[Back]

A. Alonso:
"Dependable Medium Access Control for Road-Traffic Safety";
Supervisor, Reviewer: C. Mecklenbräuker, E. G. Ström; E389, 2013; oral examination: 09-19-2013.

The IEEE802.11p is an approved amendment to the IEEE802.11 standard to add wireless access in vehicular environments (WAVE). It defines enhancements to the 802.11 required to support Cooperative Intelligent Transport Systems (C-ITS) applications. The IEEE802.11p MAC algorithm (namely EDCA) is CSMA-based using carrier sens- ing as method for monitoring the state of the channel before transmitting. In loaded vehicular ad-hoc networks (VANETs) the channel access delay of a vehicle attempting to access the channel increases unpredictably every time it is sensed busy. The Quality-of- Service (QoS) requirements demanded by safety-critical applications cannot be fulfilled. In contrast, Self-Organizing Time Division Multiple Access (SoTDMA) guarantees an upper bound on the channel access delay defined by the Selection Interval length (SI).
The European Telecommunications Standards Institute (ETSI) defines two types of safety-related messages: Cooperative Awareness Messages (CAMs) and Decentralized Environmental Notification Messages (DENMs). Both have periodic nature and each packet has a deadline to meet. In the United States CAMs are also known as Basic Set Messages (BSMs).
My research has focused on the field of vehicular communications. Initially on the evaluation of the performance of the standard MAC protocol in congested vehicular scenarios, where safety-related data is present and transient congestion control is required. Subsequently, the focus moved to the design of enhancements at the IEEE802.11p MAC layer so that dependability is assured for high-priority traffic, regardless of the vehicular traffic situation. The first step was building a simulation environment to compare the performance of different MAC protocols. My design of the MAC layer takes into account a self-managed and scalable communication environment. This work relates especially to safety-related applications, which may share real-time information and cooperate in an efficient, affordable and reliable way. The next step was identifying challenging vehicular scenarios and defining new benchmarking metrics for evaluating the performance when scheduling traffic safety data. Finally, analytical expressions for validating the simulation tool are described for further design refinement.
The simulation results show that in high-density vehicle situations a congestion control method is required in order to provide reliability and robustness to safety-related communications. Although SoTDMA is proven to outperform plain EDCA in terms of predictability, it still fails to reach the selected relibility threshold set in this thesis, which requires that 90% of the generated safety messages are correctly decoded. In ad- dition, the ETSI proposal for variable report rate is a drawback to the use of SoTDMA. Everytime the report rate is changed, a vehicle using SoTDMA enters the initialization phase and a message is dropped at the transmitter. This is unacceptable for safety-related data. My approach for achieving reliability is to design a suitable decentralized congestion control (DCC) mechanism on top of the pre-existing IEEE802.11p MAC. This work underlines the importance of a suitable parameter setting, namely the carrier sensing threshold (CST) value selection, so that reliability is achieved and sustained regardless of the traffic density. In this thesis, it is shown that the proposed multistate- active DCC mechanism achieves a robust and reliable performance when its parameters are properly selected.

Cooperative intelligent transport systems, vehicular ad-hoc networks, medium access control, EDCA, SoTDMA, vehicle-to-vehicle communications, safety-related data, CAMs/BSMs, reliability, dependability, real-time communications, decentralized congestion

http://publik.tuwien.ac.at/files/PubDat_224918.pdf