.
Figure 3: Overview of virtual city generator attributes
- Data collection and predictive modeling to realize travel preferences in the presence of future fuel/technologies in the transportation sector,
- Extensions to the existing simulation tool (SimMobility) to accommodate future mobility specifications in the transportation system,
- Data collection and city classification into a set of prototype cities representative of global urban diversity,
- Development and validation of prototypical virtual city generator that is able to generate virtual urban areas for simulation,
- Demonstration and evaluation of future mobility scenarios in two large-scale real network simulation models (Boston and Singapore),
- Scenario testing and sensitivity analyses in prototypical virtual cities to realize implications of future policies on mobility patterns, emission levels, and energy consumptions.