Car-2-Car Communication is one of the most promising technologies to make traffic in future more safe and efficient. The number of possible use cases is immense; the technology development is in progress. A lot of companies and projects in Europe, America and Asia are trying to filter out the most earning and/or efficient use cases. Beside the started or soon starting (big scale) field operational tests a lot of simulations are run to prove concept and effect of different applications with different tools and rather incomparable assumptions on environment, penetration rates, driver acceptance etc.
Since the main objective of the CAR 2 CAR Communication Consortium is to standardise and harmonise these vehicular communication systems and applications, it is a necessary step to coordinate the way towards these standards. In particular it is very important to succeed in finding a way to compare different use cases in matters of penetration rate needed, public benefit etc. Starting with a common understanding of how to simulate and evaluate results is the most promising first step.
Behind this first step is a simple thought: The system we are developing is hopefully getting better with every participant joining the network. And the same is true for our simulation efforts; better cooperation between the projects, better results.
The goals of the first C2C-CC Simulation workshop:
- promote cooperation of projects
- exchange and compare simulation results
- summarize the status quo
- address missing links and models
- show a roadmap for IPR sharing
This is not only an invitation to the workshop, but also a call for actively participating. Apart from panel discussions there is also the possibility for you to present thoughts and results in plenum or in a separated area as a booth.
09:30 - 09:45 Welcome
09:45 - 10:45 Session 1 : Applications and Goals
- introduction to C2C-CC Use cases
- goals of simulations
11.15 - 13:00 Session 2 : Models and tools
- discussion and result: list of models and tools
13:00 - 14:00 Lunch
14:00 - 15:15 Session 3 : Results and gaps
- status quo
- gap list
- discussion and result: list of results and list of gaps
15:30 - 16:00 Session 4 : Cooperation of projects
- Roadmap for sharing IPR
- Coordination of WPs = schedule of projects in matters of simulation
16:00 - 16:30 Wrap up
List of Presentations:
Starting with the ITS challenges Bart Driessen (TNO) highlighted "The role of high fidelity simulation for developing and validating new mobility". Simulation of cooperative systems includes modelling of vehicles, application, infrastructure, driver, communication channels sensors and actors. A tool-box enables multi aspect assessment up to hardware-in-the-loop simulation and testing with real vehicles. The question of how to use the nice tools started the following discussion.
Download the above presentation (1.203KB)
Thomas Fuhrmann (TU München) reported on "The BmBF AmbiComp Project - Self-Organising Decentralised Applications in Ad-Hoc Networks". Similar to the Internet AmbiComp proposes a new routing protocol for wireless peer to peer ad-hoc communication between mobile devices. Simulation results showed that the small routing algorithm is dedicated for very large networks with more than 100,000 nodes. He pointed out that simulation results are not generally comparable because of different assumptions and other aspects. To overcome the problem of comparability common scenarios and interfaces are needed for simulation benchmarks.
Download the above presentation (1.043KB)
In his talk : Verification of mobile applications by testing Nicolas Riviere (LAAS-CNRS) reported about the end-to-end verification of the highly dynamic mobile applications of HIDENETS using new testing methodologies. Testing of the first use case has been performed with strong hypotheses. The current platform showed to be not sufficient. Richer environment simulation and network simulator have to become integrated. An abstract language is planned for specifying the model based test generation in future.
Download the above presentation (238KB)
During his presentation "Requirements and challenges for the development of car-to-car channel models" Thomas Kürner (TU Braunschweig) highlighted the need for reference and worst case scenarios including traffic scenarios and environment properties (e.g. buildings, tunnels) for detailed analysis and development of realistic channel models for inter-vehicle and vehicle to infrastructure communication and their statistical characteristics. The analysis with ray tracing and ray launching methods is very time consuming. Its results will be validated with real test equipment based on IEEE 802.11.b at the moment.
Download the above presentation (1.958KB)
The concluding panel discussion pointed out a high need and interest on sharing a set of well defined same scenarios and use cases transparently described as well as modules and simulators. However depending on the research focus several (sub-)models with adopted detailing are needed. Comparable simulation results might be reached in future for a set of common use cases using an open shell and integrating the tools with adopted model detailing. The scenarios considered should include all kind of vehicles and road categories.TOP of Page
Christian Wewetzer (VW) pointed out the importance of simulation as cost-effective means of large-scale evaluation of applications regarding their effect on traffic. For this, Volkswagen has done simulations using a realistic model of the city of Braunschweig. Depending on the application under research three alternative tools couplings of VISIM and ns2 can be used from recorded movement patterns, to live simulated movement patterns up to the full feedback loop. Simulation scenarios with about 5,000 mobiles require very much computation time. Parallel computing might solve this problem. At the present only WLAN communication is modelled, but UMTS, EDGE, WiMax, etc. should be added for analysis of more general scenarios as well as further traffic environments in addition to Braunschweig. The fundamental knowledge is represented by the models while the simulator represents only the tool.
Download the above presentation (597KB)
Silja Assenmacher (TU München) reported about "Simulating impact of Car2X by VISIM & Vcom" for the 22 km motorway A5 between Friedberg and Frankfurt. The modelled driver behaviour has been calibrated with measured data based of variable message signs, etc. During the field test up to 1,000 receivers will be used. Forecast of the traffic impact of the applications under research depending on penetration rates and infrastructure equipment is done using VISIM and VCOM and an event model for communication.
Download the above presentation (381KB)
In his talk "TraNS: a Traffic and Network Simulation Environment based on SUMO
and ns2" Maxim Raya (EPFL) pointed out the challenges of modelling and validation
for getting realistic results. The shell TraNS is written in Java for integrating
the macroscopic and microscopic traffic simulator SUMO (from DLR) and the network
simulator ns2. TraNS is available from http://wiki.epfl.ch/trans.
