Carlos Lima Azevedo

Microsimulation of Demand and Supply of Autonomous Mobility On Demand

Agent-based models have gained wide acceptance in transportation planning because with increasing computational power, large-scale people-centric mobility simulations are possible. Several modeling efforts have been reported in the literature on the demand side (with sophisticated activity-based models that focus on an individual’s day activity patterns) and on the supply side (with detailed representation of network dynamics through simulation-based dynamic traffic assignment models). This paper proposes an extension to a state-of-the-art integrated agent-based demand and supply model—SimMobility—for the design and evaluation of autonomous vehicle systems. SimMobility integrates various mobility-sensitive behavioral models in a multiple time-scale structure comprising three simulation levels: (a) a long-term level that captures land use and economic activity, with special emphasis on accessibility; (b) a midterm level that handles agents’ activities and travel patterns; and (c) a short-term level that simulates movement of agents, operational systems, and decisions at a microscopic granularity. In that context, this paper proposes several extensions at the short-term and midterm levels to model and simulate autonomous vehicle systems and their effects on travel behavior. To showcase these features, the first-cut results of a hypothetical on-demand service with autonomous vehicles in a car-restricted zone of Singapore are presented. SimMobility was successfully used in an integrated manner to test and assess the performance of different autonomous vehicle fleet sizes and parking station configurations and to uncover changes in individual mobility patterns, specifically in regard to modal shares, routes, and destinations.

Modeling reaction time within a traffic simulation model

Human reaction time has a substantial effect on modeling of human behavior at a microscopic level. Drivers and pedestrian do not react to an event instantaneously; rather, they take time to perceive the event, process the information, decide on a response and finally enact their decision. All these processes introduce delay. As human movement is simulated at increasingly fine-grained resolutions, it becomes critical to consider the delay due to reaction time if one is to achieve accurate results. Most existing simulators over-simplify the reaction time implementation to reduce computational overhead and memory requirements. In this paper, we detail the framework which we are developing within the SimMobility Short Term Simulator (a microscopic traffic simulator), which is capable of explicitly modeling reaction time for each person in a detailed, flexible manner. This framework will enable modelers to set realistic reaction time values, relying on the simulator to handle implementation and optimization considerations. Following this, we report our findings demonstrating the impact of reaction time on traffic dynamics within several simulation scenarios. The findings indicate that in the incorporation of reaction time within microscopic simulations improves the traffic dynamics that produces more realistic traffic condition.

A Sensitivity-Analysis-Based Approach for the Calibration of Traffic Simulation Models

In this paper, a multistep sensitivity analysis (SA) approach for model calibration is proposed and applied to a complex traffic simulation model with more than 100 parameters. Throughout this paper, it is argued that the application of SA is crucial for true comprehension and the correct use of traffic simulation models, but it is also acknowledged that the main obstacle toward an extensive use of the most sophisticated techniques is the high number of model runs usually required. For this reason, we have tested the possibility of performing a multistep SA, where, at each step, model parameters are grouped on the basis of possible common features, and a final SA on the parameters pertaining to the most influential groups is then performed. The proposed methodology was applied to an urban motorway case study simulated using MITSIMLab, a complex microscopic traffic simulator. The method allowed the analysis of the role played by all parameters and by the model stochasticity itself, with 80% fewer model evaluations than the standard variance-based approach. Ten model parameters accounted for a big share in the output variance for the specific case study. A Kriging metamodel was then estimated and integrated with the multistep SA results for a global calibration framework in the presence of uncertainty. Results confirm the great potential of this approach and open up to a novel view for the calibration of a traffic simulation model.

SimMobility Short-Term: An Integrated Microscopic Mobility Simulator

This paper presents the development of an integrated microscopic mobility simulator, SimMobility Short-Term (ST). The simulator is integrated because its models, inputs and outputs, simulated components, and code base are integrated within a multiscale agent- and activity-based simulation platform capable of simulating different spatiotemporal resolutions and accounting for different levels of travelers’ decision making. The simulator is microscopic because both the demand (agents and its trips) and the supply (trip realization and movements on the network) are microscopic (i.e., modeled individually). Finally, the simulator has mobility because it copes with the multimodal nature of urban networks and the need for the flexible simulation of innovative transportation services, such as on-demand and smart mobility solutions. This paper follows previous publications that describe SimMobility’s overall framework and models. SimMobility is an open-source, multiscale platform that considers land use, transportation, and mobility-sensitive behavioral models. SimMobility ST aims at simulating the high-resolution movement of agents (traffic, transit, pedestrians, and goods) and the operation of different mobility services and control and information systems. This paper presents the SimMobility ST modeling framework and system architecture and reports on its successful calibration for Singapore and its use in several scenarios of innovative mobility applications. The paper also shows how detailed performance measures from SimMobility ST can be integrated with a daily activity and mobility patterns simulator. Such integration is crucial to model accurately the effect of different technologies and service operations at the urban level, as the identity and preferences of simulated agents are maintained across temporal decision scales, ensuring the consistency and accuracy of simulated accessibility and performance measures of each scenario.

Autonomous mobility on demand in SimMobility: Case study of the central business district in Singapore

Autonomous mobility on demand (AMOD) has emerged as a promising solution for urban transportation. Compared to prevailing systems, AMOD promises sustainable, affordable personal mobility through the use of self-driving shared vehicles. Our ongoing research seeks to design AMOD systems that maximize the demand level that can be satisfactorily served with a reasonable fleet size. In this paper, we introduce an extension for SimMobility - a high-fidelity agent-based simulation platform - for simulating and evaluating models for AMOD systems. As a demonstration case study, we use this extension to explore the effect of different fleet sizes and stations locations for a station-based model (where cars self-return to stations) and a free-floating model (where cars self-park anywhere). Simulation results for evening peak hours in the Singapore Central Business District show that the free-floating model performed better than the station-based model with a “small number” of stations; this occurred primarily because return legs comprised “empty” trips that did not serve customers but contributed to road congestion. These results suggest that making use of distributed parking facilities to prevent congestion can improve the overall performance of an AMOD system during peak periods.

Noise Versus Outliers

In this chapter, the reader will learn about methods for identifying outliers in a dataset, and how different methods can be compared.

Automated Mobility-on-Demand vs. Mass Transit: A Multi-Modal Activity-Driven Agent-Based Simulation Approach:

Among the new transportation services made possible by the introduction of automated vehicles, automated mobility-on-demand (AMoD) has attracted a lot of attention from both industry and researchers. AMoD provides a service similar to taxi or ride-sharing services, while being driverless. It is expected to attract a huge fraction of travelers currently using mass transit or private vehicles and will have a disruptive effect on urban transportation. While most studies have focused on the operational efficiency of the technology itself, our work aims to investigate its impact on urban mobility. Our contribution is two-fold. First, we present a flexible AMoD modeling and simulation framework developed within a multi-modal agent-based urban simulation platform (SimMobility). The framework allows the detailed simulation and assessment of different AMoD operations together with an activity-based framework that accounts for changes in demand, such as activity participation, trip making, mode, destination, or route choice decisions. Second, we focus our attention on the role of mass transit in a futuristic urban system where AMoD is widely available. Mass transit is already challenged by current ride-sharing services, for example, Uber and Lyft, which provide comparatively better and cheaper services. This trend will plausibly be exacerbated with the introduction of AMoD, which may indirectly act as a replacement to mass transit. Our simulation results show that mass transit is irreplaceable, despite the high efficiency of AMoD, in order to avoid congestion and maintain a sustainable urban transportation system with acceptable levels of service.