"Aerial Mathematics." Big data in the civil aviation world

Technology Big Data become an important foundation for the development of airlines

Introduction

According to analysts, 67% of aerospace companies implement projects based on Big Data, another 10% plan such projects. As for airlines, here the implementation of projects for February 2019 is announced by 44% of companies, and plans for such projects were announced by 25%.

These are the results of a study that FlightGlobal conducted in December 2017 regarding Big Data's role for aerospace businesses and airlines. Analysts also found out an opinion on the sharing of data on the condition of aircraft with manufacturers and companies engaged in repair and maintenance (TO)^[1]. The study involved 300 professionals from the aerospace and aviation industries. Most of them are confident that Big Data technologies can improve the operational reliability and efficiency of airlines.

Infographic with the results of the study

About half of the respondents replied that their companies use arrays of data on the state of aircraft, which helps them make more verified decisions. In the near future, the share of such companies will grow to 75%.

Sharing data with OEM/MRO is still problematic. However, 38% of airlines believe that such a model can provide them with significant business advantages.

According to a May 2018 review by Honeywell "Connected Aircraft"^[2], 47% of airlines surveyed plan to spend up to $1 million over the next year for each aircraft they operate. Most of these companies plan to meet the amount of 0.1 to 0.5 million dollars. However, in the five-year perspective, 38% of air carriers announced investments in the amount of $1-10 million for each aircraft.

Until February 2019, when airlines invested in technologies related to aviation (connected technologies), it was primarily about providing satellite communications and Wi-Fi. Now companies are ready to benefit from the data that they can obtain by using equipment directly on board aircraft. For example, such data can provide them with savings of 1% of fuel consumption, which is equivalent to $50,000 per plane per year, Honeywell analysts calculated.

Use of Big Data by foreign airlines

Big Data technologies are used to perform a number of tasks in the field of civil aviation. In this chapter, we will dwell in more detail on the main directions of using big data in aviation in a number of foreign countries. First of all, these are repair and maintenance, fuel economy, creation of digital twins, optimization of operating activities (including forecasting flight delays), formation of personal offers for passengers, etc.

Big Data technology is used by airlines around the world

Big Data and aircraft airworthiness maintenance

One of these priority areas in the near future will be maintenance (maintenance) and repair of ships. Thus, 88% of respondents to analytical studies expect that it is in this area that they will be able to get the maximum benefits from the use of technologies. Maintenance and repair are significantly ahead of all other areas in importance. Big data analysis and predictive repairs in aviation demonstrate their effectiveness and prove in practice that connected technologies work.

After maintenance and repair, respondents expect advantages from the introduction of piloting technologies related to the field, including optimization of fuel consumption and aircraft turnover time, as well as passenger service.

Thus, in the study "Sky High Economics: Evaluating the Economic Benefits of Connected Airline Operations"^[3], analysts note that connected aircraft can transmit data to the cloud or to ground servers, where this data can be analyzed using Big Data Analytics tools. Thanks to this, air carriers can, for example, identify malfunctions before they become major problems. The information obtained can be used to make more verified decisions and reduce expensive aircraft downtime (Aircraft on Ground).

In addition, with the advent of predictive modeling, it became possible to replace parts that are identified on the basis of the analysis as requiring replacement before they fail, namely during planned repair and maintenance work. All this helps to reduce costs, increases flight safety.

Digit Twins. What is it?

The theme of predictive (proactive) repairs is also closely related to the use of the so-called "digital twins." However, unlike, for example, the oil and gas industry, where CDs are already used by a number of large companies in practice, in the aviation industry this topic is still more discussed at the level of experts and analysts.

Aviation industry experts in 2019 began to actively promote the topic of using "digital twins": the management of the Swedish company IFS, a developer of software for corporate clients, including those from the aviation industry, said in April 2018 that one of the technological innovations that can help airlines ensure the efficient operation of ships while reducing maintenance and repair costs is the "digital twins"^[4]. Digital twins ― virtual replicas of physical assets that can demonstrate to engineers on the ground the operation of the engine while the aircraft is in the air. To make this possible, engineers install thousands of data collection points during the design and production phase of the engine. They are then used to create a digital model that monitors and monitors the engine in real time, providing the necessary information throughout its life cycle, such as temperature, pressure and air flow.

GE helped develop a digital twin for the aircraft's lander. The sensors were placed on parts of the chassis most prone to breakdowns. In real time, data such as pressure and temperature were transmitted to specialists, helping to predict malfunctions or remaining life. These data were compared to those of a digital twin that was subjected to similar loads.

According to IDC, companies investing in data centers are able to reduce the time required to implement critical work, including maintenance, by 30%. Analysts expect the technology to become more mature in 2019 and provide additional benefits for users.

Chat boats

Analyst company MarketsandMarkets published in May 2017 a forecast for the chatbot market ^[5]. According to analysts, the average annual growth rate of the market between 2016 and 2021 will be 35.2%. In monetary terms, the market will grow from $703.3 million in 2016 to $3.172 billion in 2021. The main goal of the introduction of chatbots in the aviation industry is the desire of companies to better understand the behavior of consumers of their services and goods.

In addition, Big Data solutions are beginning to be used in the aviation industry and for other tasks, for example, to ensure aviation safety based on the analysis of historical data arrays or, for example, which echoes the increase in the level of customer service, to optimize the processes associated with catering on board, and tracking luggage.

