SBRI: innovation in automated survey processing for railway structure gauging, phase 1
This is phase 1 of a 2-phase Small Business Research Initiative (SBRI) competition open to organisations of all sizes based in the UK, EU or EEA. A decision to proceed with phase 2 will depend on the outcomes from phase 1.
Network Rail will invest up to £720,000 (plus VAT) in phase 1 of this competition. An extra £1,080,000 (plus VAT) will be available for phase 2. This will support the development of a technological demonstrator for automated structure gauging processing. The competition process is managed by Innovate UK, as part of UK Research and Innovation.
Gauging is the process through which Network Rail ensures adequate clearance between passing trains and lineside objects and features. The objective of this competition is to:
- demonstrate innovations to Network Rail that will enhance the interpretation of point-cloud data
- provide accurate information on the location and identity of trackside features to enable accurate gauging clearance processing
The 2 phases will have the following objectives.
- training the application
- categorising 10 types of structure
- demonstrating the application
- analysis and recording of structures
For phase 1, your project’s total eligible costs must be up to £90,000 plus VAT. Phase 1 projects must plan to start by May 2020 and end by July 2020. We expect to fund up to 8 phase 1 projects.
Only successful applicants from phase 1 will be able to apply to take part in phase 2. We will monitor changes between your phase 1 and phase 2 submissions. We will not accept unjustified major changes in the project team or costs at phase 2.
For phase 2, your project’s total eligible costs must be up to £360,000, plus VAT. We expect to fund up to 3 phase 2 projects. Phase 2 projects must plan to start by September 2020 and last up to 12 months.
Your organisation must:
- be an organisation of any size
- be based in the UK, EU, or EEA
Your organisation can work alone or, if you do not have all the required expertise, you can work with others as subcontractors. Successful applicants will be expected to support quarterly progress meetings held in the UK.
Project teams must include organisations with expertise in infrastructure inspection, surveying and measurement.
We expect your demonstration to focus on data from an operational railway environment. If your organisation is new to the rail sector, we would encourage you to partner with organisations with appropriate rail experience. However, you will need to develop your own relationships with organisations with appropriate expertise.
To get help building your team, you may want to:
contact the Knowledge Transfer Network’s Transport team;
or contact railway organisations such as the Rail Industry Association.
We have allocated up to £720,000 plus VAT to fund innovation projects in phase 1. Successful projects will be 100% funded. The funding is provided by Network Rail and the competition process is managed by Innovate UK as part of UK Research and Innovation.
Applications must have at least 50% of the contract value attributed directly and exclusively for research and development (R&D) services. For this competition, your application should focus on prototyping and field-testing your product or service in a representative railway environment.
R&D does not include:
- commercial development activities such as quantity production
- supply to establish commercial viability or to recover R&D costs
- integration, customisation or incremental adaptations and improvements to existing products or processes
The aim of this competition is to develop a solution for automating railway structure gauging processing and to demonstrate this using data recorded from the live railway.
You must achieve these objectives:
1. Training the application to identify and categorise 10 types of structure and vegetation with accuracy equivalent to a human. A supporting technical specification will be made available to organisations registering for this competition, containing additional information on the structures.
2. Record accurately with cant, curvature and clearance in 5 metre slices of point cloud to meet the requirements of the standard structure gauging file template.
You must analyse successive runs on the same line taken months apart. You must record the rate and direction of movement of track relative to each structure, as well as recording any structures added or removed.
The success of the demonstrations will be based on their accuracy compared to current human-centric methods. Successful demonstrations must achieve comparable or better results than current processing methods, with the potential for cost saving and faster processing times.
The performance of the system will be compared to conventional manual inspections as part of the validation. You must share the technical specification with Network Rail as part of the deliverable report.
Your system must identify these structures:
- lineside furniture (miscellaneous minor structures)
- signal and telecommunications structures
Network Rail will give you point cloud data in .laz format or equivalent. For phase 1, you will be given a training set with identified structures, and a validation and test set. Each will cover approximately 10 miles of track and contain up to 1,000 structures.
You will not need to collect point cloud data as this will be provided by Network Rail.
If you are invited to phase 2 you will be given consecutive data sets of the same runs to enable identification of changes in the structures.
The competition aims to improve the accuracy and efficiency of gauging processing. Your project must accurately demonstrate:
- identification of existing structures by comparison with the existing National Gauging Database, which we will share if we award you funding
- correct identification of vegetation and masonry
- repeatability and reproducibility
- reading of structure distance and track cant and curvature
- identification of changes such as removed structures and added structures
- conversion of point cloud slices into structural gauging files