Intelligent moves

Tom Batchelor takes an in-depth look at the role of AI in airport baggage handling, from tag-free travel to downtime-avoiding algorithms

AI in baggage handling and the wider aviation sector has witnessed a surge in popularity in recent years
Beumer Group

The use of artificial intelligence (AI) in baggage handling systems (BHS) is fast becoming a gamechanger for the industry. At any given moment approximately 240 million kilos of baggage are up in the sky, criss-crossing countries and continents in the bellies of passenger aircraft. However, many of the existing BHS operate usingoutdated, costly and inefficient technology, say industry experts.

Barcoded tickets have been the norm for several decades, while some operators use more-expensive radiofrequency identification (RFID) tags, with electromagnetic fields to automatically identify and track bags – Hong Kong International Airport was the first airport to adopt RFID technology, in 2008. However, the most tangible improvements to BHS in the coming decade are likely to come from the more widespread introduction of AI, which relies on powerful algorithms to perform complex problem-solving tasks. Proponents say this technology, some of which is borrowed from the e-commerce sector, will streamline the experience for passengers, airlines and airports. AI applications within BHS offer improved security, enhanced passenger services and fewer lost bags, as well as helping deter the smuggling of contraband.

Impact of Resolution 753

A major regulatory shift for the baggage handling industry came in June 2018, when IATA Resolution 753 entered into force, requiring members to “track baggage at four key points in the baggage journey”: passenger handover to airline, loading to the aircraft, delivery to the transfer area and return to the passenger.

The new regulation also brought in a requirement for airlines to share the luggage tracking information with their interline partners – any other airlines operating services on part of a passenger’s journey. The aims of Resolution 753 were to reduce lost or delayed baggage, increase efficiency in baggage operations and ultimately provide a better passenger experience. But IATA Resolution 753 offered another benefit: facilitating the growth of data-hungry AI applications by requiring operators to collect and share standardised information. The use of emerging technology inevitably brings with it additional cybersecurity concerns, from hacking to data protection and incorrect application. However, experts say these risks can be addressed through careful monitoring, and that many of the systems in place do not capture personal passenger data. In addition, AI software often begins with small decisions and feedback loops that ensure the machine response is correct, allowing the technology to evolve and gradually make even better decisions.

BagsID uses an algorithm to classify baggage and then match it to a registered image set

“What’s seamless about a journey if we still have a prehistoric baggage handling method?”

Marlon van der Meer, BagsID

AI in baggage handling and the wider aviation sector has witnessed a surge in popularity in recent years, with a range of functions from advanced computer vision to complex data analytics. Here, we assess how the leading players in the market have approached AI and what BHS innovations operators can expect over the course of the next decade.

AI over bag tags

Automation and machine learning can play a role in the journey of each bag, from the passenger’s home to the airport apron, according to Marlon van der Meer, the founder of BagsID, a Dutch company using AI BHS.

“It's fascinating how much energy and money is put into what the industry calls seamless journeys,” van der Meer told Airports International. “But what's seamless about the journey if we still have a prehistoric baggage handling method, meaning you only put a tag on it. There is not even an image being taken, so we don't know what we are loading onto an aircraft. We cannot have seamless journeys if we don't crack this part of the problem.

“It starts with booking, then packing and people weighing their suitcases at home. If we don't do this right, then there is no need to focus on a completely seamless journey for passengers.”

BagsID, which alongside Vanderlande completed successful testing of AI BHS technology at Eindhoven Airport in 2021, uses an algorithm to classify baggage and then match it to a registered imageset in a library, sending back precise data including origin, type, colour, IATA classification, manufacturer and dimensions of the baggage –a much more detailed picture than traditional paper tagging can offer. If an item is damaged or its appearance alters along the journey, the BagsID system can still recognise and match it, and – crucially from an AI perspective – learn from the new data that it is presented with.

Van der Meer explained: “We use AI on multiple levels here. Based on our computer vision technology, we extract unique characteristics from an object, in this case a bag. Our machine learning and partial deep learning technology recognises a suitcase individually based on its unique characteristics, categorises it, classifies it and ranks it in a database. Right now we are extracting about 6,000 data points from a photo and feeding our systems so the algorithm is getting smarter by the day.

