Big Data for Training – and More - Civil Aviation Training

Big Data for Training – and More

The civil aviation community is using Big Data to enhance pilot training programs, and is finding use cases for large data sets beyond – throughout the sector’s ecosystem. Group Editor Marty Kauchak provides a glimpse of the surging demand.

At the same time that aviation organizations are increasing their use of Big Data to strengthen their pilots’ training performance, the civil aviation community is on the cusp of using large data sets well beyond training – in safety management systems, the build of aircraft and in an expanding number of other applications throughout the enterprise.

Foremost, new products are allowing community stakeholders to use Big Data to elevate their continua of pilot training.

CAE, in one instance, is staying ahead of the civil aviation sector’s rapid migration from paper-based training records to electronic training content – a trend which has left a number of community-wide disconnects and gaps the company is attempting to bridge. Chris Ranganathan, Senior Director for Training Strategy & Services, framed the challenge: noting that airlines generate, collect and store data on two very important parts of the pilot training continuum, selection and performance, he observed, “There is no attempt that we can see, on a continuous basis, to connect those two parts of a pilot’s journey” and rhetorically asked, “Is anyone systematically looking for correlations between pilot selection data and pilot performance and training data?” He pondered that, once a pilot is trained, whether training organizations are examining how the pilot actually performs, and using this information in the evaluation of their training programs?

The Dubai-based executive explained CAE’s solution to these shortfalls, noting the company “believes a true calculation of training effectiveness must be to also measure the operational world. In other words – does what we measure in training lead to relatable improvements in performance – safety performance or efficiency performance. No one is making that connection.” This effort will, in turn, also lead to broader discussions of what types of information should be collected from the training world, and is relatable to measurable information in the operational world.

Enter CAE Rise, a recent product offering, which is helping to take the common practice of measuring pilot effectiveness in the training world to a higher plateau. Indeed, the S&T provider notes CAE Rise allows operators to “gain access to a new data source that, combined with existing flight and training data, can further evolve the training programs.”

A second new entrant in this space is FlightSafety International’s FlightSmart integrated pilot performance evaluation and training tool.

FlightSmart applies artificial intelligence and machine learning algorithms to evaluate a pilot’s ability to perform critical tasks and maneuvers during all phases of flight. Image credit: FlightSafety International.

FlightSmart provides a glimpse of one important trend in this space: it applies artificial intelligence and machine learning algorithms to evaluate a pilot’s ability to perform critical tasks and maneuvers during all phases of flight. Any identified deficiencies result in a remedial training action path, personalized to the pilot, to increase proficiency. FlightSafety believes the tool’s capability to automatically predict student performance and identify corrective action is a key differentiator, beyond any other simulator data-driven product on the market.

Significant from another perspective, the product represents another instance of cross-pollination between the commercial and military training markets. In this case, the launch customer for FlightSmart is the US Air Force Air Education and Training Command for implementation on T6A initial and operational flight training devices.

The requirements to which FlightSafety and IBM, its product partner, responded should resonate with CAT readers: the military service was seeking to maximize the efficiency of its pilot acquisition pipeline by increasing student throughput to meet a shortage of pilots – without increasing the size of its instructor cadre and its workload.

Ideagen’s Head of Aviation, Steven Cespedes, told CAT, “This sector has helped mold our software over the past two decades. The best example of this software evolution is demonstrated in the research and development efforts that have gone into our newest cloud-based product, Coruson. During the last five years Coruson has evolved into a leader in the aviation sector thanks to projects with many leading airlines such as British Airways, Emirates and AirAsia.” His company has been working in the sector for around 20 years and enjoys a healthy client base all over the globe. “Safety, quality and risk compliance has always been a core principle of this sector and many industries look to aviation for best practice,” Cespedes added.

Standard Deviations

Ideagen’s safety management software is used in a wide range of safety-critical industries such as aviation, rail, manufacturing and food. Image credit: Ideagen.

Ideagen’s Cespedes also observed, “Aviation has always had access to numerous streams of Big Data to make significant leaps in safety performance. Aviation operations are now starting to implement the learnings from how marketing organizations analyze big data to benefit and improve their safety management system,” adding, “modern airlines are now so good at gaining safety data that their problem is having too much and not enough resources to manage or analyze it. This has resulted in the prominence of artificial intelligence, for example, to help them identify common threats, weaknesses on controls, emerging hazards, etc.” For its part, “Ideagen’s Research & Development department – based within its our headquarters in Nottingham [England] – is constantly working on themes and topics, such as artificial intelligence, as we look to further develop and evolve our suite of solutions.”

CAE’s Ranganathan emphasized CAE Rise can meet yet another emergent trend in the community – using Big Data generated in training for purposes beyond training, for safety management systems and related end uses. The CAE executive offered a use case, when a training leader tells CAE Rise to look for instances when an aircrew member in a scenario may deviate from a standard operating procedure sequence for managing a malfunction. “CAE Rise will look for deviations as opposed to following what the book says, record how often this happens, and wrap around the context of the situation when they do deviate,” he explained and noted the community has very little objective data in the flight data monitoring world to help explain how crews behave “when things go wrong.” And beyond that, having this data would allow the training organization to address this class of shortfall by developing a point-of-need training procedure correction, or bringing the problem to the attention of the operator or OEM.

