Advances in Artificial Transportation Systems and Simulation

245 Pages·2014·12.826 MB·English

by  Rosaldo J F Rosetti, Ronghui Liu

Advances in Artificial Transportation
Systems and Simulation
Edited by
Rosaldo J.F. Rossetti
Ronghui Liu
AMSTERDAM • BOSTON • HEIDELBERG • LONDON
NEW YORK • OXFORD • PARIS • SAN DIEGO
SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO
Academic Press is an imprint of Elsevier
Academic Press is an imprint of Elsevier
525 B Street, Suite 1800, San Diego, CA 92101-4495, USA
225 Wyman Street, Waltham, MA 02451, USA
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK
Copyright © 2015 Zhejiang University Press Co., Ltd. Published by Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or
mechanical, including photocopying, recording, or any information storage and retrieval system, without
permission in writing from the publisher. Details on how to seek permission, further information about
the Publisher’s permissions policies and our arrangements with organizations such as the Copyright
Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/
permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher
(other than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical treatment
may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and
using any information, methods, compounds, or experiments described herein. In using such informa-
tion or methods they should be mindful of their own safety and the safety of others, including parties for
whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any
liability for any injury and/or damage to persons or property as a matter of products liability, negligence
or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in
the material herein.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress
ISBN: 978-0-12-397041-1
For information on all Academic Press publications
visit our website at http://store.elsevier.com/
Typeset by Thomson Digital
Printed and bound in the United States
List of Contributors
Fábio Aguiar MIEIC, DEI, Faculdade de Engenharia da Universidade do Porto, Porto, Portugal
João Emílio Almeida LIACC, DEI, Faculdade de Engenharia da Universidade do Porto,
Porto, Portugal
Jean-Paul A. Barthès Université de Technologie de Compiègne, UMR CNRS Heudiasyc, France
Ana L.C. Bazzan Instituto de Informatica, UFRGS, Porto Alegre, RS, Brazil
Philippe Bonnifait Université de Technologie de Compiègne, UMR CNRS Heudiasyc, France
Sara Carvalho MIEIC, SAPO Labs, Faculdade de Engenharia da Universidade do Porto,
Porto, Portugal
António J.M. Castro LIACC, Faculdade de Engenharia da Universidade do Porto, Porto, Portugal
Paul Davidsson Department of Computer Science, Malmö University, Malmö, Sweden
Manoel T. de Abreu Netto Department of Computer Science, PUC-Rio, Rio de Janeiro, RJ, Brazil
Maicon de Brito do Amarante Instituto Federal Farroupilha, São Vicente do Sul, RS, Brazil
Carlos J.P. de Lucena Department of Computer Science, PUC-Rio, Rio de Janeiro, RJ, Brazil
Baldoino F. dos Santos Neto Department of Computer Science, PUC-Rio, Rio de Janeiro, RJ, Brazil
João Filguieras Instituto de Engenharia de Sistemas e Computadores, INESC-ID, Lisbon, Portugal
Joaquim Gabriel IDMEC, DEMec, Faculdade de Engenharia da Universidade do Porto,
Porto, Portugal
John Graham Complex Systems and Non-linear Dynamics Group, Universidad Autónoma de la
Ciudad de México, San Lorenzo, Del Valle, México, D.F. México
Shaza Hanif Department of Computer Science, KU Leuven
Milton Heinen Instituto de Informatica, UFRGS, Porto Alegre, RS, Brazil
Sergio Hernandez Postgraduation Program in Complex Systems and Non-linear dynamics,
Universidad Autónoma de la Ciudad de México, San Lorenzo, Del Valle, México, D.F. México
xi
List of Contributors
Johan Holmgren Faculty of Computing, Blekinge Institute of Technology, Karlshamn, Sweden,
Department of Computer Science, Malmö University, Malmö, Sweden
Tom Holvoet Department of Computer Science, KU Leuven
Zafeiris Kokkinogenis LIACC, DEI, Faculdade de Engenharia da Universidade do Porto,
Porto, Portugal
Zhengjiang Li State Key Laboratory of Management and Control for Complex Systems, Institute of
