Network Rail Limited Public Funding

**The evolution of Atypical Air Environment through UTM based SORA mitigation** NR, Dronecloud and Railscape (t/a RUAS) have worked together to create a Centrally Managed Drone Hub at NR's Rail Operating Centre (ROC) in Birmingham, leading to Operational Authorisation for Beyond Visual Line of Sight (BVLOS) under the Civil Aviation Authority's (CAA) Atypical Air Environment (AAE) policy (CAP 3040) to improve incident management and asset inspection in a limited real-world rail environment. This proved AAE as a viable mechanism for low volume, low altitude, short distance BVLOS to gather operational experience and data but stopped short of delivering a route to scalable, long distance, higher altitude BVLOS across CNI. The key (technical, regulatory and operational) constraints of AAE are: 1.Constrained to extremely low altitudes, close to linear assets and structures increasing ground risk 2.CAA require a semi-manual notification (NOTAM) for each flight, limiting rapid response and scale 3.Restricted volume through lack of regulatory mandate to share flight information amongst airspace users. 4.Industry lacking regulatory guidance on technical & operational solutions required scale out of limited AAE **Project Aim** Project SOCNI aims to use the UK SORA 2.5 framework to identify and demonstrate key mitigations based around Uncrewed Traffic Management UTM to enable the scaling of safe BVLOS, initially under the AAE but with the view to expanding operations into AAE+ and potentially other structures such as Temporary Restriction Areas (TRAs) & Transponder Mandatory Zones (TMZs). **DELIVERY** Engaging with Network Rail, British Transport Police (BTP), Transport for Wales (TfW) NATS en Route (NERL) and drone operators Railscape t/a RUAS over the CNI network, we will create a structured approach to designing, deploying and testing safety mitigations. We will test and assess the effectiveness and impact of each mitigation and provide a roadmap for deploying the strategy over CNI to enable scalable BVLOS with realistic milestone delivery targets mapped to a CNI specific roadmap.

This project aims to develop a productivity-enhancing AI solution for Structural Health Monitoring (SHM) of composite components, considering manufacture and service life, for the construction industry. Using AI in SHM is identified as a promising approach because of the need to collect, process and analyse large amounts of data, and recognise patterns which allow the assessment and accurate classification of key structural health-defining events which determine the condition of civil infrastructure and other engineering structures. Fibre-reinforced composites have been widely applied in automotive, aerospace, construction, energy, etc. The use of composite materials is set to grow due to their superior performance versus weight characteristics. However, drawbacks include high inspection rates, with inspection during service accounting for one third of operating costs. Long-term health monitoring, in the form of SHM, is required to assess the performance of composites from manufacture, through the point when a part enters service, until its end of life. Although the requirement for composite material monitoring is clear, the current technology used to monitor composite structures is unreliable (due to the fragile nature of existing sensors), the data generated is difficult to interpret and humans can miss critical events. The InteGraph project will address these limitations by using a novel graphene-based sensor system and developing a well-trained, bias-free, validated, transparent and safe AI solution based on Long Short-Term Memory (LSTM) neural networks for SHM.

**Rail Anomaly Detection and Reaction** The RADAR project team will be led by Transmission Dynamics (TD) and delivered in collaboration with a fellow SME, Komodo Digital (KD) and the RADAR Project Board; Network Rail (NR). Angel Trains (AT -- Rolling Stock Leasing Company) and Avanti West Coast (AWC - Train Operating Company). The partners have already delivered exemplary collaboration to develop and rollout a market leading pantograph and overhead line monitoring system (PANDAS V(r)). RADAR will build on this success to deliver a solution to the global Rail industry's critical need, to monitor and resolve asset and infrastructure anomalies more efficiently, via data driven decision making and automation of administrative tasks. RADAR will deliver self-learning automated pantograph and overhead line anomaly detection capabilities and demonstrate the ability to automate reactive measures to critical incidents. These rail market leading capabilities will be delivered through extensive Machine Learning, Artificial Intelligence and Generative Prescription Transformers, to enhance on board edge processing protocols via TD's existing, multi-award-winning, pantograph and overhead line monitoring solution, PANDAS V(r). This game-changing project will deliver disruptive innovations that are train, pantograph and location agnostic. This truly globally advanced Industrial Internet of Things solution will improve efficiency and productivity in asset and infrastructure anomaly detection and reaction. Moving way beyond leading remote condition monitoring and predictive maintenance decision support solutions, RADAR offers the ability to use live asset/infrastructure intelligence in real time, to instantly reduce risks, disruption and costs associated with prolific anomalies that otherwise deteriorate exponentially over time. With upto 50 trains per hour passing every anomaly and exacerbating them every time, RADAR's intelligent and self-configuring pantograph network interaction will incomprehensibly impact safety and likelihood of faults leading to further damage, risks of service effecting failures (delays) and potential for expensive to rectify (£1.5m each) dangerous and disruptive de-wirements. In addition to delivering the monitoring and automated reporting capabilities above, RADAR will develop an Artificial Intelligence driven Graphical User Interface to improve user adoption. It will learn individual user's requirements, level of engineering knowledge and role in the complex rail supply chain to give them the exact information they need, in a form they can understand, in order to promptly utilise it. This complex, user lead GUI work package will ensure the presentation of RADAR intelligence dynamically evolves to meet their individual needs, improving overall user experience and engagement, to maximise the speed of reaction to alerts.