Examples of practical implementation

Let's consider examples of practical implementation of Big Data solutions by foreign air carriers. The main emphasis is placed on large-scale projects in recent years, as well as on those business cases that provide data on quantitative or qualitative effects from implementation. These business cases relate to the promising areas described above.

Big Data Repair and Maintenance

British low-cost airline EasyJet, an operator of Airbus A319 aircraft, plans to implement about 50 different algorithms for predictive maintenance and repair on board its ships. The company's technical management announced in mid-June 2018 that work on the entire fleet should be completed by the end of 2019^[6]

EasyJet plans to implement around 50 different algorithms for predictive maintenance and repair on board its vessels

To implement the forecast repair and maintenance model, two solutions will be used ― an on-board data exchange system for flight operations and maintenance (FOMAX) and off-port tools for analyzing aviation data Skywise. FOMAX, a server from Rockwell Collins, collects data on the maintenance and performance of the aircraft, automatically sends them to engineers and technicians. SkyWise, which runs on a cloud platform, provides analysis of various data, was developed jointly by Airbus and Palantir Technologies.

The FOMAX system retrieves all data from the on-board FDIMU (flight data interface management unit). FOMAX has the functionality of a 4G router: after the ship lands, all data using 4G Gatelink antennas is transferred to the Skywise analytical platform and analyzed by Airbus specialists. For analysis, Airbus specialists have independently developed special models that can predict the occurrence of system problems. After the analysis, its results are forwarded to EasyJet specialists who already independently decide on the need for predictive maintenance or repair. Using the information received, the air carrier's specialists can create algorithms that will make it possible to predict the occurrence of a particular problem on any of the aircraft in the future.

Aircraft of the A320 model with FOMAX are capable of collecting more than 24,000 parameters, that is, providing 100% collection of information from aircraft systems and components. Aircraft without FOMAX collect 400 parameters, that is, 2% of the available information.

Meanwhile, Delta Air Lines announced in April 2018 that it plans to adapt its maintenance system with predictive functionality for the latest generation of aircraft that generate significantly larger amounts of data than previous ships (the airline is preparing to accept regular aircraft models such as Airbus A350 and Bombardier CS100)^[7].

According to Delta management, the predictive maintenance program used helps the airline significantly reduce the number of outages: Over the past 12 months, the use of proactive maintenance has helped to avoid 1,200 flight delays or cancellations.

The program uses data Boeing coming from various systems, such as Aircraft Health Management from, for example, Airbus and GE systems. At the same time, the basis of the airline's fleet of aircraft is aircraft, which were developed even before the extraction and analysis of data became a "must-have" function. After analyzing the data obtained, the program makes recommendations for replacing parts and mechanisms. According to Delta Air Lines experts, the program used demonstrates a 95% level of accuracy in terms of recommendations for replacing parts.

Honeywell in May 2017, together with Cathay Pacific, tested the use of Big Data Analytics solutions on the aircraft of the Hong Kong air carrier^[8]. The tested solution ― GoDirect Maintenance Service ― is focused on the use of predictive analysis methods in terms of aircraft maintenance.

Cathay Pacific provided its A330 aircraft for testing the solution. Data from aircraft equipment was transmitted through existing channels to the airline's divisions, as well as to providers providing maintenance services. Thanks to this, the maintenance personnel on the ground could prepare in advance for potential maintenance work and also order spare parts in advance, without waiting for the onset of a critical situation.

During testing, it was possible to reduce the number of breakdowns of aircraft equipment by 35%, which made it possible to reduce maintenance costs, reduce the number of departure delays and aircraft downtime. The accuracy of predicting breakdowns was 99%, Honeywell management said.

In June 2017, the companies entered into an agreement to implement the GoDirect Connected Maintenance solution, which uses Big Data and sensors to predict faults, on all the airline's Airbus A330 aircraft. According to the contract, GoDirect Connected Maintenance systems will be installed on more than 60 Cathay Pacific aircraft. According to Honeywell, Cathay Pacific is considering using the system also on Boeing B777 aircraft.

Eastern neighbors are also not left out. In the fall of 2017, Korean Air began using artificial intelligence technologies to support specialists conducting^[9]: Korean Air uses Machine Learning technology to process large amounts of data generated by the company's ships. The solution relies on IBM's Watson platform based on artificial intelligence and is capable of reading past records regarding natural language maintenance, as well as processing data from the aircraft itself and technicians.

To solve such problems, Korean Air uses content analysis technology based on WatsonExplorer and Natural Language Understanding (NLU), which helps to process structured and unstructured data. This reduces the time required to determine the causes of a possible future malfunction or problem by 90% compared to conventional methods.

Examples of Fuel Cost and Emissions Reduction in the Transition to Big Data Technologies

In 2016, the American airline Southwest Airlines decided to test an analytical platform that reduced the second largest item of costs ― fuel costs, which range from $4 billion to $6 billion annually^[10]. Initially, forecasting fuel costs at Southwest Airlines was based on information from several systems, including Ariba, Allegro's fuel management environment, as well as the company's own corporate repository of historical data. All this data was then summarized into one bulky table. Specialists generated 1,200 forecasts of fuel demand every month. The financial analyst spent three days a month working on these forecasts, which, moreover, did not meet the company's requirements in terms of accuracy.

