DECEMBER 2025AEROSPACEDEFENSEREVIEW.COM19essential part of this plan was the upgrade of the existing passenger forecasting model into machine learning models, with a high level of automation, allowing frequent reruns and the expansion of the scope of the forecasting team to several other processes, such as bags, bussing, PRM and many more."Stein Van Haver, Data Scientist at Airport Intelligence and responsible for the development of these forecasting models, continues, "The power of Artificial Intelligence is the combination of mathematics, data and technology. It is a technique that for very specific, repetitive tasks ­ due to the immense loads of data and calculation power ­ will perform very quickly and much better than a human. This makes it a very powerful technique to use as application in business processes. One of the best-known applications of AI is predictive analytics also known as Machine Learning, a technique that can predict what will happen in the future when it is exposed to a large set of data."In airports, this technique has proven to bring efficiency, speed and safety. Korijn explains how their machine learning passenger forecast is the base of decision making. "Our models are able to understand years and years of historical data and capture trends on the lowest level. The outcome is the amount of local and transfer passengers per flight for the (near) future which is a wealth of information to pro-actively optimise the airport capacity." These models give more weight to more recent data and can pick up any new situation very quickly. As a result, the technology remained reliable during the current pandemic and the outcome of the models became even more critical to make sound decisions on which infrastructure was needed and how many resources should be deployed in the most cost efficient manner. The passenger forecast at Brussels Airport is used to optimise staffing and equipment at different processes throughout the entire passenger journey. The historical passenger arrival flow patterns are combined with the passenger by flight forecast to know at what time which passenger load is to be expected at a certain process in a particular location of the airport. As such, the airport knows very accurately the expected load throughout the day at the screening process, which has for a long time been a bottleneck.As the passenger satisfaction is critical in this stage, Brussels Airport aims for waiting times that never exceed 10'. This is only possible because the passenger forecast is shared with the screening provider and Brussels Airport orders a fluctuating number of lanes and staffing depending on the traffic peaks. The results of the forecasting models are integrated in the overall Airport Operations Plan, putting all data from pre-OPS to post-OPS at the disposal of all airport stakeholders as the basis for pro-active decision making and continuous review of the processes. This overall approach on screening resulted in a cost reduction of 10 percent per year and an increase in spend per passenger of 1 percent.Artificial Intelligence as Part of a Way of WorkingBut these forecasting models, driven by AI, do not do the trick on their own. These models are part of a system and a way of working within the airport ecosystem. It needs a data and performance driven culture of the whole organisation. It needs people to understand the input and trust the output. It requires change... With the Operational Coordination Center founded in 2015, Brussels Airport moved towards pro-active and data-driven decision making already long before AI came in.Airport Intelligence strongly believes in this philosophy and implements this integrated cycle of forecasting, planning, monitoring, mitigating and performance review and improvements in every process review. The introduction of an AOP (Airport Operations Plan) is the base of data-driven decision making in an airport operation. Thanks to the AOP of Airport Intelligence all necessary data is available and a single version of the truth can be guaranteed. Airport Intelligence offers a service of monitoring and running the models for airports (insights-as-a-service). At Brussels Airport this is done directly in the Operational Coordination Center by a dedicated data driven team. This team is the bridge between AOP, the data scientists and operations. Airport Intelligence focuses on bringing their tools, knowledge and expertise to other airports in a very modular way. They realise that it brings cultural change and have the right knowledge to help airports in this trajectory and set up collaborative decision making and coordination. Korijn Defever
< Page 9 | Page 11 >