Automation, Chinese Academy of Sciences, Beijing, China
Nuno Machado MIEIC, DEI, Faculdade de Engenharia da Universidade do Porto, Porto, Portugal
Peter T. Martin Department of Civil Engineering, New Mexico State University, Las Cruces, New
Mexico, USA
Antonio Neme Complex Systems and Non-linear Dynamics Group, Universidad Autónoma de la
Ciudad de México, San Lorenzo, Del Valle, México, D.F. México
Omar Neme School of Economics, Instituto Politécnico Nacional, México, D.F. México
Eugénio Oliveira LIACC, DEI, Faculdade de Engenharia da Universidade do Porto, Porto, Portugal
Lúcio Sanchez Passos LIACC, DEI, Faculdade de Engenharia da Universidade do Porto,
Porto, Portugal
Linda Ramstedt Sweco, Stockholm, Sweden
Rosaldo J.F. Rossetti LIACC, DEI, Faculdade de Engenharia da Universidade do Porto,
Porto, Portugal
Luís Sarmento LIACC, SAPO Labs, Faculdade de Engenharia da Universidade do Porto,
Porto, Portugal
Ivana Tasic Department of Civil and Environmental Engineering, University of Utah, Salt Lake City,
Utah, USA
Fenghua Zhu State Key Laboratory of Management and Control for Complex Systems, Institute of
Automation, Chinese Academy of Sciences, Beijing, China
Milan Zlatkovic Department of Civil and Environmental Engineering, University of Utah, Salt Lake
City, Utah, USA
xii
Preface
Intelligent Transportation Systems (ITS) have evolved enormously within the last three
decades. More recently the rapid growth in new technologies has allowed for the practical
implementation of more interactive and pervasive ITS solutions. Because of the important
role the transportation systems play in society and economy, industry has engaged actively as
a great promoter of ITS’ technological development and governments all over the world are
prioritizing mobility as a key ingredient of their social and economic growth. Increasingly,
ITS developments become not solely technological, but are across a wide span of different
disciplines.
Despite of the exponential growth in computer power and communication technologies
underlying ITS, meeting the future challenges in transportation will require focusing on
social and environmental aspects where user preferences are a central concern. Human
interactions with ITS solutions have now gained a new meaning. As technology is able to
behave more intelligently than before, services become peers of users: they perceive, make
decisions and reason about the results of their actions, all the while seeking to benefit all
parties. In fact, rather than increasing service capacity, one underlying approach of ITS-based
solutions nowadays is to ensure productivity and mobility by making better use of existing
infrastructure and services, furnishing them with smarter, greener, safer, and more efficient
solutions.
The complexity of contemporary ITS technology and its wider social and environmental
impacts demand new modeling paradigms that incorporate cooperation and collaboration
among intervening parties, and support the design and practical deployment of future
mobility solutions. In response to the need to understand the interplay between technology
and social interactions, Prof. Fei-Yue Wang proposed the concept of Artificial Transportation
Systems (ATS), just over a decade ago, during the 2003 IEEE International Conference
on Intelligent Transportation Systems. With the ability to integrate different transportation
models and solutions in a virtual environment, ATS are an extension to traditional modeling
and simulation methodologies that deal with transportation issues from the complex systems
perspective and in a systematic and synthetic way. They provide a natural platform where
new approaches can be experimented while avoiding natural drawbacks of dealing directly
with real-life critical domains. Building on the theories and metaphors developed in a wide
xiii
Preface
spectrum of disciplines, spanning from social sciences, artificial intelligence, and multi-agent
systems, to distributed computing and virtual reality, many important issues arise in ATS that
challenge and motivate researchers and practitioners from multidisciplinary technical and
scientific backgrounds.