Southwest Airlines cuts fuel costs at the expense of Big Data

The pilot used the Alteryx Designer platform. She helped build 8 different predictive models that included a time-series regression modeling function and neural networks. For each month and each airport, the system could generate 9,600 forecasts. The processing time for making forecasts was reduced by 60%. In addition, the accuracy of the forecast was increased. The project helped Southwest Airlines figure out that acquiring fuel from a single supplier would provide greater benefits than several, as previously practiced. The speed of making forecasts was reduced from 3 days to 5 minutes.

Danish airline Thomas Cook Airlines Scandinavia, which provides charter services, chose in May 2017 the GoDirect Fuel Efficiency solution developed by Honeywell to improve fuel efficiency and reduce harmful emissions^[11]. The SW ensures that the EU emission standards are met.

Thomas Cook Airlines Scandinavia uses Honeywell developments to improve fuel efficiency and reduce harmful emissions

The comprehensive GoDirect solution, which includes a number of services and applications, provides operators, ship crews and repair teams with software for managing large amounts of data and analytical tools to improve the efficiency of the aircraft. The efficiency of this solution for fuel economy differs depending on aircraft and airline models, but existing customers talk about annual savings of up to 5%. For large airlines with large fleets of ships, this equates to savings of tens of millions of dollars annually.

As of February 2019, GoDirect Fuel Efficiency software is already used on all Thomas Cook ships.

Predicting flight delays with Big Data

In April 2018, the venture capital arm of JetBlue Airlines, an American low-cost airline, invested in Lumo (formerly FlightSayer) startup^[12]. According to management, a startup, the company is able to predict flight delays a few hours and even a few days before departure.

JetBlue Airlines predicts flight delays with Big Data

The company's solution is based on technologies in the field of artificial intelligence. Users can check the accuracy of the Lumo solution for free online. To do this, they must enter the flight date, flight number and airline name on the startup website. After that, users will receive a forecast indicating the duration of the possible delay. The use of the Lumo solution allows airlines and travel agencies to rebook passengers' tickets, thereby maintaining their loyalty, as well as preventing airport crises arising from flight delays.

Among the possible problems in the implementation of the solution, experts highlight, first of all, the presence of competing solutions from large players in the market, as well as insufficient integration with the systems of travel companies. As for competing solutions, experts call SITA FlightPredictor (in April 2017, SITA management announced the appearance of the starting version of the solution by the end of summer 2017); offered since 2016, the so-called Schedule Recovery System from Amadeus, the first user of which was Qantas; Google program (since November 2017, designed for passengers). A number of start-ups also offer similar programmes, such as Cambridge-based Freebird.

When presenting the decision, the startup indicated that in the spring of 2018 a pilot with a large international air carrier should be implemented. However, no additional information has been published on this topic.

Application of Big Data for analysis and forecasting of demand for air transportation

Southeast Asia can be considered a strong enough player in predictive demand analytics for air travel.

So, in December 2016, Philippine Airlines chose the PROS solution to optimize the Global Revenue Strategies^[13]. The PROS Origin & Destination cloud solution allows you to predict demand and optimize revenues taking into account market trends, changes in demand and proactive analysis.

Philippine Airlines Optimizes Revenue Based on Market Trends, Demand Changes, and Proactive Analysis

Philippine Airlines is a national carrier in the Philippines. As of February 2019, the company's aircraft fly to 30 local and 43 international destinations. The company operates 79 aircraft ― Boeing 777-300ER, Airbus A340, A330, A321, Bombardier Q400 and Q300.

Aviation technology developer Sabre Corporation, a developer of solutions for the global tourism industry, announced in November 2017 the signing of a multi-year agreement under which Hong Kong Airlines will receive the Big Data Solution MIDT (Market Information Data Tapes)^[14]. This solution is a database that provides access to historical and forecast (depth up to 11 months) booking data. The possession of this product allows airlines to analyze the impact of measures in the field of tariff formation, marketing programs.

Hong Kong Airlines Uses MIDT's Big Data Solution to Execute Marketing Strategy

Hong Kong Airlines expects to use the product when implementing plans to start work in the North American market. The product allows you to generate reports and has analytical functionality, which gives the user the opportunity to identify the optimal channels for the implementation of a marketing strategy. Hong Kong Airlines will have access to data from Sabre agents around the world for all departure and destination points where the airline is present.

Big Data to improve customer satisfaction and personalization

British Airways, which is part of the TOP 10 in terms of passenger traffic, has been using Big Data Analytics since 2013 to improve the level of service for its customers: The carrier collects various passenger data in a special storage facility, and then uploads it to a program called "Know Me"^[15]. The purpose of the program ― to learn and better understand the needs of customers, as well as to use the data accumulated during all kinds of contacts with these customers to improve their service level.

British Airwaysuse Big Data Analytics to boost its customer experience

"Know Me" contains various data on passengers: behavior during online orders, wishes when making purchases, preferences when choosing a seat. All this information is automatically generated and automatically used on the next booking made by the customer.

The program works using analytical software from Opera Solutions. Image Google Search is also used, which allows airline employees to recognize especially important and many flying passengers already at the moment when they enter the airport or business box, and, accordingly, offer them high-end service.

