Tutorials & Workshops
WS1: The First International Workshop on Intelligent Public Transports – Toward the Next Generation of Urban Mobility WS01 programme
Organizers: Luis Moreira-Matias, luis.matias at neclab.eu (Neclab EU, DE), Oded Cats (TU Delft, NL), Marc Barthelemy (CEA, FR)
Content: Nowadays, everything is being built up taking advantage in sensor’s data (e.g.: bridges, computers, houses, vehicles). The Public Transport is not an exception. By being highly dependent on the dynamics of the human behavior (both drivers and passengers), it is intrinsically connected to the data derived from them as well. In the past, this was completely unrealistic. The Data Miners worked closed on their labs with their impractical Machine Learning algorithms - as there was no large-scale data to apply them. On the other hand, the Civil Engineers aimed to model such dynamics assuming theoretical levels of stochasticity and/or optimistic scenarios. Such models comprise a fair but still inaccurate approach to such dynamic behavior. Today’s reality increased dramatically the availability of the mobility-based data by the multiple social infrastructures that already use intelligent sensors and real-time communicational frameworks (e.g. 3G).
The availability of these type of data (e.g. smartphones, traffic light sensors, APC/AVL, fare-based, etc.) on a large- scale changed the way that both Civil and Computer Scientists faced the problematics around Public Transportation. It enables a whole new bunch of possibilities which are still far by being fully explored. On the other hand, it also brings novel issues regarding each individual’s and/or company’s privacy that are worthy to be discussed and analyzed. Where are we going? Where do we want to go? Which are the current trends? How can we explore these data to improve the Public Transportation? Which can be done to improve the bus transfers coordination? How about the taxi dispatching? Preventive Maintenance? The Planning and The Control of Public Transportation in general?
These problematics are addressed by this workshop’s scope. The researchers/engineers are encouraged to participate and take advantage of this opportunity to exchange ideas and to share their R&D findings/experiences.
Organizers: S. Ilgin Guler, iguler at engr.psu.edu (PSU, US), Monica Menendez (ETHZ, CH)
Content: Modal variations in traffic operations could be one of the most viable solutions to traffic congestion problems in urban areas. By considering alternative modes, the use of personal vehicles as a means of transportation can be reduced, freeing up roadway space. However, often these alternative modes are thought of as inferior to personal vehicles due to their perceived low speeds and reliability levels. Both of these issues can be addressed by implementing intelligent transportation systems with a focus on improving multimodal transportation operations. The goal of this workshop is therefore to discuss the latest developments in intelligent transportation systems used to monitor, model, and/or control multimodal traffic operations.
Some of the alternative modes (e.g., buses), can often provide rather detailed and close to complete information on their location and schedule. This data could be used to monitor and model the operations of these modes, or even control them when combined with intelligent transportation systems. Moreover, this data could also be used to monitor and model the interactions between the alternative modes and personal vehicles, leading to control decisions which could improve the system holistically. This could lead, in general, not only to a more efficient service and more attractive alternative modes, but also a more sustainable transportation system overall.
List of topics:
- Transit signal priority as well as other signal control algorithms that account for multiple modes (with a special focus on public transport);
- Dynamic transit lanes and other mode-specific lanes;
- Shared spaces with multiple interacting modes;
- Car2car communications across different types of modes; and
- Any other algorithms to monitor, model or control the interactions between multiple modes on urban or interurban networks.
Organizer: Aleksandar Stevanovic, astevano at fau.edu (FAU, US)
Content: Urban traffic control is one of the most important factors that affect traffic congestion and its externalities (number of accidents, air pollution, noise, etc.) on urban roads. Traffic signal systems, as the core piece of any urban traffic control are going through a transitional phase. Caught between the crossroads of traditional practices and conventional equipment and new technologies and paradigms, field of the traffic signal systems is experiencing a mixture of ideas and intelligent solutions. While the need for recently emphasized performance monitoring is driving the process of retrofitting existing signal systems, new wireless-communicationbased technologies are emerging. This is all happening in spite of the fact that the ‘next generation’ of traffic signal systems (a.k.a. adaptive traffic control) has not been utilized to its full potential yet. In such a mixture of operational conditions, Intelligent Transportation Systems researchers and professionals (who design, operate and maintain traffic signal systems) are faced with multitude of challenges, including those on methodological, technological, and institutional sides. This workshop is intended to address such contemporary issues that researchers, industry professionals, and institutional users face in their signal-related activities and decision-making processes. The workshop is envisioned as a forum of ideas coming from a mixture of professionals from academia, industry, and government. Invited speakers will present on a variety of up-to-date research topics and innovative field developments, with orientation towards emerging technologies and practices. Then, the floor will be opened to audience to emerge into discussion with the speakers.
