CBTC offers many benefits to metro operators. It can deliver very short headways enabling a high-frequency service to be offered which boosts capacity. Port Authority Trans Hudson (Path), which is installing CBTC on its busy metro linking New York with Newark, says CBTC will increase capacity by at least 20%. It also offers high reliability because of the level of redundancy built into it, automatic train protection, and full automation allowing driverless operation leading to potential savings in manpower and more operating flexibility.

The construction of new metro lines around the world, particularly in China, and the need to increase capacity on existing lines as more and more people switch to metro travel are the two main drivers for the introduction of CBTC.
In theory, nobody should be building driver-operated metros today, but they are. The rush to complete new lines as quickly as possible sometimes works against the installation of CBTC. The Indian government, which still remembers the horrendous problems and delays building the country's first metro in Kolkata, has banned the introduction of new technology on metro projects in order to complete them as quickly as possible following the highly-successful Delhi model. The government fears systems such as CBTC could slow metro construction.

Signalling is also a very small part of the investment in a new metro, which means some cities do not give it sufficient priority. As Mr Piero Galileo La Scala of Turin Metro, pointed out, it can also be a challenge to get new systems approved by the relevant safety authorities because, in their mind, new systems represent a danger.

There is no denying the introduction of CBTC represents a challenge for operators as it is a step-change in technology. Mr William Fellini, Path's assistant director of capital projects, said a peer review of their CBTC project by six other CBTC operators threw up two things they hadn't thought of: the need for a test track and the fact that it would take six years to implement the project rather than the four years which Path had estimated, due to the complexity of the technology.

It was pointed out that it is only really possible to introduce driverless operation on new metros or on existing lines where new trains are being introduced. This is because of the difficulty of converting old trains to automatic operation, the poor reliability of older trains, and the fact that even if a fleet of old trains was built at the same time, the trains tend to become different from one another over time due to repairs and modifications, which means the fleet is no longer homogenous.

Some operators seem to be taking the seemingly easier path of retaining drivers on fully-automatic trains rather than trying to negotiate a change in working practices to redeploy them to more productive work. But failing to grasp this nettle when CBTC is introduced is likely to make it far more difficult to eliminate drivers later. Retaining drivers will rob a metro of the real benefits of driverless operation such as the ability to increase the frequency quickly to cope with a surge in demand. One metro discovered that trains were being delayed because drivers stopped to chat during a crew change at a station. This wouldn't happen with driverless operation.

Nevertheless, Mr Gerard Yelloz of Siemens said it is possible for operators which cannot switch to full driverless operation to obtain some of the benefits by adopting it for empty train movements. This would cover train turnround at terminals, automatic wake-up at the start of service, and automatic operation in depots and through train washers. "Recovery after disruption is also much quicker than with manual operation," he said.

Mr Stephan Shirlaw of Alstom believes that CBTC should eventually become a global system with some standardisation. "There was an attempt to form a user's group early on, but London and Madrid jumped ranks," he told me. Shirlaw believes that CBTC should eventually become part of a communications-based transport system covering passenger information and overall control of a network.

Shirlaw says as the cost of computers is falling steadily, it should eventually be possible for operators to replace their CBTC systems every 10 years to reap the benefits of advances in technology. This would get round the problem of the increasing speed at which technology becomes obsolete and having to maintain systems well beyond their "use-by" date.

Perhaps the idea of almost throw-away systems will fill some operators with horror, but this is probably the future as computer-based technology continues to evolve.