Another major market player, Virgin Australia in late 2017, revealed that it was optimising How machine learning^[16] boost^[17] frequent^[18]For these purposes, the company attracted the startup DataRobot. American the startup has developed a predictive analytics platform to quickly create and implement predictive models. This platform is already helping Virgin Australia reduce the time to create predictive models by 90%, while the accuracy of forecasting increases by 15%.

Virgin Australia is working to optimize its Velocity Frequent Flyer loyalty program with predictive analytics

As of February 2019, the airline is working on optimizing its Velocity Frequent Flyer loyalty program, introducing predictive analytics into it, which should support the company's customers when they choose the best time to use the points received. DataRobot is tasked with building forecasts/models of who is most likely to go on a trip, what price and what type of travel the traveler prefers. In general, we are talking about increasing the level of service for participants in the airline loyalty program.

Using smart chatbots

In September 2017, Finnish national airline Finnair launched an artificial intelligence-based chatbot for its Facebook account Messenger^[19]. The Finn chatbot is able to sell tickets, provide information on departure times and how much luggage can be taken on board. In addition, the bot can direct passengers to the Manage My Booking airline page, where they can purchase additional services. Also, the bot is able to answer the most frequently asked questions. If it is not possible to give an answer, it redirects the passenger to the client service staff.

Finnair launches AI-powered chatbot

Finn understands and communicates in English and Finnish. The company is considering using the bot on other platforms, such as Wechat in China. The chatbot was created in conjunction with Caravelo, a company that develops solutions for the aviation industry.

The AI-powered chatbot was also launched on Air New Zealand under the name Bravo Oscar Tango (Oscar) on the airline's website in New Zealand. In October 2017, it entered the overseas market - the Australian market ^[20]. The task of the chatbot is to answer FAQ questions from site visitors.

Smart chatbot also uses Air New Zealand

The launched beta version of the bot was the initial development of Air New Zealand in the field of artificial intelligence. Since launch, he has been trained on the basis of questions from site visitors. According to the airline's technology executive, on its first day of operation, the chatbot was able to successfully respond to 7% of requests. As of October 2017, this figure reached 67%, and by February 2018 ― 75%, while the chatbot continues the learning process. As the airlines found out, passengers mainly turn to the chatbot when they require quick answers to questions that arise on the day of departure, or about booking tickets. When booking tickets, the most popular topics are such as receiving a reservation confirmation, the permissible weight of luggage and air miles. In general, the chatbot is able to conduct conversations on more than 380 different topics. In August-October 2018, the chatbot held about 55,000 conversations. On average, he received 300-350 requests per day. On some days, the number of requests exceeded 1,000.

Before launching in Australia, the chatbot learned some expressions from Australian jargon and a number of sayings, Air New Zealand officials said. Along with answering questions, the chatbot can sing and tell jokes. It was recently launched on the Air New Zealand mobile app.

Another Australian airline, low-cost Jetstar, launched the virtual assistant "Jess" on Facebook Messenger^[21] February 2018^[22]The chatbot originally started working on the airline's webpage in 2013. In November 2017, testing of the solution began on Facebook. Jetstar became the first airline in the Asia-Pacific region to expand its chatbot from a Facebook Messenger webpage.

Jetstar launched a virtual assistant "Jess"

"Jess" answers customer questions regarding their bookings, provides information on luggage and places. It is available to customers from Australia, New Zealand and Asia.

The solution uses Nuance Enterprise's Natural Language Understanding technology, an advanced resolution technology that allows it to conduct real-time conversations. However, the chatbot is not able to initiate conversations or make suggestions.

The airline reported the results of the chatbot's work on the platform: Its performance (the share of questions to which the chatbot was able to give answers that satisfied customers) was 73%. Answer times for a number of questions have dropped from 17 hours to 0 minutes, said Jetstar's head of customer services. As for the overall performance of the chatbot on two platforms, it is involved in 250,000 dialogues every month, and since its launch, it has participated in 9 million dialogues with Jetstar clients.

In April 2018, at the Passenger Experience Conference, the English company Aviget demonstrated a chatbot with advanced^[23] functionality]. Using machine learning technology, the solution allows you to interact with passengers based on various platforms such as Facebook Messenger, Viber and WeChat. The solution uses Natural Language Processing to recognize quite complex complex phrases, for example, "find me a flight from Heathrow for under $150 that arrivals in Jakarta on Boxing Day."

Providing Aviation Security with Artificial Intelligence and Big Data

In 2018, British air carrier Virgin Atlantic announced that it plans to introduce programs of the British company Osprey^[24]. We are talking about the Flight Risk Assessment System, which uses artificial intelligence (AI) and machine learning ( Machine Learning) technologies to process large amounts of data in order to improve the efficiency and reliability of aviation operations.

Virgin Atlantic applies artificial intelligence and machine learning technologies to process large amounts of data to improve the efficiency and reliability of aviation operations

The system collects data from more than 200,000 sources in 60 different languages. Then this data is transmitted to the database, which contains information on 380,000 events in the field of aviation safety and reliability. Data is collected from open sources such as messages in, MEDIA from social networks , and from industry websites. Osprey provides rapid threat response and long-term data analysis to identifications address the causes of these threats. The purpose of the system ― to provide participants in the aviation industry with information necessary for making correct decisions regarding flight safety. With the ability to analyze historical events using machine learning and artificial intelligence, the system is able to predict the likelihood of where and when they can happen again.