Inspired by the concept, IEEE ITS Society soon after created and has since hosted the
Technical Activities Sub-committee on Artificial Transportation Systems and Simulation
(ATSS) with the mission of motivating and promoting research and practice in ATSS. As
part of this effort, a series of successful biennial ATSS Workshops has been organized by
this committee and integrated in scientific programs of IEEE ITSC, the flagship conference
series of the society. In addition, every other year, alternating with ATSS workshops, a series
of special sessions have been organized as part of ITSC to consolidate ideas and trends
discussed in previous ATSS workshops. This book is a major outcome of the ATSS sub-
committee, bringing to the readers a collection of selected papers presented during the ATSS
Workshop held in ITSC 2010, in Madeira, Portugal, and during the ATSS Special Session
held in ITSC 2011, in Washington-DC, USA. The papers herein included report on important
aspects and technological advances underlying the concept of ATS and establishing the basis
for the implementation of appropriate simulation methodologies and tools.
Each chapter in this book consists of an extended and updated version of papers previously
presented in the above two ATSS events. A total of 12 papers have been selected that
cover different aspects of ATS and span across topics such as simulation tools, modeling
methodologies, practical applications, alternative data sources, crowd sensing, and
participatory simulation. Starting with tools, Bazzan et al. present ITSUMO, an open-source
microscopic traffic simulator whose implementation relies on the agent metaphor. Differently
from other similar tools, ITSUMO integrates both demand and control perspectives. It
follows a bottom-up behavioral approach, offering sufficient flexibility for the development
of different algorithms and techniques to test with route assignment and re-planning, driver
behavior, and traffic control strategies and coordination.
Netto et al. discuss issues related to the representation of complex domains, and propose a
framework as a reusable solution for building decentralized self-organizing systems, based on
major architectural patterns found in the literature. Considering a higher level of abstraction,
their approach provides extensibility features to develop new interaction and coordination
mechanisms between agents and the environment, which is demonstrated in an automated
guided vehicles scenario. Passos et al. also tackle complex system analysis from a multi-
agent perspective. In their work, authors analyze the adequacy of traditional approaches in
the field of Agent-Oriented Software Engineering to create adequate multi-agent systems
in the specific domain of transportation. They devise a novel methodology where the concept
of services is considered as peer of agents, ambience, and processes, and becomes prominent
elements in the modeling phase. The approach is illustrated in a typical transportation domain
xiv
Preface
scenario. Holmgren et al. explore modeling specificities in the supply chain domain. The
authors propose a method based on a framework of supply chain roles, responsibilities,
and interactions, which can be used to represent different types of organizations involved
in providing and using products and transport services. Their method is illustrated through
five different supply chain simulation models, which are analyzed to demonstrate validity
and generality of their approach. Also talking about modeling issues, Hanif and Holvoet
illustrate how design patterns can support the design of complex agent-based solutions to
Pickup and Delivery Problems. In particular, authors use the so-called delegate multi-agent-
system patterns to build agent interaction behaviors, and show through simulation that a
structured and reusable pattern can significantly reduce the system design and implementation
complexity, yet achieve interesting quality characteristics.
Reporting on the applications, Machado et al. demonstrate how the ATS concept can be
used in practice as means to assess and evolve the organizational structure performance of
airline companies. The authors build on the empirical knowledge gained through interviews
with airline operators and develop an analytical framework so as to evaluate current as
well as hypothetical organizational structures. To illustrate their approach, real pre- and
postoperational data is used to support the simulation of different operation scenarios
allowing for different metrics to be analyzed. Neme et al. address issues in modeling
pedestrian dynamics and study passengers inside high-capacity buses. Through ATS they
observe uneven density distributions leading to high discomfort to the passengers. An agent-
based model was devised to represent the interactions between passengers and the bus
interior, comprising seats, aisle, and access doors. The authors analyze different schemas
and policies so as to gain insight into how to improve passenger comfort on buses. In a
different perspective, Almeida et al. study pedestrian dynamics in relation to the behavior of
crowds in emergency situations, and present ModP, an agent-based pedestrian simulator. The
tool is flexible enough to allow different behaviors to be modeled through a simple syntax,
providing designers with a productive environment to rapidly prototype and test with different
scenarios. The authors carried out preliminary studies to illustrate usability of their tool.
