Inspiration
Train operators often have a large range of assets on their networks, which can be difficult to keep documented and maintained. Network Rail, for example, have over 17,000 building assets alone, however there are large gaps in these databases. In this case, around 40% of Network Rail building assets have ‘unknown’ location data, meaning there is no official record of where these assets are to an accuracy of less than 100m. Staff trying to locate these assets must rely on basic ELR data, and other information such as postcodes, which can often be broad and inaccurate. This wastes the time of engineers who often get lost trying to find these assets, and leads to high lost staff hours.
In addition to this, any work that is conducted on these assets is difficult to track. Problems are often discussed via private email chains and are not accessible to other staff members. This leads to ambiguity in the overall condition of assets, which has knock on effects regarding overall safety and reliability of these often critical structures.
What it does
AssetLoc8 is a combined mobile application and web base dashboard solution that provides an asset management system, allowing engineers and staff to efficiently identify and attend Network Rail assets such as buildings. This solution delivers benefits in four major areas:
Traceable audit trail: AssetLoc8 stores records which document every step of a workflow process following maintenance tasks on an asset. These records are stored in a shared environment which allows all staff members to view the job history of that asset, including comments and photographs.
Standardisation: AssetLoc8 provides a platform to input data surrounding an asset using a set array on input fields. This results in a standardised and regular way to add asset information to an asset profile.
Accurate GPS definitions: One of the main challenges faced in this area is that, as mentioned, around 40% of all Network Rail building assets have no specific location information. Therefore a vital characteristic of AssetLoc8 is the ability to define the location of “location unknown” assets, either by using the users current GPS location, or by placing a marker on a map combined with images and details describing nearby landmarks and identifiable scenes if GPS is not available.
Active feedback loop: Staff are encouraged to update assets with additional information and pictures where possible, promoting a consistently up to date profile portraying the condition and work history of all assets.
How we built it
Database: Firebase database was used to host our data and facilitate the communication between the mobile application and the dashboard.
What3words API was also used to convert lat/long point data to a 3m x 3m square grid to accurately find locations and add them more quickly, easily and with less ambiguity.
Mobile: Our mobile application is constructed using Android Studio and particularly making use of the Google Maps API/SDK.
Dashboard: Our dashboard solution is constructed using HTML and Leaflet JavaScript library with embed spatial data.
Challenges we ran into
Combining separate workstreams on the app (photo capture, map data capture, and general UI) in to one single application. This was particularly challenging when it transpired that identical file names had been used by different members of the team, which caused errors when files were merged.
Accomplishments that we're proud of
Reaching beyond the initially provided scope and integrating solutions to additional problems to our design, in particular the ability to closely manage and monitor the condition of assets and the required work to maintain them.
Additionally, we were proud with the speed at which we turned around a working tool ready for demonstrating, following a string of setbacks and challenges getting to that stage.
What we learned
A key area which was explored and developed by the team was the use of Google Maps API and spatial data. In addition, Android app development for different functions (eg. maps specific, basic templates, etc) was built upon quite heavily due to the integral nature of this aspect of the project.
PostGres is an open source database which allows spatial data to be pulled from mobile application and dashboards. We recognise that any organisation implementing our solution would likely use the database infrastructure they currently have.
What's next for AssetLoc8
We have some longer-term pipeline developments such as:
Develop augmented reality capability and utilise the mobile phone and iPad cameras to overlay asset metadata over the assets it is looking at.
The development of a chatbot which could be accessed via the app in the event that the engineer encounters a problem with their work. This chatbot would operate as a Machine Learning algorithm, and so with further use it could potentially learn how to solve simple and common problems, negating the need for additional engineering resources to get involved.
Using drones as a support to the AssetLoc8 system, collecting data remotely on behalf of engineers, who are controlling the drones from a distance. This increases the safety aspect of the system as data can be collected without staff having to get dangerously close to the lineside.
The potential to combine image recognition on the front of trains with GPS data, allowing lineside building assets to be recognised in video footage, their identification confirmed by GPS location, and then an image of the asset uploaded to the profile.
Log in or sign up for Devpost to join the conversation.