The Osprey system divides the globe into blocks not bounded by country boundaries or Flight Information Region boundaries. These blocks are dynamically updated. Other datasets, such as terrain height and population density, may be added to the flight information in these blocks. This allows you to identify existing or emerging trends.

Analyze and evaluate marketing initiatives with Big Data solutions

In May 2018, analytical solutions developer Neustar, Inc. reported that Scandinavian Airlines selected its product to evaluate and measure the impact of marketing on^[25] to^[26]. The solution, called Neustar MarketShare, will provide the carrier with analytics that will allow daily planning of marketing activity and support in making marketing decisions.

SAS (Scandinavian Airlines) selected Neustar product to evaluate and measure marketing impact on sales

Marketing The Marketing Mix Modeling (MMM) approach, implemented by Neustar in its solutions, allows airlines to analyze key sales channels and major regions to choose the optimal model for the implementation of their products. In addition, in order to give Scandinavian Airlines a holistic view of their activities in the media and the impact of various drivers on the company's economy, the Neustar MarketShare solution analyzes various resources - MEDIA, non-media resources, data on economics, competition, seasonality.

Big Data and other civil aviation industries

Big Data technologies find their application in other areas of civil aviation.

Developers of Spafax, a provider of entertainment solutions for airlines, created in June 2018 basic working prototypes of artificial intelligence-based solutions for their integration into an onboard entertainment platform with personalization functionality, which is used by a number of airlines, including American Airlines, Lufthansa and SWISS^[27].

The first solution ― a chatbot model that is close to human communication. At the same time, an application based on machine learning called LUIS (Language Understanding Intelligence Service) was used to improve dialog capabilities. In addition, cognitive services are integrated into the chatbot, in particular face recognition. Thanks to this, airline customers will be able to request a list of films in which a certain actor plays for viewing on board. To do this, you only need to upload a photo of this actor to the application.

The second solution ― an artificial intelligence-based application for analyzing video content using machine learning. The platform was able to identify certain objects, scenarios or content with age restrictions, which is often required in accordance with airline content requirements. For example, artificial intelligence is able to detect content related to scenes of plane crashes or adult content and filter it out.

In April 2018, FoxTripper first demonstrated a program with a "roaming card"^[28]. The program provides passengers with information about the places over which the plane flies, and allows passengers to make reservations at their destinations. The data collected in flight in combination with the airline's passenger data allows you to make forecasts regarding the total of which products and services are relevant to him.

Another interesting example ― Gogo Air. This in-flight infotainment company uses artificial intelligence and machine learning to help airlines improve How^[29]. Gogo Air uses the Adobe Analytics series toolkit, including Virtual Analyst ― a machine learning-based tool, to gather customer information for a number of major airlines.

By providing entertainment content and Wi-Fi access in flight, Gogo Air collects information about passengers using these services. This information is then processed and analyzed. As a result, airlines receive the data that helps them improve customer service and, often, offer their passengers more targeted products. Airlines will find out what products their customers may be interested in during the flight, what devices they use in flight, how much time they are ready to spend on the Internet or what entertainment they prefer on the plane.

Airlines use the received data to personalize services based on the situation context, for example, adapting the screens of infotainment systems in the aircraft for the client, depending on the flight length used by the passenger of the devices, destination.

Food management technologies on board also do not stand aside. So, in April 2018 in Hamburg, the company Black Swan Data, which develops solutions for data analysis, entered into a cooperation agreement^[30]. The goal of the collaboration ― to analyze passenger data and trends on social networks to predict which menu on the plane passengers will choose. Passengers will be able to order and expect to receive their favorite dishes after boarding. The pilot project of the two companies showed good results: It was possible to reduce waste for food by 50% and increase productivity by 15%.

In May 2018, ― aviation solutions developer SITA proposed a baggage tracking and management system. The BagJourney technology she developed allows you to manage luggage operations more and more SITA's^[31]. In the first six months of 2018 alone, more than 20 carriers chose this solution. SITA BagJourney ― one of the main solutions that helps the aviation industry comply with IATA Resolution 753, which spells out the requirement to track luggage at each stage of the journey.

The SITA BagJourney solution is used each year to handle hundreds of millions of luggage items. According to users, the solution reduces the number of errors by 30%. BagJourney is compatible with various hardware, including mobile scanning devices or stationary devices.

According to BahamasAir, one of the users of the solution, after its implementation, within 7 days it was possible to carry out the process of complete transition to mobile devices to track all baggage in the two most loaded areas in terms of baggage ― Nassau and Maya. According to the results of six months, the number of complaints about problems with luggage in the busiest direction decreased by 60%. The airline plans to implement the solution in all directions and expects that by the end of the year it will fully comply with the requirements of resolution 753.

Interviews with experts

Kirill Bogdanov, Aeroflot: We plan to continue the forced development of information technologies

big data He told TAdviser about the possibilities of analysis technologies for aviation growth and the project to create a platform for processing customer requests, built using Big Data, in an interview with TAdviser. Kirill Bogdanov CIO Aeroflot

Dmitry Bulenkov, "RAMAX": Civil aviation is the leader of digitalization in many countries of the world

Dmitry Bulenkov, Vice President for Sales of RAMAX Group of Companies, in an interview with TAdviser spoke about the use of Big Data technologies in civil aviation and projects carried out by the company in this area.