Autonomous flight is a general purpose technology with enormous commercial potential. As a result a lot of resources are being put into solving for Logistics and Advanced Air Mobility (AAM). We want to provide these companies with the world's most advanced solution for autonomous unmanned flight and we took the first steps in FFP2\. In FFP2 the consortium led by [sees.ai][0] has developed a technology that enables BVLOS flights on demand, at low altitude and close to obstacles. In our system the Pilot designs the mission; our UAS senses and maps the world in 3D using Lidars and Cameras, allowing it to fly autonomously close to and amongst infrastructure. The 3D map is sent in real-time to the pilot who supervises the mission remotely for safety. This technology provided the necessary safety and risk mitigations that allowed us to obtain the first authorisation for routine BVLOS approved by the CAA. Although the system is designed to work in any congested area (including urban) the authorisation is limited to a number of predefined industrial sites in which we will build our safe flight record. In FFP3 we are pushing the technology and our operational safety case to extend our current capabilities to: * Enable Atypical Airspace (AA) BVLOS inspection of assets in the public domain by: * Advancing the capabilities of our DAA (Detect and Avoid System) so it can detect vehicles and people on the ground and any approaching aircraft * Embedding our operations into the wider aviation ecosystem by integrating them into a commercial UTM system. * Developing a system that can provide comms between UAS and pilot in areas with poor or no 4G/5G coverage. * Leveraging our advanced spatial awareness and our integration into the aviation ecosystem to create a solid Concept of Operations that will allow us to obtain one of the world's first approvals for AA-BVLOS. * Enable a pilot to control multiple UAS, an important step towards increasing the efficiency and scalability of UAS operations. We will also be aiming to be one of the first companies to obtain regulatory authorisation to fly multiple UAS simultaneously in AA. With these advancements our consortium will be hoping to contribute to the BVLOS infrastructure of the future. At the end of the project we will have a number of systems that will be tried and tested and ready to be deployed regularly by our clients. [0]: http://sees.ai/

The project will develop a new Location-Based Service (LBS) utilising novel technology to track the locations of trackside workers in the rail industry. The service provides the ability to define safe working zones and provide robust alerts if workers stray outside zones. The use of the new technology makes it applicable in cuttings, tunnels, under bridges and underground. The project will last 2 years with a total value of £991K. The consortium includes Guidance Navigation, Blue Frog Design, Network Rail, and Oxford University. There is a clear need for a technology that works, providing a LBS in GPS-denied terrain to provide awareness and management of worker location on the track, a dangerous working environment.

The properties of glass fibre reinforced plastic have been exploited to develop a revolutionary design of railway bogie which incorporates the suspension and damping functions within the glass fibre bogie frame. The design has been verified at 1/5 scale by more than 7 years of successful service operation of two bogies under a passenger wagon with Lakeside Railways, Eastleigh. Having developed the manufacturing procedures for moulding very thick components, full size components have been manufactured and the various sub-assemblies have been loaded statically and dynamically to ensure that the design concepts are fit for purpose. The next step will be to mould the full size bogie frame and test the preproduction bogie under a vehicle on a vehicle shaker rig and then to proceed to track testing and service evaluation. The benefits of such a generic design include a higher payload due to reduction in bogie mass, better load equalisation between wheels on twisted track, less contact between wheel flange and rail and lower audible noise and ground borne vibration emissions. The various design concepts can also be adapted to reduce the mass and improve the performance of existing bogies.

The project focusses on Enviromentally Friendly Transport within the Modern Built Environment. It will deliver a modelling and simulation tool to enhance safety and cost effectiveness of railway transportation through accurate understanding of in-service behaviour, component interaction and degradation. The lack of knowledge of in-service behaviour is a barrier to targetted innovation and is a reason for wasted research effort. The project deliverables will make a significant contribution to the design and manufacture of vehicles and the track system components. It further enhances the functionality and applicability of existing vehicle dynamics models by incorporating an accurate track system model. A much needed contribution will come form a scientifically robust model to assess the criticality of defects within the rail to enable knowledge based management of costly maintenance and rail renewal.

Train borne inspection of railway track has until recently been restricted to Ultrasonic or Electromagnetic testing and measurement, neither of which is capable of resolving shallow Rolling Contact Fatigue (RCF) cracks of a few mm long. Hence, these techniques are reactive to events & are not advanced enough to measure the key characteristics of RCF, i.e. shape, angle, length, linear density & position of crack initiation. They cannot be used to either enhance rail life or predict /prevent broken rails. It is this type of defect which was evidenced at Hatfield and is currently a subjective measurement resulting from manual visual inspection. This proposal details the development of an image acquisiton & analysis system enhanced by laser illumination & video imaging of the critical rail-wheel interface, in particular the contact band and the characterisation of visible rail head defects. This will be a preventative measure offering both safety and cost benefits. Image acquisition at high-speed, i.e. up to 125mph is a considerable challenge, requiring significant expertise. This is offered by the proposed consortium, along with ~8 years of rail defect data already obtained by Corus.

Awaiting Public Summary