Future versions will include a software bridge to SWANS and an implementation
of security mechanisms in ns2. TraNS is partially supported by the SeVeCom project.
Download the above presentation (358KB)
Elmar Schoch (Uni Ulm) reported on Efficient and realistic VANET simulation
with JiST/SWANS for investigation of the mutual influence of Car2X on highway
and urban scenarios. Because of the better performance compared to ns2 JiST/SWANS
can be used for network simulation of vehicular ad hoc networks, which typically
exceed several hundred nodes. The software is provided for download from www.vanet.info.
Upcoming extensions will include a more detailed radio model based on ray tracing
and security mechanisms.
Download the above presentation (390KB)
In his presentation "Extended ns-2 for simulation of inter-vehicle communications" Marc Torrent-Moreno (Uni Karlsruhe) informed about the modular design and improvements of the network simulator with focus on safety applications. Especially the radio channel model, the interference model and the reception model have been enhanced. However the propagation model still has to be improved further. The model of IEEE 802.11.p will be available from autumn 2007.
Download the above presentation (625KB)
In his presentation "DLR Simulation Environment m3" Matthias Röckl (German, Aerospace Center, DLR) addressed the urgent demand of a common simulation environment for the ITS Society in order to obtain a reliable comparison of novel algorithms and to show the benefit in safety, efficiency and/or comfort of novel ITS applications. This simulation environment has to incorporate various models for the driver, vehicle, environment as well as telematics in a highly adaptable and aligned way (from course-grained to high-fidelity modelling) to approach the different objectives of simulation.
Furthermore, Matthias Röckl gave an overview on the currently available simulators of the DLR which will be combined in a new integrated simulation environment called multi-modal multi-vehicle mobility simulation (m3). m3 complies with the above mentioned requirements and will be an independent modular test and certification environment which is inevitable for the progression of ITS.
Download the above presentation (624KB)
During his presentation "Protocol conception and validation: the experience of INRIA (MANET, NEMO and MANEMO)" Arnaud de La Fortelle (INRIA) started with a vision on inter-vehicle and vehicle to infrastructure communication followed by a recommended procedure for validation of protocol development, using simulation, but not only (e.g. experimental tests using various vehicles). He gave some examples of simulation applications starting from protocol development up to cooperative applications, illustrated by the COM2REACT project.
Download the above presentation (1.503KB)
Arne Kesting (TU Dresden) reported on "Analytic and simulated statistics and applications of inter-vehicle communication" for traffic jam prediction. Simplification of channel model and nodes distribution enables analytical analysis. The analytical results fit quite well with those of the simulation. The local prediction of a traffic jam based on the information received is not time-critical and store & forward showed more robust than multi-hop for the information distribution.
Download the above presentation (1.205KB)
In his presentation "Traffic Light simulation for Car-2-Infrastructure Communication" Floria Weichenmeier (GEVAS) pointed out the challenges for proper simulation results. The traffic light market can be characterised as conservative. Traffic light control algorithms in Europe are based on different methods (e.g. Germany: stage based, Netherlands: signal based) and depend on sensor signals as well as on public transport priority. Complete simulation of real traffic lights is not always possible, because not all information like dynamic data and control strategy are available. Upcoming standards will facilitate traffic light simulation in future.
Download the above presentation (493KB)
Bart van Arem (TNO / Uni Twente) reported on "Simulation of the impacts of a congestion assistant based on vehicle-vehicle communication on traffic flow". Driver needs investigated range from warning about the down-stream conditions up to driver support during the congestion. The calibration of the traffic model turned out as critical task using upstream and downstream measurement of a test area. Warning 500 m in front of the congestion showed high acceptance and good results. Further research on interaction with active gas pedal and cooperative ACC are planned.
Download the above presentation (1.103KB)
The concluding panel discussion stated that an ultimate integrated simulation is neither possible nor useful. Simulation models have to be adapted to the focus under research. However there is a big need in interchanging the use cases, the results and providing transparency on assumptions, modelled features and limitations of models. Verification of simulation results with real world experiments is important for establishing confidence in the results.
Rudolf Mietzner (COMeSafety, C2C-CC) illustrated the relation between the Car2Car Communication Consortium and the COMeSafety project and informed about their missions and objectives. Current major tasks are the harmonisation of the several C2X-architectures as well as the spectrum allocation in 5.8 / 5.9 GHz frequency band. IPR issues and cooperation policy are very important for the European Commission. Most of the 6FP projects have already cluster activities articles in their consortium contracts! The proposed roadmap focuses on consolidation and feed-back / dissemination of the project results. Based on identified needs of the projects also the proposal of a new consolidated project might be possible.
Download the above presentation (715KB)
The concluding wrap up and discussion pointed out that community cooperation without project is not appropriate for all kind of tasks. The simulation workshop uncovered some methodical gaps and further research needs related to flexible, hierarchical simulation models, the verification of results and the establishment of confidence in the results. Furthermore common use cases, data formats / interfaces are needed for tool integration / tool linking. The central questions and a proposal to overcome these needs shall be formulated and prepared by the following interest group:
- Mr. Jürgen Rataj (DLR), volunteered as initiative leader
- Mr. Bart van Arem (TNO)
- Mr. Thomas Kürner (TU Braunschweig)
- Mr. Jose Herrero (GMV)
- Mr. Manfred Miller (Nordsys)
- Mr. Schürer (Swisscom Innovations)
- Mr. Lübke (Volkswagen)
- Mr. Manuel Fünfrocken (HTW Saarbrücken)
- Mr. Hellbrück (Uni Lübeck)
Within its scope COMeSafety will support this initiative as well as further cooperation initiatives between the projects.