List of Topics: Adaptive traffic control systems; Traffic signals in connected vehicle environment; Performance-measurement-based traffic signals; Traffic signal systems and operations in multimodal environment; Innovative traffic signal control algorithms and methods; Traffic control as a congestion mitigation measure; Freeway ramp metering and control; Traffic signal control for oversaturated operations; Modeling and simulation exercises with advanced traffic signal systems; Incident-responsive traffic signal control; Traffic-responsive signal control; Traffic signals in environment of Big Data and Cloud Computing; Learning reinforcement and other soft-computing methods for urban traffic control
WS4: Advancing the Microscopic Traffic Simulations Towards Realistic Modelling of Driver Behaviour and Reliable Evaluation of Safety WS04 programme
Organizers: Haneen Farah, H.Farah at tudelft.nl (TU Delft, NL), Hans van Lint (TU Delft, NL)
Content: Microscopic Traffic Simulations Towards Realistic Modelling of Driver Behaviour and Reliable Evaluation of Microscopic traffic simulations have proven to be a useful tool for evaluating different traffic systems and ITS applications and are widely used in the transportation community. However, the underlying mathematical models in these traffic simulation tools for car-following, lane-changing and gap-acceptance are still far from fully replicating human driving behaviour. Although many of the assumptions in these models make intuitively sense, many of these are not yet sufficiently supported by evidence (microscopic traffic data) under controlled conditions. The result is that most microscopic simulation models are able to reproduce observable macroscopic phenomena in traffic flows (breakdown, the capacity drop, wave propagation, lane distributions, capacity distributions), but still have limited validity in reproducing the actual individual driving behaviour, particularly related to lateral movement, gap acceptance and path-planning.
The consequence of all this is that most microsimulation models on the market (and in research labs) cannot be used to predict the microscopic effects of for example different geometric designs, drivers’ fallacies, or specific control measures. Since the underlying behavioural models are by design collision-free (safe a few dedicated research models), predicting safety effects under all sorts of circumstances is beyond the scope of these models yet. As such, derivation of surrogate safety measures from current microscopic simulation models might lead for inaccurate or wrong conclusions regarding the safety of an entity. These research gaps are becoming critical with the rapid advancement of the intelligent and automated vehicles which will add more complexity to the already complex system of the driver-road-vehicle system.This workshop will be dedicated to presenting empirical insights, experimental ideas, behavioural theories, and simulation models of drivers’ behaviour under thoseconditions in order to uncover new and innovative ideas and share knowledge and information that can benefit the transportation research community.
List of topics:
- Driver behaviour modelling
- Microscopic simulation models
- Surrogate safety measures
- Driving simulators studies
- Naturalistic data
- Experimental methodologies
- Intelligent transportation systems
- Automated vehicles
WS5: New Tendencies of Artificial Intelligence in Dynamic Modeling and Optimization of Traffic Network WS05 programme
Organizer: Mauro Dell’Orco, dellorco at poliba.it (TU Bari, IT)
Content: It is well-known that dynamic modeling and optimization of traffic networks involves assignment problems not easy to handle, due to the "combinatorial explosion" of alternative paths.In this workshop, we solicit novel contributions and breaking results on this problem, supported by methods of Artificial Intelligence (AI), like Expert Systems, Swarm Intelligence, Neural Network or Soft Computing methods. Anyway, contributions on all aspects of Dynamic Network Modeling and Optimization are welcome. We are sure that this workshop possesses a great potential for giving rise to robust simulations of traffic networks for real-time solutions.