The BagsID system can even identify items of baggage damaged during handling at the airport
Baggage 360 from Siemens Logistics uses advanced machine learning and AI to anticipate luggage volumes
Siemens Logistics
Vanderlande is exploring the use of advanced vision AI and machine learning for predictive maintenance

“Then you can look at the pre-screening processes. We want to get to the point where we can see whether a bag has been externally modified or prepared for all kinds of security and smuggling prevention by comparing the cases with information from the manufacturer, for example. And then you might not even need so many x-ray machines any more, or you could at least do it a little bit smarter.”

The benefits to streaming airport baggage processing are evident, but what about the cost? Does the implementation of sophisticated AI technology risk shrinking profit margins or pushing up prices for passengers? Van der Meer is clear that the implementation of AI will save not only time but money. The current cost per printed bag tag is estimated at around €0.15, while BagsID claims to be able to offer ticketless luggage tracking for 10% of the cost.

AI-driven predictions

Siemens Logistics’ Baggage 360 uses advanced machine learning and AI to anticipate luggage volumes over a 24-hour period and to predict the flow of baggage across the airport. Major hubs in Europe, the Middle East and Asia are already using the equipment and additional trials are planned for 2022. The company also has predictive maintenance products, such as Sorter 360, Carousel 360 and Motor 360, which use condition monitoring and analytics to ensure a safe and reliable operation. Speaking to Airports International, Khaled Nabli, head of digital services at Siemens Logistics, said: “Machine learning algorithms allow our customers to determine the ideal time for servicing and avoid unexpected and unplanned downtimes – for higher system availability and reliability.”

Beumer Group offers more than a dozen baggage handling products, one of them seen here at Singapore’s Changi Airport
Beumer Group

In order to increase the spread of AI in BHS, operators will require access to even greater quantities of data. “We believe that airports, airlines and concessionaires can form powerful data ecosystems if they would be more open and collaborate to the benefit of customers and sustainability,” said Nabli. “We at Siemens Logistics support these initiatives by utilising open approaches, such as Aviation Data Hub and MindSphere, and adopting these global industry data standards in future opensource projects.”

The benefits for the industry are numerous. Operators can use the technology to better monetise elements of their business, such as dynamic fees for baggage, providing more appropriate levels of staffing or helping to avoid shutdown of the BHS due to jammed bags or equipment failures, which can cause delays to departures. AI technology also allows for continuous improvements in the system as the technology is able to 'learn' new features.

France’s Alstef Group uses AI technology in its self-service bag drop and check-in technology to screen bags prior to their entry to the back-of-house BHS, checking for irregularities (such as extended handles, loose straps, soft or cylindrical bags) that may cause issues when the bags travel through the system. The Alstef solution can determine whether a bag needs to be placed in a plastic box or even be directed through the oversize or fragile bag route. It can determine if a bag is already placed in a tub and is then able to deduct the weight of the tub at check-in, ensuring the correct weight is recorded. Alstef also provides a predictive maintenance solution that uses AI to learn and determine the optimal time for maintenance of BHS equipment to avoid costly system failures. The company’s self-service bag drop is in operation at Vancouver Airport, while its predictive maintenance AI system has been rolled out at Zagreb Airport, Croatia, and the new Felipe Ángeles International Airport in Mexico City, which is due to open this year.

Siemens’ machine learning and AI solutions can be found at major hubs across Europe, the Middle East and Asia
Siemens Logistics

“Airports, airlines, and concessionaires can form powerful data ecosystems”

Khaled Nabli, Siemens Logistics

Versatile technology

Germany’s Beumer Group, an intralogistics firm with more than a dozen baggage handling products, employs AI for a range of functions, including predictive maintenance and to monitor which part of the BHS needs inspection. AI is used to 'supervise' the system logs of its 24/7 Hotline engineering support team to identify abnormalities and automatically create a notification to the service team, including guidance to where in the log they should look for the issue.