Application of Big Data technologies in Russian civil aviation

As of February 2019, Big Data is already actively used in Russian airlines. Aeroflot, S7 and other major market players work with big data. The first developments in the field of Big Data appeared in the domestic segment in the mid-2000s, they are associated with the creation of customer service services, ticket booking and aircraft repair.

Big Data technologies find use in Russian airlines

In this chapter, we will dwell in more detail on the market review of domestic developers and system integrators in the field of civil aviation, highlight projects based on Big Data and tell you about the most remarkable of them. Classic vendors working in the aviation segment ― Sabre, Lufthansa Systems, SITA, INFORM, Amadeus. These companies are developing specialized solutions for the industry.

At the same time, vendors such as IBM and Oracle also work in the field of development related to big data. Russian vendors working with Big Data in the aviation segment are VK (formerly Mail.ru Group) with the Tarantool Platform in-memory computing, Innodata with the Kribrum technologies ^[32] and Hadoop^[33].

Also in Russia there are quite a few system integrators involved in implementations in the aviation industry. Among them:

• RAMAX Group of Companies (25 years in the market, a technological consortium of system integrators and development companies covering the entire range of customer needs - from strategy development to support of complex solutions);
• Integro Technologies (a dynamically developing system integrator that provides services in the development, integrated implementation and support of specialized IT solutions, as well as software localization. The company offers a wide range of comprehensive solutions for implementing large-scale projects for the transport sector, in particular for airlines, airports and ground services. Included in RAMAX Group of Companies);
• GC "Technoserv";
• Gazpromneft-Avia;
• SAP CIS.

Russian airlines Aeroflot, Russia and S7 are actively introducing Big Data technologies. Next, we will touch upon the most promising and innovative projects based on this technology.

Automated airworthiness maintenance system using Big Data technologies

In 2017, Aeroflot put into commercial operation an automated system for maintaining airworthiness, maintenance and repair of aircraft on the AMOS platform from the Swiss vendor AMOS Swiss Aviation. Thanks to this project, it became possible to track the real picture of the state of air transport. Data was consolidated and transformed from a huge number of sources, including informal data; Integration of warehouse stock maintenance systems is implemented involvement of many structural subdivisions; increased requirements for translation quality for design needs are met.

Aeroflot introduced an automated system for maintaining airworthiness, maintenance and repair of aircraft

The use of AMOS significantly reduced the cost of maintenance of the aircraft due to indicators such as an increase in labor productivity of employees of the relevant departments by 4%, a reduction in warehouse stocks and optimization of planning and material support for MR. Improve product planning and accounting by up to 4% by integrating AMOS with SAP production systems and providing high quality data.

It is also necessary to note the reduction of labor costs when working with the system and ensuring the automatic formation of work cards for vehicle maintenance due to the loading of aircraft documentation, a completely different landscape and the organization of a reliable data backup system made it possible to significantly increase the fault tolerance of the solution, improve the quality and efficiency of accounting and management reporting to be provided to aviation regulators.

author '= Kirill Bogdanov, CIO PJSC[34] " This project allowed us to have at hand a real picture of the state of airworthiness and maintenance and repair of Aeroflot aircraft. With the help of the AMOS system, we got a global picture of all M&R processes in a single integrated system: procurement planning, repair planning and execution, online integration with Boeing and Airbus systems. This made it possible to reduce labor intensity, improve the quality of MR processes and maintain the airworthiness of aircraft. Our immediate goal is to reduce aircraft downtime on the ground through the transition to predictive maintenance. We will launch the implementation of this task in 2018.

A platform for analyzing and processing big data from social networks

In 2017, the RAMAX Group of Companies represented by Integro Technologies introduced a platform for analyzing and processing passenger requests on social networks using IBM BigInsights and IBM PureData platforms.

Integro Technologies has introduced a platform for processing passenger requests in social networks in Aeroflot

Working with a client reputation is of great importance for transport companies, including aviation. Social networks allow you to collect in real time feedback from passengers and respond quickly to them.

The advantages of this system ― the possibility of continuous monitoring of company satisfaction and interaction with users on social networks; security and identification of terrorist organizations, extremism and other problems; continuous improvement of the offer for the customer through analysis of big data in social networks and the ability to communicate with the operator directly; maintaining the reputation of the airline through prompt contact with the audience on social networks; analysis of user preferences and preparation of individual product offers, as well as successful targeted advertising. More information about the project is here.

Big data technologies are the basis for all current and future changes in the customer experience. Thanks to Big Data, Aeroflot PJSC has additional opportunities to conduct efficient business.

Domestic DBMS Tarantool in the big data analytics project

Aeroflot has introduced predictive big data analytics algorithms as part of a platform project for analyzing and processing passenger requests on social networks. As a DBMS, the domestic development ― the Tarantool solution from Mail.Ru Group was used.

Aeroflot used Tarantool DBMS in big data analytics project

The complex consists of a large number of modules that cover both functional business requirements and integration modules into the existing IT infrastructure of Aeroflot PJSC and various channels for receiving calls (social networks, e-mail, official website, personal account).