CAE Rise has been delivered to four civil sector customers for initial, dedicated training requirements. “Quite a few more are collecting CAE Rise telemetry data and are determining how they want to use the data – they want to try it before they buy it,” Ranganathan added.

Matt Littrell, Product Director for AI & Adaptive Learning at FlightSafety, was asked about the launch customer’s interest in using FlightSmart beyond training. He said at this early point, the Air Force is taking a phased approach to the product’s use. He explained, “They have their assumptions of the value they will be able to derive and are looking at FlightSmart from a holistic perspective – beyond efficiency of training.”

When CAT met with the FlightSafety team, the company reported significant early interest in FlightSmart from an array of civil sector organizations: in primary training, from training organizations supporting ab initio training, universities, regional airlines and the major airlines. “They all have similar, but unique, requirements and value they see in FlightSmart,” Littrell recalled.

In one case, the needs of the training organizations and universities approximate those of the Air Force launch customer – to tailor training to the individual, as well as increase throughput and other program efficiencies. In another instance, regional airlines have an interest in FlightSmart for applicant screening and to meet training requirements for two disparate groups: newly minted pilots who have excellent academic credentials, and former pilots who are returning to aviation after being furloughed during the recession of 2008-11. And initial major carrier interest in FlightSmart has been generated by airlines’ desire to focus on evidence-based training and competency-based training and assessment. “They are looking for that objective evidence to improve the efficiencies of the training they are providing, be it AQP or EBT,” the FlightSafety director added.

Much like the US Air Force customer, initial queries about FlightSmart are training-focused. Littrell concluded unofficial discussions with prospective customers have included their “visions of where they can go. This could definitely include using Big Data to make more informed decisions not only for pilots, but maintainers, cabin crews, dispatchers and others – the list of additional opportunities is endless.”

Bert Sawyer, FlightSafety’s Director of Government Strategic Management, noted yet another prospective user group in the civil market is FlightSafety type rating and recurrent training customers for regional airlines and business aircraft. “FlightSmart has stirred up a lot of interest within the customer base. They are asking: when will I get my FlightSmart; and will I see this when I am doing my FlightSafety training?” He indicated FlightSafety’s strategy for civil rollout of FlightSmart will take shape “in early 2020.”

Ecosystem Perspective

The real story with Big Data’s evolution in this sector is its expanding use beyond training organizations and airlines – up to the OEM level and then through the broader civil aviation ecosystem.

The Skywise open data platform developed by Airbus and Palantir enables digital transformation to gain operational efficiency improvements, such as production ramp-up or aircraft in-service feedback. Image credit: Airbus.

Skywise is Airbus’s open data platform developed in partnership with Palantir. Airbus is primarily using Skywise for its own digital transformation needs to gain operational efficiency improvements, such as non-quality management, production ramp-up or aircraft in-service feedback. Frederic Sutter, Digital Transformation Leader at Airbus, noted these represent most of Skywise’s benefits for Airbus, as a foundation for investing in Big Data. In addition, Airbus uses these large data sets as a support to its safety network for further enhancing its risk management approach to select the priorities, then dig further on technical issues and solutions for them.

Indeed, in one user case, Sutter said, “It is estimated that Skywise directly contributed to the A350 production ramp-up in 2015-2016. Today, Airbus Engineering teams and design office have access to in-services operational data that are leveraged to improve their design models and ultimately will benefit all current and future Airbus aircraft programs.”

Skywise is following the upward trajectory of using large data sets. In just two years Skywise was reported by Airbus to have steadily proved its value with the OEM’s employees, then airlines, then suppliers and now with world-leading information management integrators (IBM, Capgemini, Sopra, Accenture, FPT). Sutter added, “With more than 100 airlines on board on Skywise, representing a fleet of more 9,500 aircraft from all major OEMs, Skywise is fast becoming the platform of reference for all major aviation players to improve their operational performance and business results, as well as to support their own digital transformation.”

The business case for investing in Skywise is significant, with other main, high-level outcomes being efficiencies, savings and enhanced flight operations achieved by reducing operational interruptions, burning less fuel, operating more efficiently, reducing workload and leaving room for third-party services. And of course, there is the underlying, compelling monetary value to embrace Skywise – with Airbus estimating that the costs of inefficiency in its industry along its entire value chain are worth $50 billion. Sutter explained this is mainly because of the lack of data continuity induced by organizational and operational siloes at every level in many companies. “This is why the Skywise program as an open aviation platform that airlines can use for their entire fleets (not only for Airbus aircraft) is a real revolution in our industry.”

Exponential Expansion

To be certain, the use of Big Data will continue to be the underpinning of programs throughout an aircraft’s life cycle… well beyond the training of any individual – from aircrews to maintainers to cabin crew and others – who “touches” an airframe.

Yannick Vanhecke, Head of Safety Enhancement at Airbus, emphasized: “The use of large data sets is part of our digital transformation and this will have a direct impact in all phases of an aircraft development, production, support and safety management. This has started by definition of means to further support the risk management of our safety process via text mining, machine learning, AI, but also on further enhancing identification and control of our safety threats.”