List of topics:
- Nature-inspired algorithms for the solution of complex transportation problems
- Agent-based methods for simulation of transport networks
- Multivariate Logic in studies of Travelers' Behaviour
- Other emerging methods for Dynamic Network Modeling and Optimization
Organizer: Ke Han, k.han at imperial.ac.uk (ICL, UK)
Content: The traffic research community and practitioners are now faced with massive, multi-source, and heterogeneous datasets, which demand paradigm-shifting innovations in traffic theories, algorithms, and applications. These include, but are not limited to, formulation and calibration of fundamentally new models, traffic state estimation and prediction enhanced by statistical learning and data mining techniques, more robust and reliable decision making informed by historical and real-time traffic data, new travel behavior modeling facilitated by information and communication technologies, and new data management and visualization paradigms. This workshop will bring together expertise from multiple disciplines, and provide a platform to share and synergize cutting-edge research and practice.
List of topics:
- calibration and application of parsimonious traffic models
- robust traffic estimation and prediction with heterogenous data sources
- decision support systems based on machine/statistical learning techniques
- dynamic traffic assignment modeling enabled by real-time data
- visualization of big data in traffic systems
- optimization of transportation system driven by big data
- data-driven robust optimization
- interactive, adaptive, and smart infrastructure
WS7: Mapping the Landscape of Social Media/Big Data for Transportation Systems Analysis WS07 programme
Organizers: Constantinos Antoniou (NTUA, GR), Francisco Pereira, camara at mit.edu (MIT, US), Loukas Dimitriou (Univ. of Cyprus, CY)
Content: Data collection for transportation applications has moved in the last decades from a limited pool of traffic surveillance technologies and travel surveys to a wealth of high volume heterogeneous, opportunistic data. Being a particularly rich dataset about human behavior, the Internet has brought us several new angles to analyse. It comprises social networks, institutional, private and personal websites, crowdsourced content, and multi-platform mobility services (e.g. waze). A large number of researchers with different backgrounds have used such data to approach transportation application problems from different perspectives, resulting in a large number of potential solutions to the same problems. The objective of this workshop is to contribute towards the organization and structured mapping of this landscape, in an effort to consolidate existing data and techniques and suggest the most promising future avenue to pursue. Researchers from different backgrounds will present their views and there will be ample time for the exchange of views, in response to the presented material.
List of topics:
- Social network information for inference of travel behavior
- Mobility pattern recognition by Social/Public data
- Mining internet for contextual data about travel behavior
- Big data analysis for anomaly/extreme events detection and prediction
- Internet data analysis for anomaly/extreme events detection and prediction
- Using social networks to map urban structures
Organizer: Jeffrey Miller, jeffrey.miller at usc.edu (USC, US)
Content: With the availability of driverless vehicles being around 5-10 years from now (if not sooner), people need to be educated on how to operate them and the technologies that are involved. Nearly everyone will be affected by this drastic change in transportation, so everyone will need to be educated in the effects of ITS technology. This includes researchers, the public, kids, other drivers, pedestrians, bicyclists, and a number of other groups. This workshop will focus on defining the who, what, where, when, why, and how that will go into educating people about the impact of ITS and driverless vehicles in the world.
List of topics:
- Education in ITS
- Education in driverless vehicles
- Training in ITS technologies
- User acceptance strategies for ITS
- User acceptance strategies for driverless vehicles
Organizers: Fei-Yue Wang (CASIA, PRC), Wei Chen (ZJU, PRC), Pu Wang (CSU, PRC), Xiao Wang (CASIA, PRC), Xinhu Zheng, zheng473 at umn.edu (UMN, US)
Content: With the fast developement and application of sensing, computing, and networking technologies, social media, wearable and mobile devices have produced huge volumes of real-time human-generated real-time signals for social transportation, an emerging field of transportation research and applications. These signals, from drivers’ GPS coordinates to mobile phone billing records and messages from social networking websites, record our daily spatio-temporal information and form huge data sources for traffic and transportation analytics. Transportation systems, as one of the most complex man-made systems, face a number of complicated and cost/time/space-critical tasks. The goal of this session/workshop is to encourage people to solve or address these tasks by introducing crowdsourcing mechanisms and social transportation into ITS, which can make better use of the pervasive real-time social signals and distribute the cumbersome and time consuming tasks among the crowds. The main idea is to apply the location-based services and to use the real-time user feedback social signals to conduct traffic planning, assignment, navigation, localization and transportation analytics with lower cost, higher accuracy, faster speed, more agility and full response. We hope this special session/workshop can attract transportation researchers and practitioners to join us along this promising new direction in ITS.