The first module is responsible for identifying the client based on a comprehensive analysis of the data, both the text itself and the author's profile data. The number of requests can reach several thousand per day.

The second module is designed to search for duplicate cases. Copy text to be placed on different resources or sent by mail. Semantically similar posts are defined to identify clusters that are incidents. Processing several such posts at once leads to a significant reduction in the load of responsible employees.

The third module "InfoGuides" is one of the most important in the system. Its main feature ― to predictively identify dangerous posts even before the growth of activity begins. Thus, the built-in algorithms indicate a potential "info bomb" and make it possible to offset reputational losses.

The estimated number of data accesses was several thousand requests per second with the required response of a couple of milliseconds. To meet the high requirements of the customer, such as the prescribed time limit of three seconds for enriching circulation with various properties, the use of high-tech software was required. Based on the results of the tests on performance, data storage quality and functionality, it was decided to use domestic development ― the Tarantool DBMS.

Tarantool is used in the Platform as an operational database, in which cases are stored in the form of special data structures necessary for analytics algorithms. Extremely high performance and the presence in the database of such properties as secondary indices and support for a large number of connections without loss of performance made it possible to successfully implement the above-described functional modules without exceeding the set time frame.

The use of domestic developments in such a large company as Aeroflot is extremely important. Russian software is often in no way inferior, and, as in our case, surpasses foreign counterparts. That is why Tarantool was chosen. And, of course, an important import substitution factor is being fulfilled, which for our company is one of the key priorities for the coming years.

As a result of the implementation, the Customer significantly increased the efficiency of the process of processing complaints and customer requests by responsible employees of Aeroflot PJSC using the Platform, radically reduced the delivery time of the appeal and the time for processing/resolving the issue due to the mechanisms for enriching the appeal with context, tone, topics (tagging), author's profile, etc. All this is aimed at achieving a positive economic and reputational effect at almost all stages of providing services to Aeroflot PJSC. Based on successful experience, all project participants will continue to use Tarantool software in their projects and strengthen partnerships^[35][36].

Big Data Analysis for Customer Segmentation

In 2018, the system integrator Technoserv implemented a project to create an intelligent customer segmentation system for Aeroflot on the IBM platform. The system, using Big Data analysis and machine learning models, segments customers by multiple characteristics. The IT solution helped the airline implement a marketing strategy, actively develop online sales, as well as increase the number of passengers.

Technoserv has implemented an intelligent customer segmentation system for Aeroflot

Technoserv confirmed that Big Data technologies are generally in demand in the transport industry, and this is confirmed by an increase in the number of projects using these technologies. At the same time, the topics of the projects, according to her, are completely different. These are the tasks of increasing the personalization of communications with customers, proactive equipment repair, demand prediction and other tasks solved using machine learning algorithms and analyzing large volumes of structured, unstructured and semi-structured data^[37]

Conducting calculations on the blockchain

In July 2017, Alfa-Bank launched a blockchain project based on the Ethereum platform to conduct settlements between S7 and its ticket agents. At the same time, Alfa-Bank acted as a settlement bank. Through such interaction, it was possible to solve the problem of distrust in the event of scaling the system and connecting other banks or airlines to this platform. The implementation of the platform opened up opportunities for significant optimization of business processes for both the airline and its partners. The speed of calculations increased from 14 days to 23 seconds.

Alfa-Bank launched a blockchain project based on the Ethereum platform to conduct settlements between S7 and its ticket agents

author '= Pavel Voronin, Deputy General Director for Information Technology of S7 Group (S7 Group of Companies[38] ' We carried out a deal to buy an air ticket through an open blockchain api to the bank, but I am sure that such a scheme will very soon be used by many companies around the world. The blockchain platform allows you to significantly optimize business processes. It automates any settlement scheme, even a very complex one - for example, warehouse deliveries. With such a mechanism, the participation of a person is practically not required: there is no need to issue invoices, carry out reconciliations, write acts. Potentially, suppliers of onboard power, fuel, airport services can be connected to the platform - all those companies with which S7 Airlines constantly works and not only.

Air refueling on the blockchain

In August 2018, Gazpromneft-Aero, operator of the jet fuel business Gazprom Neft, and S7 Airlines developed and implemented joint smart contracts (Aviation fuel smart contracts, AFSC) based on blockchain. The project allowed to automate the planning and accounting of fuel supplies and is designed to increase the speed of mutual settlements when refueling aircraft.

Gazpromneft-Aero and S7 Airlines have implemented joint smart contracts based on blockchain.

According to the statement of representatives of Gazprom Neft, this is the first experience for the Russian aviation market in using distributed register technologies. With their help, the airline was able to instantly pay for fuel directly when refueling on planes without prepayment, bank guarantees and financial risks for the participants in the transaction. This approach increases the efficiency of financial operations and reduces labor costs, according to the oil and gas company.

Predicting breakdowns of S7 Airlines planes with machine learning and big data analysis

In early March 2018, S7 Airlines developed a predictive maintenance system. According to the company itself, it became the first Russian air carrier to complete the development of such a system.

S7 Airlines has developed a predictive maintenance system

Initially, it is used for Airbus A319 aircraft. In the future, the system will be adapted for the entire fleet of aircraft.

The predictive maintenance system involves the analysis of an array of historical data on aircraft maintenance and the operation of individual components.