Leonardo, the Italian aerospace specialist, provides a range of BHS solutions including its flagship product: the Multisort Baggage Handling System (MBHS). This, the multinational says, is the only cross-belt-sorting system suitable for baggage handling which is currently in operation at airports. The MBHS offers high-precision sorting and can handle all types of baggage, including fragile or irregular-shaped bags or those with highfriction or sticky surfaces. While Leonardo currently does not have AI products within its portfolio, it is looking at how it can deploy the technology, including in allocation and sortation. A spokesperson said the company was “moving towards the digitisation of our portfolio, which will naturally include improving and updating our software, but also embracing new technologies, one of which is, of course, AI.

Leonardo’s proprietary high level IT solution for baggage handling systems is called EasyAirport

“There are some obvious applications in allocation and sortation, and AI coupled with other technologies: vision systems, for example, could support the eradication of bag tags so that the equipment intelligently characterises a bag based on size, shape, colour and characteristics.”

Logistics automation company Vanderlande’s Fleet product uses AI to optimise the routing and handling of baggage from the check-in areas through to the apron. The company is also enhancing airport robotic loading solutions with AI to improve operational performance, such as increasing the fill rate of BHS and reducing load times. More specifically, Vanderlande is exploring the use of advanced vision AI, deep reinforcement learning and machine learning for predictive maintenance. By integrating the entire baggage information f low, the company aims to reduce the number of mishandled bags and improve the experience of the passengers.

In the UK, Aberdeen, Glasgow and Southampton airports will see improvements in security operations with the help of technology company Pangiam and Google Cloud's AI and machine learning computer vision tools. The three sites, all managed by AGS Airports, will use the technology to look for threats concealed within baggage and other shipments. Project Dartmouth, as it has been named, seeks to replicate “human intuition” by scanning for items that "just don’t look right", using Aggregated Threat Detection (ATD) software to look for complex or co-ordinated attempts to breach security, such as dismantled weapons.

Khaled Nabli of Siemens Logistics predicts an increase in on-demand BHS sortation capacity
Siemens Logistics
Shenzhen Bao’an is the biggest BHS delivered to date by Vanderlande in mainland China

More immediate changes

AI is already disrupting the aviation sector and industry leaders predict a major shift in BHS. The capability of AI systems is being rapidly extended, with improvements in decision-making accuracy enabling a wider range of automation opportunities. Improved technology will mean end-toend baggage delivery, luggage storage and on-demand travel items at destination to improve airport efficiency.

“New processes and players will innovate baggage and passenger journey modes and we will see more on-demand BHS sortation capacity such as those with autonomous vehicles,” explained Siemens Logistics’ Nabli. “Besides new technological trends, we see major shifts in how airports approach the business of baggage handling. The main drivers here are digitalisation and, currently, the pandemic. We see increased acceptance for new business models such as baggage handling as a service and data-driven value-added baggage services. Sustainability will play a major role, and the industry will strive to reduce the carbon footprint of travel and baggage journeys.”

Looking ahead

Asked what innovations the industry can expect over a five-to 10-year period, Beumer Group's data specialist, Per Engelbrechtsen, told Airports International: “Decision science will be the next big thing. When looking at BHS charts and graphs, the system will automatically provide the user with a recommendation to which decision to make. It might be that nothing should be done, but help and guidance is issued as to what to decide. In data analytics, decision science will be the helping hand to know what to do, for example, to address unwanted human behaviour, which is when BHS staff may have bad habits or are not adequately trained. With the quality of data improving all the time, the capabilities for optimising operations will keep improving.”

Vanderlande and Shenzhen Bao’an

Dutch baggage behemoth Vanderlande has installed a brand new BHS at the satellite terminal of Shenzhen Bao’an International Airport in China to facilitate the site’s expansion as it targets millions more passengers each year.