The implementation of cognitive capabilities in intelligent vehicles interfaces is addressed
by Barthès and Bonnifait as a means to allow vehicles to interact collaboratively with their
drivers in operating conditions. The interactive interface is built on top of a multi-agent
system and tested in an Advanced Driving Assistance System scenario providing speed
warnings whenever dangerous areas are approached. The authors test their approach in a
real-world scenario resorting to an experimental vehicle. Zhu and Li look into new methods
to address transportation problems from new perspectives using the Artificial Societies,
Computational Experiments and Parallel Execution approach, built upon the ATS premises.
The authors emphasize on synthesizing artificial societies and modeling environmental
impacts, whereas their architectural approach relies on a cloud computing infrastructure.
They illustrate the concept with an ATS case study and present preliminary results of their
xv
Preface
methodology. In addition, Zlatkovic et al. illustrate one important premise of ATS, namely its
ability to allow for software- and hardware-in-the-loop simulations. Their work present and
discuss an implementation of software-in-the-loop (SIL) simulation of the Advanced System
Controller series 3 (ASC/3) in transit signal priority scenarios. The authors test two options of
ASC/3 using a VISSIM simulation model of a bus rapid transit solution in West Valley City,
Utah. Results are encouraging and demonstrate how SIL simulation can offer many options
for testing custom-defined traffic control strategies.
Alternative data sources, crowd-sensing, and participatory simulation are also topics of
major concern in ATS. Kokkinogenis et al. discuss a new type of mobility studies resorting
to the growth in popularity of opinion mining in social media, and social media itself. Users
are considered to be sensors of the mobility dynamics, capable of providing insight into the
flaws of mobility networks, user preferences, and other unexplored sorts of information.
Their approach builds upon sensing real-time traffic-related information using microblogging
messages posted on Twitter (by users in transit), to which a text classification approach
is proposed. This work opens up vast swathes of opportunities to explore alternative data
sources leveraged on mass participation and collaboration.
We are very pleased with this interesting and motivating collection of papers reporting on
the basic concepts, state-of-the-art developments, and a myriad of potential applications and
future trends in Artificial Transportation Systems and Simulation. We hope this book will
serve as a source of inspiration and insight for many researchers, practitioners, and educators,
and leverage further progresses in the field of ATSS. We conclude with a word of appreciation
to all authors who have submitted their contributions to ATSS events (in both workshops
and special sessions) throughout the past 10 years, with clear visions, novel research, and
significant results. Finally, the contributions included in this book are results of laborious
and time-consuming work of many reviewers who have helped us with their expertise,
suggestions, and recommendations, and to whom we are greatly indebted. Thank you!
October 2014
Rosaldo J.F. Rossetti
Laboratório de Inteligência Artificial e Ciência de Computadores,
Departamento de Engenharia Informática,
Faculdade de Engenharia da Universidade do Porto,
Rua Dr. Roberto Frias, S/N, Porto, Portugal
Ronghui Liu
Institute for Transport Studies, University of Leeds, 34–40 University Road,
Leeds, United Kingdom
xvi
CHAPTER 1
ITSUMO: An Agent-Based Simulator
for Intelligent Transportation Systems
Ana L.C. Bazzan*, Milton Heinen**, Maicon de Brito do Amarante†
*Instituto de Informatica, UFRGS, Porto Alegre, RS, Brazil; **Instituto de Informatica, UFRGS,
Porto Alegre, RS, Brazil; †Instituto Federal Farroupilha, São Vicente do Sul, RS, Brazil
1.1 Introduction and Motivation
The second half of the last century has seen the beginning of the phenomenon of traffic
congestion. This arose due to the fact that the demand for mobility in our society has
increased constantly. Traffic congestion is a phenomenon caused by too many vehicles trying
to use the same infrastructure at the same time. The consequences are well known: delays,
air pollution, decrease in speed, and dissatisfaction (which may lead to risk maneuver thus
reducing safety for pedestrians as well as for other drivers).