Software for data analysis and mathematical model building was developed by S7 Airlines specialists together with the Russian company Datadvance, specializing in the development of solutions for predictive analytics.

In March 2018, an array of data for the period from 2012 to 2017 was already available for analysis. These are data recorded in aircraft telemetry systems, the database of the S7 Technics aircraft maintenance and repair holding and meteorological data.

The main tasks that the company expects to solve with the help of predictive maintenance are reducing the number of sorties delayed for technical reasons, improving flight safety and the effectiveness of ship maintenance, predicting the likelihood of possible breakdowns for each aircraft in the company's fleet.

Equipping Pobeda aircraft with RFID tags

Pobeda Airlines has implemented the world's first project for equipping aircraft with RFID tags, within the framework of which radiometers are installed on all emergency rescue equipment of each side, read using a tablet computer.

Pobeda Airlines equipped aircraft inventory with RFID tags

Several hundred RFID tags in each of the planes are attached to literally everything that is not pinned down - from life jackets to seat belts. Also, labels are attached to heat-resistant gloves, megaphones, oxygen cylinders, masks, fire extinguishers, etc.

The goal of the project is to speed up the inventory of emergency rescue equipment that occurs after each flight. One of the flight attendants launches a special application on the tablet and passes through the cabin, scanning RFID tags. Each detected tag is answered with a short beep, and at the end the application generates a report on the availability of all emergency and rescue equipment. The report is immediately uploaded to the server: SIM-cards are installed in tablets, and the cloud part is implemented on the basis of Microsoft Azure.

If there is no equipment, this is immediately visible in the report, respectively, in this case, a command is not given to depart apron buses with passengers and they are checked.

Without equipment, the aircraft cannot be allowed on the next flight (that is, if there is not enough life jacket on board, it means that one of the passengers will be denied transportation). Manual inventory takes much more time and effort: only vests under the chairs - 189 pieces, and they need to be checked. Thus, thanks to the Victory RFID technology, it was possible to reduce the minimum turnover time of the aircraft from 30 to 25 minutes. This is one of the key KPIs in passenger aviation: the sense is that the less time it takes from arrival at the airport to departure on the next flight, the higher the airline's economic efficiency, since the plane generates income only when it flies, and does not stand on the ground. With the size of the Pobeda fleet of one and a half dozen aircraft, reducing the inventory time of each side by 5 minutes makes it possible to perform at least one additional flight without increasing the fleet of aircraft.

Establishment of a civil aviation innovation center to strengthen Big Data expertise

In 2017, Innodata, a Russian software developer in the field of innovative technologies, and Innopolis, a Russian IT university, created the Innovation Center in Civil Aviation (TSIGA). The purpose of the association ― the development of a technological and digital presence in modern civil aviation, promoting the disclosure of the essence and significance of modern technologies that affect the demand and supply for players in the aviation industry, and integrating the innovations of the digital world into current civil aviation technologies. In 2018, RAMAX Group became a partner of the Center in order to strengthen the existing expertise in the field of Big Data technologies, as well as the development of specialized solutions for the aviation industry.

Innodata and Innopolis IT University have created a center for innovation in the field of civil aviation to strengthen expertise in the field of Big Data

The main areas of activity are the development of existing and the creation of new solutions for the aviation industry, respectively. The center conducts both educational activities and project activities, be it the implementation of projects in the scientific and technical, innovative or information and analytical plane. CIGA is also open to pilot projects to promote advanced technologies and solutions and is ready to support development.

The basis of the economy of the city of Innopolis itself ― high-tech industries, therefore, it is important for us to actively participate in such a association, one of the tasks of which is to convey the importance of modern technologies and their active introduction into the aviation industry.

The Center works in the field of solving such problems as the development and implementation of augmented and virtual reality technologies to combat aerophobia, navigation at the airport based on virtual reality technologies, behavioral analysis of employee activities in the information field, prediction of the purchasing power of passengers and the formation of dynamic recommendations for changing the cost of tickets, flight schedule planning and seasonal schedule optimization analysis, predictive passenger traffic management, personnel management at airports, development of a system for selecting personal package offers of airline and partner services, and aircraft surface scanning techniques during post-flight maintenance, analysis of the runway, management of the overbooking level, analysis of the interests of passengers and formation of proposals for them.^[39][40][41]

Conclusion

The examples discussed above show that airlines ― no longer just aircraft, carriers that we managed to get used to. An important basis for their development ― big data technologies that make it possible, for example, to personalize services. Individual offers that make each passenger's trip as comfortable as possible. Search for travel information, order tickets, search queries - any actions on the network leave digital traces that can be analyzed to form the most targeted package of services. In addition, working with big data can increase customer loyalty, for example, by quickly responding to passenger requests.

Even more data is generated by production systems. Aircraft, rail locomotives and trains are the source of a huge stream of technical data that comes from sensors installed in engines and life support systems. Detailed analysis of this data allows you to identify and predict the need for repair of a particular part. Thus, big data allows you to increase the level of security, as well as save significant money for carriers. The required repair time is reduced and the aircraft can be used for its intended purpose for a longer period of time.

The proposed material touched upon some of the possibilities and practical results of using Big Data technologies in the aviation industry, in reality, there are more and more such opportunities for development every day.

Notes