Vanderlande supplied the BHS for Shenzhen Airport’s Terminal 2B in 2004, then for Terminal 3 in 2013. The latest project, awarded in 2019 and completed in December 2021, saw the company’s technicians continuing to work throughout the pandemic. Shenzhen Bao’an was the biggest BHS delivered to date by Vanderlande in mainland China, with some 20km of conveyors across the system, including the TUBTRAX individual carrier system (ICS) in the satellite terminal and Terminal 3, the high-speed TUBTRAX (running at 7m/sec) ICS in the tunnel area, ten transfer check-in desks, plus conveyors and carousels. Shenzhen Bao’an now ranks as the third busiest airport in mainland China, with 37.9 million passengers in 2020, and the new BHS caters to the growing demand for flights from the region.

Doney Xu, executive MD for Vanderlande China, said the 29-month project was achieved “despite the challenging conditions and timescales” presented by COVID-19. The company has secured a four-year contract to deliver roundthe-clock maintenance, operations and IT services. Vanderlande also provided Qingdao Jiaodong International Airport in Jiaozhou, Qingdao City, China with its BHS, which opened last year, comprising 16km of conveyors stretching 12m underground to 12.8m above ground at the check-in area. The system includes a 1,500 bag-capacity storage facility and an inbound system with 18 arrival lines and 14 reclaim carousels, capable of handling 15,000 bags per hour.

Among Leonardo’s recent installations at Milan Malpenso Airport was a Multisorting Baggage Handling System (MBHS) cross-belt sorter

And the benefits AI can bring are by no means limited to BHS. From air traffic control to flight delay prediction, applications driven by artificial intelligence have the power to reshape the modern aviation landscape. For example, airports in Osaka, Japan, and Abu Dhabi have explored autonomous check-in facilities, while Seattle and Miami in the US use sensors to monitor security queues. New York City-based NLX has developed its Voice Compass software to verbally guide customers through tasks from password reset to rebooking flights and is already in use with Copa Airlines.

COVID-19 has disrupted the aviation industry like nothing before it, and with airports having adapted to reduced levels of staffing, the use of data-driven asset management and AI agents has increased. Driven by this new way of working, AI can assist airports in making better decisions by using data analysis to help them operate their BHS in the most efficient way. The pandemic has also placed an emphasis on seamless journeys where interaction between passengers and ground staff is minimised, making self-check-in, bag drops and boarding a more attractive proposition. In addition, airlines, particularly those in receipt of COVID-linked government aid, are under pressure to be more sustainable. The use of data-driven AI applications is proven to help in this endeavour, not least through the possible future eradication of paper tags. While the full impact of the pandemic remains uncertain, it is clear that when it comes to the application of AI within BHS, the future promises even more opportunities for the world’s leading technology and logistics companies to innovate. AI

Tackling the cyberthreat

Alstef Group was one of the partners responsible for the Security of Air Transport Infrastructure of Europe (SATIE) project that helped develop a toolkit to protect critical airport infrastructure from cybersecurity threats. This large-scale project brought together 18 partners from ten European countries with a common goal: improving the prevention and detection of cyberphysical attacks at airports.

The partnership worked specifically on the cybersecurity of the baggage handling system – critical infrastructure for any airport. The cybersecurity measures were tested through a series of simulated attacks on a digital twin of the baggage system at Zagreb Airport, including some of the most common threats such as a ransomware attack or a DDOS (distributed denial of service). The results were conclusive, as all of the attacks were detected, both in the simulation tests and in the demonstration.

The SATIE project has resulted in the creation of an interoperable defence toolkit that detects, prevents/mitigates and responds to airport threats, thereby protecting critical systems, sensitive data and passengers. The toolkit has led to innovative advances in securing critical airport infrastructures. These solutions offer pioneering opportunities, from threat prevention and detection to incident response within baggage handling systems.

“Today, it’s no longer a question of ‘if’, it’s a matter of ‘when’ a cyberattack will occur. It is therefore essential to know how to detect cyberthreats so that SOCs [security operation centres] can be alerted and act quickly,” explained Eric Hervé, the chief information security officer at Alstef Group. “The innovations and expertise developed through our involvement in the SATIE project can now be passed on to our airport and intralogistics customers and provide the assurance of thoroughly tested and proven cybersecurity solutions.”

Leonardo recently undertook to bring the baggage handling system at Milan Bergamo in line with ECAC Standard 3, including the MBHS machine