The aviation industry is quickly recognising the value that advanced technologies such as artificial intelligence (AI) and machine learning (ML) can bring to the industry as a way to optimise ground handling operations. INFORM’s comprehensive software suite GroundStar incorporates both of these technologies to deliver valuable benefits to ground handlers. In an interview, INFORM Aviation chief technology officer Ekkehart Vette discussed how AI and ML are driving significant improvements in operational productivity, operational efficiency, Service Level Agreement (SLA) performance, and passenger satisfaction.

Why are AI and machine learning so critical for today’s aviation industry?

Ekkehart stated that: “All members of the aviation industry can no longer afford to be reactive, but must be proactive in addressing and effectively managing various disruptions which are becoming more and more frequent. Striking a balance between an ordinary day of operation needs with ad-hoc resource demand prompted by an unexpected contingency requires decision makers to have access to sophisticated planning tools that support optimised, real-time decisions. The days of relying on gut instincts yesterday’s decisions no longer suffice.

“AI plays a vital role in helping to leverage data and extracting data insights to support enhanced decision making and increased productivity. ML techniques predict missing data, for example, expected passenger volume. Working together, these two technologies enable complex situations and disruptions to be analysed, followed by recommendations generated from historical decisions and simulations on how best to manage a particular disruption.

“Additionally, the know-how-driven approach of AI leverages human expertise, along with data-driven methods such as Operations Research and mathematical optimisation. When combined with ML, these powerful technologies create a Hybrid AI concept which is incorporated in GroundStar. Using Hybrid AI, we can achieve the best results and deliver maximum advantages to our customers.”

Are you saying that Hybrid AI goes beyond machine-driven intelligence?

Ekkehart: “Traditionally, optimisation was based on algorithms developed by humans. Instead, AI and Machine Learning do not require any human intervention, and automatically learn and develop algorithms that are based on the outcome of data analytics.

“GroundStar uses know-how-driven Operations Research, mathematical optimisation, Machine Learning, and predictive analytics.”

Can you cite the benefits and operational results of Hybrid AI?

Ekkehart: “Hybrid AI facilitates efficient staff scheduling by automatically learning from the past and taking into consideration prior, similar scenarios. It plays a vital role in managing unexpected disruptions such as weather-related, staff shortages, or ground support equipment (GSE) problems. Subsequently, the number of stand conflicts can be significantly reduced, along with ad hoc changes on the day of operations, by assessing predicted flight delays and generating a robust stand plan. Among the benefits derived are better predictions of passenger numbers, and the maximisation of passengers on pier positions, which support improved customer satisfaction and saves on resources such as passenger buses. Additionally, Hybrid AI enables cargo airlines to identify optimum plans and real-time strategies to maximise efficiency, customer SLAs, and sustainable goals concurrently. An overall benefit of AI and the repeated, automatic learning is the increased quality of business rules and the substantial reduction in workload related to manual maintenance of these rules.”

What is included in the GroundStar Machine Learning Service?

Speaking on GroundStar and ML, Ekkehart said “this service is integrated directly into the GroundStar landscape and incorporates the following components: data processing, model generation, monitoring, continuous learning, and model hosting in the Cloud. Also included are a set of standard ML models and use cases. The current GroundStar version also includes passenger and delay prediction. Customer-specific models can also be provided as a service.”

How does AI work in GroundStar?

Ekkehart: “GroundStar uses operational data to make predictions. Flight, task, turnaround, load, and shift data is captured from the central GroundStar RealTime (GS RealTime) systems historical databased. These GS RealTime systems include GS RealTime Staff and Equipment, GS RealTime Stands and Terminal Resources, GS Connect, and GS TurnManager. All of the data collected by GroundStar is combined with external systems data such as weather data. The GroundStar Machine Learning Service provides ML Ops and the complete ML pipeline (i.e., data processing, model generation, monitoring, continuous learning, and model hosting in the Cloud) for continual learning. To ensure sustained accuracy, the models are continuously retrained. GroundStar modules are able to request predictions and apply them for strategic and tactical planning in GS Planning, for example, or pre-planning in GS RealTime.

Why is it important for GroundStar users to choose the GroundStar Machine Learning Service?

Ekkehart: “This service is integrated into the existing GroundStar suite as a new standard component. Accessing its predictive results directly in GroundStar requires no additional or external software or interfaces.”

Can GroundStar Machine Learning also support sustainability goals?

Regarding GroundStar’s potential for sustainability, Ekkehart said “Absolutely. For example, mixed GSE fleets, which are now commonplace in all airports, can be more sustainably managed. Applying the service’s advanced AI and predictive analytics, airlines, airports, and ground handlers can achieve optimised planning of their mixed GSE fleets in order to reduce driver-based vehicles’ and autonomous equipment’s fuel consumption and related CO₂ emissions, at the same time they are achieving operational stability.”

Please provide a preview of future AI developments within the GroundStar suite?

In terms of future developments for AI and GoldStar, Ekkehart commented: “Right now, we are working on using AI to enhance our optimiser functionality, using reinforcement learning. Our goal is to develop a self-learning, self-improving optimiser based on actions and key performance indicators (KPIs) derived from these actions. In addition, we are looking to apply AI-driven complex computer vision within GroundStar, which would be used in people and object counting, as well as predictions of turnaround milestones.

“Given the many disruptions the industry faces, ranging from labour shortages and rising inflation, to high fuel costs and supply chain challenges for cargo airlines, it is more important than ever that aviation remains focused on better disruption management. With that said, the aviation industry must also be striving to meet net-zero sustainability goals, maximum resource management, and an enhanced passenger experience.

In summary, Ekkehard stated that “Hybrid AI is the leading technology driving these objectives’ fulfilment. We at INFORM are ready to continue supporting the aviation industry by integrating this powerful concept into our GroundStar suite and looking to the future to further leveraging Hybrid AI.”