The increase in transportation demand can be met by providing additional capacity. However,
this might no longer be economically or socially attainable or feasible. Thus, the emphasis
has shifted to improving the existing infrastructure without increasing the overall nominal
capacity, by means of a better utilization of this capacity. Two complementary measures can
be taken. In traffic engineering terminology these are associated with management of the
demand (users, drivers) and supply (infrastructure, control). The set of all these measures is
framed as Intelligent Transportation Systems (ITS).
In the last years there have been some proposals for simulation platforms that are flexible
enough to test ITS techniques and approaches. Some (e.g., Paramics, AISUM, VISIM,
EMME2, Dracula) are based on classical models of simulation and are commercial tools.
With the appearance of a new simulation paradigm – agent-based simulation – it is now
possible that traffic experts and other users develop their own applications. This has been
achieved to some extent (e.g., Dresner and Stone, 2004; Rossetti and Liu, 2005; van
Katwijk et al., 2005; Balmer et al., 2008; Bazzan et al., 1999; Burmeister et al., 1997; Tumer
et al., 2008; Vasirani and Ossowski, 2009, 2011) but these tools are goal-directed meaning
that they were built for (more or less) specific purposes. One of the notable exceptions is
MATSim (www.matsim.org). However, MATSim’s simulation paradigm is queue based, traf-
fic lights are very simple, and drivers are not fully autonomous (e.g., during replanning). We
remark that besides the commercial simulators mentioned above, there is also the possibility
Advances in Artificial Transportation Systems and Simulation.
1
Copyright © 2015 Zhejiang University Press Co., Ltd. Published by Elsevier Inc. All rights reserved.
2 Chapter 1
to use SUMO (http://sumo.sourceforge.net/) as a starting point to investigate traffic scenarios
using microscopic traffic simulations. However, SUMO does not yet allow native tools for
implementing agent-based solutions.
In short, most of previously mentioned works have one or more of the following drawbacks:
they are not fully agent based; they rely on strong simplifying assumptions; they do not consid-
er both control and assignment of demand as a whole process (except in (Vasirani and Ossows-
ki, 2009, 2011), but here the integration only refers to their specific market-based approach).
Therefore, there is a lack of support for traffic experts who want to implement and test their
own solutions (e.g., artificial intelligence (AI)-based approaches for optimization or broadcast
of recommendation). These experts can neither extend commercial tools (except for some
API-based modules, which (1) are not totally flexible and (2) represent an additional purchase
cost), nor use the available free tools as they deal only with pieces of the whole problem.
This way there is still the need for an integrated platform that is as follows: fully based on
the autonomous agent paradigm of simulation; open source and user friendly; and considers the
effects of both control measures on driver’s reasoning and vice-versa. The present chapter
describes ITSUMO (Intelligent Transportation System for Urban Mobility), an open-source
tool that addresses these issues. It allows the modeling of traffic actors (drivers, traffic lights,
and even autonomous vehicles) as autonomous agents; it deals with short-term control of
traffic lights and with en-route replanning by drivers; thus it permits the study of coeffects
of both demand and supply. This is achieved by means of AI techniques in general and of
agent-based techniques in particular.
With the increased dissemination and computing power of mobile devices, it is now possible to
execute distributed AI applications for various situations: intelligent routing using algorithms
that do not rely on full knowledge; planning under constraints, and restricted communication
and information; distributed optimization of traffic lights. For instance, it is possible to define
drivers as intelligent agents and to plug each driver model. This approach is different from
current models, which are purely reactive and ignore drivers’ mental states (informational and
motivational data). Also, it is possible to plug reinforcement-learning-based control for traffic
lights.
An earlier version of ITSUMO was presented as a demo in the AAMAS conference (da Silva
et al., 2006b). However, although ITSUMO has also been used to investigate route choice
scenarios, the focus has been primarily on control. The current version was extended in the
sense that it now allows modeling of both control measures and drivers reaction to them, as
well as routing techniques. Moreover, this is provided as noncommercial code and is one of
the few tools that are truly agent based (thus microscopic). As shown in the next section, the
simulation kernel is responsible for handling the movement of vehicles. Other modules
support the agent-based modeling of demand and supply.