Taking a logical approach offers interoperability benefits
INTRO: Recent advances in data communications are enabling off-the-shelf equipment to replace purpose-designed links in transmission-based train control. The adoption of a common architecture based on standardised interfaces and modular design will enable operators to integrate proven equipment from different suppliers
BYLINE: Peter Ludikar
Product ManagerAlcatel Transport Automation Solutions
A CONFERENCE on Transmission-Based Train Control in Washington DC last year suggested that it was an immature technology. But as far as urban rail is concerned, the truth is exactly the opposite. For almost two decades TBTC has been alive and kicking, in the form of SelTrac¨ moving-block signalling. Although some suppliers are only now attempting to emulate moving block, SelTrac has lived up to its promise of high reliability and performance and dominates the world map of TBTC installations.
Traditionally, SelTrac has been applied in a centrally integrated fashion, offering automatic operation and other management benefits. Major greenfield applications include the Vancouver Skytrain, London's Docklands Light Railway, the Scarborough RT in Toronto, Ankara's heavy metro and Putra in Kuala Lumpur, amongst others. Vancouver now has over 15 years of practical experience (p309).
But SelTrac has also been applied to existing lines and extensions. One of the most interesting is San Francisco's Muni Metro light rail subway. Last year Alcatel TAS successfully completed a 'proof of throughput', running 48trains/h in one direction through the subway in full automatic operation. Muni's light rail operation is one of the most complex in the world, with over 600 trains entering the automated section each day from the manually-driven street routes. Adding to the complexity is the need to provide automatic protection and routing of unequipped trains on the same lines.
But can this 20-year-old approach still maintain its relevance? Over the last couple of decades electronics has consistently outpaced itself. The original hardware was bigger, slower and significantly less complex than that available today. In practice, the newer breed of processing platforms has made little difference to the actual train control application; SelTrac was quite advanced for its day. But the new hardware has opened up several possibilities that will affect the way signalling is carried out in the future.
Logical block overlays
Alcatel's 'logical block' technology is applying the SelTrac concept to the resignalling market, where it is proving effective in meeting urban interoperability and mixed-mode requirements. For overlay applications, the technology is designed to work in conjunction with existing interlockings. It is applied in a modular, distributed fashion for Speed Enforcement (Toronto Transit Commission), Supervisory Performance Overlay, Automatic Performance Overlay (RATP, Paris), and Automatic Management & Operation (Las Vegas Monorail).
With modular ATO implementation, the design is accomplished using software logic and resident parameters to define block length, train characteristics and other critical system features. But the system is not reliant on physical track circuit blocks. Vehicle-centric 'block' logic is built into the software, and can be customised to suit different types of train and track layouts. This allows more efficient block lengths to be set within the parameters of the fixed-block locations, enabling trains to operate closer together safely and improving throughput.
The single biggest advance in TBTC technology is the widespread use of modern data communications to carry messages from one controller to another. SelTrac still uses vital on-board computers that communicate with vital computers at the trackside. But the purpose-designed dedicated inductive communications-based link used in earlier applications can now be replaced by a high speed data network.
A major challenge facing urban rail operators is the maintenance and replacement of infrastructure, with the cost implications ruling out frequent wholesale renewals. The number of suppliers of conventional relay-based signalling has dropped significantly in the last 20 years, but maintenance costs are not falling.
Today's alternative to relays is electronics, but there are as many options as there are suppliers. So one of the major procurement decisions is whether to risk being married to the same supplier for the life of the equipment. A related concern stems from the recent phenomenal advances in electronics. Equipment bought five years ago is unlikely to have a modern drop-in replacement, making the long-term supply of spare parts an inevitable concern.
A systematic industry-wide interoperability standard would help overcome these worries. Modern data communications capabilities offer a major opportunity to move towards interoperable signalling equipment. Networking technology will ease the process, assuming that other suppliers are willing to make it happen. This is not an unreasonable assumption given the precedent set by the GSM cellular telephone suppliers.
'Interoperability' will rapidly become part of our common train control vocabulary. The driving force is not the fact that the metro in City X wants to start through running with suburban rail services in Region Y. It will be the economies of scale inherent with a flexible approach to technology design, application and operation. Until an acceptable standard is established, each supplier will continue to provide its own proprietary solutions.
On the technical side, train control architectures based on standard internet protocols (IP) now permit vital train control processors to be physically and functionally separate from the communications network. This means that the train control infrastructure can potentially be shared by other systems, such as health monitoring, staff management and even public information displays. Alcatel has shown this approach to be viable in three successful field demonstrations.
In June 1998, a demonstration of radio-based train control included a simple computer conferencing session between on-board and trackside computers. This included full driverless automatic control of a train and sending video pictures between the computers at the same time. During a second demonstration for the Canarsie TBTC project in New York (RG 2.02 p83), test operators on the train kept in contact with staff on the trackside by sending text messages over the network while trains were being safely controlled.
In the last demonstration on Paris metro Line 8, the network was used to download programs to computers on the train from a central location, as well as providing network and train control diagnostics capabilities. The Paris test also included a demonstration of a third-party video surveillance system. This clearly exemplified the benefit of using a standard networking interface, because it took seconds to connect and operate the system over the train control network. All of this occurred while trains were being controlled over the same network.
The recent award of a contract to resignal Paris Line 13 has validated Alcatel's approach in offering interface characteristics that naturally lend themselves to interoperability. This uses open standard network protocols at all intra-system interfaces. Thus any vital unit that needs to send messages to other vital units simply addresses messages to the appropriate destination and the data communications system takes care of delivery.
Distributed communication is provided by a spread-spectrum radio/data system meeting IEEE standards. The radio communication between the train and the trackside is completely transparent to the vital computers. In fact, the radios can be replaced by hard-wired connections in the laboratory without any change to the vital controllers.
Demand for interoperability
The future for TBTC is certainly bright, but in today's competitive mass transit signalling environment operators are increasingly looking for interoperability to avoid being locked into proprietary solutions. As the technology concepts become more viable, standardisation programmes such as IEEE RTVIS WG #2 and Europe's UGTMS have surfaced to deflect the danger of the industry identifying and settling on an old system as the interoperable standard. This closely reflects to the parallel ERTMS initiatives in the main line sector, which are addressing exactly the same concerns.
To-date, however, there remains disagreement and misconceptions on the definition of 'interoperability' in the urban sector, and the expectations of the technology. These range from functional compatibility with subsystems and external communication systems, to physical interface interchangeability. There has been much focus on enabling technologies such as data communication standards, radio air gaps, and balises, which has only served to confuse the real interoperability issues - safe train control behaviour at interfaces, guideway modelling, and train control messaging protocols.
We believe that to achieve true interoperability, the system interfaces must be radically redesigned. Ideally, the industry should adapt common and stable commercial standards for the rail environment. By using a layered communication approach, most of the difficult interfaces can be easily dismissed and the resulting interface specifications can focus on the critical train control information exchange needed, not the detailed protocols.
With its modular architecture, robust interface design and behaviour description documents, SelTrac focuses on independence from low-level proprietary interfaces and concentrates on application-to-application interfaces, isolating lower layers to commercial standard solutions (box p308).
The pioneering concepts of SelTrac are still relevant today, and, in conjunction with advances in data communications standards, have led directly to a real basis for an interoperable standard. SelTrac has come a long way and will definitely light the way to the future of TBTC.
CAPTION: Alcatel has recently been awarded a contract to resignal Paris metro Line 13
Photo: RATP/ J F Mauboussin
CAPTION: SelTrac has been successfully applied to two routes in Vancouver
Interoperability solved with application interfaces
Modular architecture for train control systems: