Network Rail claim that the new network has to be based on GSM-R technology. But GSM-R does not necessarily require masts (let alone 33-metre masts): other, less obtrusive, technologies may be used.
One from the same manufacturer, Siemens, is called the leaky feeder cable or Euroloop. On their website Siemens say that the leaky cable works under water, under snow, and under dust. London Underground and the Severn Tunnel have used leaky cable for years, as have the Swiss. Network Rail stated to the Tarka Action Group that leaky cable is a viable alternative system after all, which could be used instead of masts along the Tarka Line.
The other GSM-R-compatible solution uses satellite technology. Several GSM-over-IP (Internet Protocol) implementations are available, and satellite links can (and do) carry internet traffic (see below), so it would not be difficult to interconnect the existing GPS satellite system with GSM-R.
However, the European Union is creating its own satellite network, called Galileo, which is specifically intended to link-up with ERTMS (particularly the European Train Control System component). According to the section of the Galileo website devoted to rail applications, “The introduction of satellite navigation within the ETCS/ERTMS will… contribute to increase the performance on the high-density line[s] and lower the cost on low density and regional lines.”
Galileo uses 30 satellites, distributed over three planes in Medium Earth Orbit, “which ensures a uniform performance both in terms of accuracy and availability. Unlike GPS, it will be possible to receive Galileo in towns and in regions located in extreme latitudes.” Galileo is scheduled to offer an operational service from 2008 onwards - that is, at least seven years before ERTMS becomes operational in the UK.
But both have disadvantages, according to Network Rail (who obviously don't consider the disfigurement of our countryside with masts a 'disadvantage').
In the words of John Armitt, Network Rail's Chief Executive, in a letter to Dr Caroline Jackson, an MEP for the south-west: “The two basic altemative technologies capable of delivering GSM-R standard communications have considerable drawbacks. First, satellite technology is expensive and does not provide a solution for tunnels, cuttings and highly built-up areas. The second option, leaky feeder cable, is equally if not more expensive and poses real operational issues in terms of the amount of lineside installation required.”
A combination of the two technologies could well be the solution. Leaky feeder cable is particularly suited to tunnels and cuttings, and satellite could handle (most of) the rest.
Regarding the problem of using satellite technology in “highly built-up areas”: the press release on the Strategic Rail Authority's website (cited earlier) says, in item 7 of the Notes to Editors, “A particular challenge for the UK will be delivering the system in busy commuter and mixed traffic areas, where a technical review of the workings of the GSM-R mobile communications network has highlighted a lack of capacity and resilience as a major technical risk. Currently nowhere in Europe is anyone understood to be planning to implement ERTMS in this type of rail environment, which is common in the UK.”
A recent report on the weaknesses of GSM-R in such environments can be found in an article called Rail comms system could delay trains.
Another, and much cheaper, solution would be for Network Rail to make use of the existing mobile phone network for cab-to-signaller voice communications, with a much smaller number of additional masts to fill in gaps in the existing coverage, and satellite technology for ERTMS-related data communications. As mentioned previously, the Cullen Report was quite favourable towards the use of standard mobile phones by drivers and conductors.
Providing uninterrupted coverage to rail passengers is something of a Holy Grail for mobile phone operators. One or more operators would almost certainly be willing to provide Network Rail and the rail operators with secure in-cab communications at a heavily subsidised price in return for Network Rail's cooperation.
The Tyne and Wear Metro provides one example - Nexus, the Passenger Transport Executive that operates the service, cooperated with the four main mobile operators to install microcells (small antennae) on the network, giving passengers continuous mobile reception. According to an article on the project, “Nexus said that the technology provides additional safety capabilities, allowing train drivers and station staff to make calls to emergency services and control centres.” According to another article, Orange is working "in partnership with the necessary rail authorities" to provide mobile reception in tunnels on the East Coast, West Coast, and Great Western lines.
From another direction, GNER is planning to install Wi-Fi networks on its rolling stock - according to a BBC article, “Satellite dishes on the trains will connect to a variety of networks along the route, including digital TV connections and mobile base stations.” Virgin are planning something similar to offset the bad reception caused by the silvered windows of their new Pendolino trains.
ip.access is a Cambridge-based company that specialises in providing GSM and G3 access over IP (the internet protocol) using 'nanocells' - tiny antennae the size of a Smartie packet. IP is satellite-compatible, so a directional dish on top of the train could pick up mobile communications signals and distribute them throughout the train. A version of their system is operational on the Kiev Metro. There's more information here. Similar GSM-over-IP technology has been developed by Sweden's Telia Bright.
The most compelling reasons Network Rail's Head of Communications, GSM-R Project, could adduce for not using third-party masts are that “the GSM-R network will be a key element of our operational safety apparatus and as such, the masts will be dedicated to this use to minimise any operational risk. Particularly in sparsely populated areas mobile coverage is poor and masts tend not be in the right places for our intended purpose.” We imagine that most people would trust Vodafone or Orange, or almost anyone else, to manage a communications network better than Network Rail. And erecting new masts to fill in the gaps in coverage would be vastly cheaper and less disruptive than building a whole new network, and could be implemented far more quickly. Unfortunately, it seems that Network Rail is institutionally averse to cooperation - see Whatever happened to DART?
And communications technology is advancing all the time. GSM-R, which is based on GSM technology developed in the 1980s and standardised in 1991, is already obsolescent, overtaken by G3, and G3 in turn is threatened by several alternative technologies.
Here's a quotation from the Final Report on West Coast Route Modernisation to the Office of the Rail Regulator, published in June 2000:
The GSM system is being superseded by the third generation mobile telephone system (UMTS [G3]). Although one telecoms operator has stated that it will not guarantee GSM beyond 2009, it is likely that GSM-R will not become technologically obsolescent for about 15 years, by which time the systems would in any case be life-expired.
West Coast Route Modernisation, Report to Office of the Rail Regulator, Booz Allen & Hamilton, June 2000, Section 3.58 (949kb PDF)
15 years from the date of publication takes us up to 2015 - which is the earliest date that ERTMS can be implemented. (See Delays?).
The committee of the UIC International Union of Railways that developed the GSM-R standard is well aware that GSM-R is in its twilight years. Here is a rather tetchy and defensive quote from the FAQ page on their GSM-R website:
Of course, there might be better solutions in theory today if we were starting again. But we're not. GSM-R is today's mobile communications system for the railways. The objective was to find a solution that meets the railways' overall requirements (economically and technically). GSM-R is this solution.... In any case, if we were to start from scratch now we would embrace other possible solutions in the study [our emphasis] - software radio for instance or UMTS [G3].
Bearing in mind that GSM-R is only required by ERTMS Level 2, that the UK is the first (and so far the only) EU country to commit to implementing ERTMS Level 2 on all its track, and that the UK timetable for implementation has been constantly deferred, by the time GSM-R is brought into operation the technology it's based on will be 20 years old or more. “Starting from scratch now” sounds like an excellent idea (though it may still be slightly premature!).
Another promising technology is stratospheric broadband using High Altitude Platforms (unmanned airships or solar-powered gliders flying at an altitude of around 20 kilometres - well above normal aircraft but substantially below orbiting satellites), which has much higher capacity than satellite but doesn't require the masts needed by terrestrial networks, and is cheaper than either. One application the developers consider ideal for this technology is high-bandwidth communication with moving trains using directional antennae.
There's more about the Capanina project (which is EU-funded) here.