Specifically, this workshop will focus on the following areas: 1) traffic or transportation analytics with big data and social signals using data mining, machine learning, and natural language processing methods; 2) crowdsourcing mechanisms for transportation based social media, social networking, and the Internet of Things (IoT) or even the Internet of Everything, especially the coming V2X, that is, vehicles to vehicles, websites, people, infrastructures communication; 3) new services beyond location-based services (LBS), such as transportation knowledge automation, especially decision-based services (DBS) or task-based services (TBS) that collecting required information in real time for transportation decisions or tasks, and information or intelligence-based services (IBS) or knowledge-based services (KBS) that recommending agents or organizations who might find the identified intelligence or knowledge useful for solving their traffic problems or improving the transportation performance; 4) web-based agent technology for transportation control and management, such as software robots, knowledge robots or web surrogates for traffic monitoring, safe driving, vehicular heath and energy management, at this point, the effort should be directed in developing various smart apps that collect social traffic data and link people to traffic and cars in real time; and 5) real applications and feedback for more research and development.
Organizers: Xiangmo Zhao, xmzhao at chd.edu.cn (CHD, PRC), Fei Hui (CHD, PRC), Yinhai Wang (UW, US), Alois Knoll (TUM, DE), Feihu Zhang (Fortiss GmbH, DE)
Content: The combination of data from internal sensors on-vehicle and external sensors from infrastructure and other traffic participators plays an important role in modern transportation domain. An intelligent analysis of this data can help to optimize the utilization of the transportation efficiency and mitigating the impacts of traffic congestions. With the development of Vehicle-2-Infrastructure (V2I) and Vehicle-2-Vehicle (V2V) techniques, sharing information from various devices, infrastructures, services and participants becomes possible across the whole network. These new heterogeneous communication networks can be utilized to improve the situational awareness. Achieving the mentioned goals requires combining computer science and engineering techniques for further developments in transportation domain. This workshop aims to provide a forum for researchers to present work which focused on the algorithms, applications and the developments of the Connected Vehicles and Cooperative Systems.
Organizers: Samy El-Tawab, eltawass at jmu.edu (JMU, US), Mona Rizvi (NU, KZ) -->link to website
Content: The goal of this workshop is to discuss the issues concerning building applications on Vehicular Networks. Due to the recent decision by the U.S. National Highway Traffic Safety Administration (NHTSA) to move forward with vehicle-to-vehicle (V2V) communications technology for light vehicles, vehicular ad-hoc networks (VANET) applications will receive more attention. Further, the new communications capability to be required in vehicles will provide the potential for significantly more smart vehicles application development.
With the nearly ubiquitous use of smart phones by consumers, most car manufacturers already provide smart phone apps that integrate with their in-vehicle navigation, information and entertainment systems. As vehicles become progressively smarter and the era of self-driving vehicles nears, further integration between consumers' personal devices and vehicle systems will be required. Research done in the field of Vehicular Networks and Intelligent Transportation has covered several types of applications including incident detection, weather alerts, traffic conditions, safety applications, and more.
This workshop will bring together researchers on VANETs to present and discuss recent advances in the development of applications and related technologies. The workshop will cover all theoretical and practical aspects related to applications, modeling, simulation and testbeds.
Organizers: Joaquim Ferreira, jjcf at ua.pt (UA, PT), Matteo Petracca (CNIT, IT), Elena Cordiviola (INTECS, IT) -->link to website
Content: Transportation systems have received widespread attention from scientific community and emerged towards Intelligent Transportation Systems, where there is closed loop interaction between vehicles/drivers and the transportation infrastructure empowered by cooperative V2X communications, cellular networks and wireless sensor networks. Cooperative applications with data sensing, acquisition, processing and communication provide an unprecedented potential to improve vehicle and road safety, passenger's comfort and efficiency of traffic management. In order to support such visionary scenarios, applications running in the vehicles are required to communicate with other applications in the vehicle or with applications deployed in the back office of the emergency services, road operators or public services. These applications run unattended, reporting information and taking commands from counterpart applications in the vehicle or network, and are therefore referred to as Machine-to-Machine applications (M2M). Current ETSI M2M standards do not fully address scenarios in which mobility of entities are involved; such as the case of vehicular network nodes roaming through a network of various RSUs connected to M2M gateways. In this networks, devices operating inside vehicles are called On Board Units (OBUs), while devices operating on the side of the road are Road Side Units (RSUs), and have different requirements and modes of operation. A cooperative system for smart mobility is a pervasive system based on wireless sensor networks and vehicular networks composed by control centres, RSUs, OBUs and personal handled devices.
