Tadorea embraces Knowledge Discovery and machine learning solutions, based on massive data analytics for the aviation sector.
Tadorea aims to extend and apply the research results achieved by its founding entities to the needs of aviation stakeholders.
We leverage the research results in the aviation field and our aviation data infrastructure built during the last few years, including a variety of simulation models built in the domains of mobility, GHG emission assessment, delay propagation, passenger connectivity and network disruption management.
We build scalable approaches to data acquisition and analytics, capable of growing with the increase of data and demand for data-driven decision making. This is done through meticulous interfacing with end users and addressing issues of digitalisation, deployment infrastructure and industrialisation.
Data comes in different formats and with different confidentiality requirements. Whether it be structured or unstructured, the lack of explicit semantics and data silos of heterogeneous data sources without interoperability are common issues.
At Tadorea, we ensure the quality of the data, data provenance and data lifecycle management processes.
Dedicated servers, cloud computing or distributed ledger technologies (blockchain) provide infrastructure solutions to different requirements.
We work on a variety of architectures to manage heterogeneous data at-rest and data-in-motion, making sure the performance of our algorithms provides scalable, and when needed, provide, real time analytics.
There are different analytics frameworks which provide business analytics and intelligence. Integrating those analytics with the business processes is key.
We work with state-of-the-art machine learning techniques to design the predictive analytics that provide optimal value to the business processes.
We work with privacy-aware technologies which allow machine learning on fused datasets while maintaining data privacy.
We design cryptographic solutions, including secure multi-party computation when needed, as a much more powerful alternative to de-identification.
Cryptography allows the design of token networks, to remove friction by aligning network participants to work together toward common goals. Efficient mechanism to incentivize network participants allow to optimise the use of common resources.
We use visual metaphors as a tool both for communication and data mining. We develop interactive interfaces that ensure visual data discovery and exploration for our users.
This constitutes a complete air transport simulation tool including a wide range of performance and mobility metrics, for a variety of uses for airlines, network managers and policy makers. This simulator is capable of modeling passenger connectivities and a wide range of flight and passenger prioritisation scenarios. This simulator has been built using airline and other industry inputs, and captures airline decision-making including their related costs, by fusing a variety of data sources.
Click here to visit Mercury’s website.
See here a video presentation about Mercury.
Tadorea efficiently processes flight information from different data sources, including more than 20 million flights that have operated in Europe within the last three years. A production-ready infrastructure enables data mining for the development of accurate predictive analytics, which are ultimately available for use by airlines, airports and air navigation service providers for strategic decision-making or real-time oversight.
We can monitor runway performance, runway occupancy time, and aircraft operations in order to predict safety and performance issues. This is all possible by taking into account the aircraft type/operator, meteorological conditions, runway configurations, threshold times, touchdown points, rotation points, line-up times, exit/entry points, excessive runway occupancy, runway over run, missed exit points, and other similar factors.
We develop insightful performance metrics for air transport scenarios to manage the cost of disruption of a delayed flight in the network context. Monitoring flights and the connections passengers are making, we identify the likelihood of potential problems and the associatedcosts. This is achieved by running continuous simulations with all the information available. Sources of data to be used will come from, inter alia, network (traffic) data, data on the current status and forecast of potential air traffic congestion (using real time traffic data), passenger inventory systems (for itineraries and connectivity data), and flight planning systems.
Through data mining tools, techniques and procedures we’re able to optimise fuel use in airline operations taking into account the complexity of the scenarios which presently characterise air transport operations. This is completed through observing the airspace capacity restrictions and flow control, airport congestion, meteorological conditions, fleet and equipment mix, personnel training, and other factors.
We are able to perform various computations of aviation emissions based on a requirements-based, quantitative approach. This also includes a validation of CO2 estimates even when the data is incomplete or of unknown quality. For this, we utilise emissions modeling, aircraft performance models, mathematical modeling, statistical analysis and operational expertise.
We are able to examine airspace performance by exploring the relationships between sector capacity, workload and complexity. We do this through machine learning techniques that are used to analyse large quantities of historical data, that are ultimately used to provide a set of predictive analytics that provide the air traffic controller teams with a deeper understanding of the most likely scenario to occur considering the information available.
Our technology enables evaluation under novel flight and passenger prioritisation scenarios, plus future regulatory environments (such as changes to Regulation 261), exploring trade-offs and characterising delay through the network.
This analysis is based on data-driven metrics for safety monitoring and safety intelligence. This leverages unique data in which we can develop predictive analytics for safety, including safety performance indicators for a variety of scenarios, such as: unstable approaches, real time approach congestion monitoring, airprox, lack of proper separation with terrain, or adherence to air traffic control procedures.