Signalling design is going through many changes and one of the overriding aspects of this is the integration and expansion of software tools. Desktop signal sighting is one area that Gioconda has been developing over the past five years; it hasn’t been as straightforward as it may seem.
Traditionally the signal sighting process has involved lengthy site visits involving both engineers and operators. Quite apart from the expense and safety implications, it also suffers from the historic situation of being based on its own chainages which often do not correlate with the other design disciplines.
In March 2010, RSSB authorised a UK-wide derogation to allow initial signal sighting to be completed using desktop tools – a major stepping stone in the use of such technology. Prior to this, any project wishing to use virtual or 3D imagery had to apply for a project-specific derogation. Network Rail’s future aspiration for the signal sighting standard is that desktop solutions become the norm.
These tools can be split into two main categories – 3D virtual reality (VR) and high definition (HD) video.
A model approach
3D VR systems have been around for 15 years but have only been widely used over the past five years due to the availability of computing power and reductions in cost. They are often closely compared with training simulator technology because of their visual similarity but are actually just 3D CAD models with some added functionality and libraries. This is good news because, as time goes on, more disciplines are expected to be working in 3D CAD and thus be able to provide at least a basic model for wider use.
Gioconda’s approach is to offer a service solution – ‘we build, you operate’. In practice that means that the client does not need broad 3D knowledge to build all the elements to visualise a railway, just how to operate it and become proficient in making simple changes such as moving a signal to new chainage or changing the height and offset of an aspect. When it comes to major updates like a new p-way alignment or alterations to structures then Gioconda does this back at base or via its resource in India for a very quick turnaround.
Models are based on real-world data and this is where the process usually gets held up. There are many sources available but, in general, they are reliant on what the client can provide as outputs from surveys and other design disciplines. The problem is that, on many projects, these outputs are programmed to be available concurrently or after signal sighting is to have taken place. To overcome this, Gioconda adopts an iterative approach and simply uses the best available information. After all, if the process were to be done manually, the same problem would exist.
The company delivers a Grip 3 model for initial assessment, a ‘bare-bones’ Grip 4 model for sign-off by the Signal Sighting Committee (SSC), a detailed Grip 5 update and another ‘as-built’, as well as driver briefing resources. Amongst the source materials are scheme plans and designs, point clouds and other survey data, together with OS maps.
A principal benefit of the full 3D approach over HD is accuracy – the software is capable of mm precision. The main criticism of it is the sometimes ‘unreal’ view it gives. There is no need for this with currently available software but constraints such as budget and time often play a part.
So Gioconda’s aim with its process is to try and get the scene to look recognisable. Aside from accuracy, this also seems key to getting a Signal Sighting Committee to accept and buy into the use of VR, because of course they still have the right to attend a site visit if they see fit. To achieve visual accuracy, the company will endeavour to obtain HD video through the existing routes for reference and is finding that ‘Bird’s Eye’ views available on the internet are a fabulous resource.
Tracking the video footage
The HD video approach offers something quite different. The end result is a completely real view of the railway but with the assets removed or added to reflect the proposed signalling and signage layout.
Once the material is captured using a calibrated camera setup, Gioconda applies tracking technology to correlate the video runs with specified chainages. It’s at this stage that the raw footage for each run is turned over to the project team. It is supplied within the Gioconda Route Assessment Tool (G-RAT) interface, enabling the signal engineers to reconcile early draft scheme plans with the actual terrain. Simple measurements of offset and height to assets and structures can be taken and recorded. The project team can then start to build up a library of associated files with locations along the route so if the chairman has been on site with their camera, it is a simple task to import and associate pictures with correct chainage locations in the video. Similarly, PDFs, Signal Sighting Forms (SSF) and any other common-format files can be associated with a location down the route.
Using the in-built tools, the design team can use G-RAT to assess possible locations for signals and signage prior to developing the scheme plan. By now the project team will have produced working plans and early draft SSFs. These are used by Gioconda’s 3D modelling department to begin creating the new signals and other virtual infrastructure. These complex models will either be tracked into the video footage or placed within a full 3D VR model of the terrain.
The virtual signals now have to be motion-tracked to ensure they remain in the correct position within the scene. This is a complex process that involves the creation of a virtual camera in synchronisation with the camera used to capture the footage. The process is further complicated by the need to accurately simulate obstructions. These factors are essential if the desktop tool is to provide reliable sighting information.
Learning through experience
A more recent development is the use of the G-SST (Gioconda Signal Sighting Tool) for signalling immunisation on the Edinburgh-Glasgow Improvements Programme. This has proved to be a tremendous success and allowed the SSC to be completed largely as a desktop exercise, saving many man-hours walking the track. As time goes on, Gioconda’s technology is helping projects to achieve well over 90% of signal sighting in this way; it has already been tried and tested on many major schemes, bringing a host of benefits.
Through this practical experience, the company has ironed out many of the pitfalls in both the process and with data inaccuracies, and has been able to educate the engineers concerned as to what they can and cannot achieve with these tools. However the wish list is long, with new ideas emerging from every project!