This workshop will bring together academic and industrial stakeholders to identify and discuss technical challenges, and be a venue for disseminating innovative solutions in the field of cooperative sensing for smart mobility applications. These solutions will encompass all aspects of ITS associated with cooperative sensing: from wireless network infrastructure and mobility, to physical layer transmission techniques, middleware, M2M architectures, data models and real life implementations.
Organizers: Hilmi Berk Celikoglu (ITU, TR), Javier Sanchez Medina, javier.sanchez.medina at ieee.org (CICEI, ES),
Content: The current technological progressions on electronics and computer science possess a great potential to extend the application of computing methodologies in research and industry. The increasing power of computers has advanced the modeling, simulation, and optimization of complex systems such as dynamic transportation networks. This gave rise not only to the incorporation of various existing theories and methods to network problems, but also to the robust simulation of interactive user-network behavior for real time solutions. Considering the need for the efficient modeling and simulation of vehicular network traffic within temporal domain with reasonable computation load, this workshop solicits novel contributions and breaking results on all aspects of high performance computing applications in network traffic simulation.
Workshop aims to setup a platform for a multidisciplinary community that’s composed of researchers, educators, and practitioners. Workshop is open to contributions expressing and discussing views on trends, challenges, and state-of-the art in parallel computing applied to traffic network problems.
Organizers: Cristobal Curio, cristobal.curio at reutlingen-university.de (Reutlingen Univ., DE), Christoph Stiller (KIT, DE)
Content: The effective interaction with other traffic participants is an open challenge for automated vehicles. This is particularly true for urban environments that are not primarily dedicated to traffic and are populated with vulnerable road users like pedestrians and bicyclists. In order to cope with the wide variations in traffic situations and behaviour of traffic participants scientific progress is required in perception, prediction and interaction techniques.
Current driver assistance systems support the recognition of pedestrians in dynamic scenes, typically based on the visual detection of static and dynamic cues of individual persons. Based on this information the behaviour of pedestrians is coarsely predicted and a suitable driver assistance function is triggered. Social interaction between automated vehicles and humans is generally not explicitly addressed.
The focus of this workshop is to discuss the state-of-the-art and emerging methods in understanding, prediction and planning of urban traffic situations, social actions and interaction. The workshop will gather experts from the disciplines relevant to this field from academia and industry. The long term vision is a "social car" that judges and communicates complex interactive situations between a vehicle and other traffic participants and establishes a dyadic exchange of socially relevant information for seamless cooperative automated driving or comprehensive driver assistance in populated areas.
List of topics:
- Behaviour Models for Cars, Bicycles and Pedestrians
- Social Interaction Models
- Long Term Behavior Prediction
- Behavior Change Detection
- Intention Modelling
- Cooperative Motion Planning
- Pedestrian Detection and Tracking
- Motion Capture
- Perception of Crowds
- Building a simulation network from publicly available data sources;
- Specifying multi-modal traffic demand;
- Visualizing the Simulation;
- Exporting traffic data for further analysis; and
- Advanced interaction with the simulation.
Prof. Peter Wagner and Dr. Jakob Erdmann (German Aerospace Center, DLR)
The number of participants to this Tutorial will be limited. We will enable a free registration system. Stay tuned.
- Important concepts – time, location, and system partitioning,
- The Unified Implementation of the Connected Vehicle Reference Implementation Architecture,
- Tools and implementation support activities, and
- Progress on Connected Vehicle Pilots that make use of the architecture.
Walton Fehr (ITS Joint Program Office, US Department of Transportation)
The number of participants to this Tutorial will be limited. We will enable a free registration system. Stay tuned.
We will be publishing here more information about Tutorials & Workshops soon.