Title:
Performance Analysis and Evaluation of Advanced Designs for
Radio Communication Systems for Communication-Based Train
Control (CBTC)
Supervisors
Principal supervisor: Prof Lars Dittmann
Evaluation Board
Lector Ying Yan, DTU Fotonik
Research Director Marion Berbineau, IFSTTAR, France
Signal Systems and R&D Manager Matias Sevel Rasmussen, Metroselskabet & Hovedstadens Letbane, DK
Master of the Ceremony
TBA, DTU Fotonik
Abstract:
Communications-Based Train Control (CBTC) is a modern signalling system that uses radio communication to enable the exchange of high resolution and real-time train control information between the train and the wayside infrastructure. A vast majority of CBTC systems worldwide use IEEE 802.11 Wi-Fi as the radio technology mostly due to its cost-effectiveness. The trackside networks in these systems are mostly based on conventional infrastructure Wi-Fi. It means a train has to continuously associate (i.e. perform handshake) with the trackside Wi-Fi Access Points (AP) as it moves. This is a time-consuming process associated with a certain delay. Additionally, these APs are connected to the wayside infrastructure via optical fiber cables that incurs huge costs.
To address these problems, a novel design of the CBTC trackside network was proposed at Siemens. In this design, trackside nodes function in ad-hoc Wi-Fi mode, which means no associations have to be performed with them prior to transmitting. A train simply broadcasts packets. A node upon receiving these packets forwards them to the next node and so on, forming a chain of nodes. Following this chain, packets arrive at the destination. To minimize the interference, transmissions are separated on multiple frequencies. Furthermore, redundancy is introduced in the design as a node forwards packets to not only one but two of its neighbors.
The research work presented in this thesis investigates the performance of this new design using computer-based simulations. A large number of scenarios were investigated, in particular with the objective of studying the resiliency, redundancy and scalability supported by the design. The results from the first phase of the study show that due to the frequency separation and redundancy inherent in this design, significantly large numbers of packets can be successfully transferred across large networks. Nonetheless, the results expose two shortcomings of the design as well. They show that the train node undermines the frequency separation guaranteed by the chain nodes as it is required to transmit on all frequencies, and, the design under-estimates the interference produced by distant nodes in ideal propagation conditions despite the frequency separation.
A large number of potential solutions to minimize these shortcomings were subsequently investigated, including adjusting the transmission range of the train, employing a lower number of radios on the train, employing more robust modulation and coding schemes, among others. Additionally, two extensions to the design were proposed that involved extending the frequency separation distance by employing additional frequencies, and, introducing a separate frequency for the train-to-trackside communication. The results show that substantial improvements can be achieved as a result of these solutions.
In the last phase of the study, scenarios to investigate the impact of parameters such as the number of trains, train speed, headway distance, train's location on the track, train direction, and track layouts, were carried out. One of the objectives of this study was to investigate if high data rates can be supported by this design, to enable non-CBTC applications on top of the typical CBTC traffic. The results for these scenarios show that while the design can successfully support the data requirements of typical CBTC traffic, enabling higher data rates is challenging when the number of data traffic flows involved becomes large.