Smart Gates Suite

Flight disruptions i.e. variation in actual versus scheduled times for arrival or departure have a domino effect across the aviation ecosystem. As per the US Department of Transportation, disruptions caused by aspects within the airline’s control – aircraft cleaning, baggage loading, crew problems, fueling, etc. are the second most common reason for hold-ups after aircraft arriving late. In 2016 alone, unexpected delays cost the global airline industry nearly USD 25 billion.    

What if there was a way to minimize disruptions by improving the predictability of gate operations? Imagine a smart tool that analyzes historical data, identify patterns and improve predictability using machine learning tools. The cost and time savings would be immense, not to mention the significant impact on efficiency.

Solution Overview

ADB SAFEGATE is working on a suite of applications that combine our Airport Systems offerings with our advanced analytics capabilities, study historical data using machine learning, to accurately and predict arrivals and departures in a reliable manner.

We are currently focused on the following areas of application - taxi time, crew allocation, and disruption management and recovery.

Our value propositions

• Precision, by combining data from our own data sources, trusted global partners as well as public sources.
• New granular data could be generated integrating sensors on ground crew vehicles and our camera images
• Advanced machine learning algorithms developed with our (commercial and academic) partners

Benefits

• Better prediction of landing time, which allows airport stakeholders to plan and manage resources better  
• Better incident management
• Resource optimization
• Faster, more accurate decision making at gate and airport level, thereby reducing the impact of disruptions and savings in terms of time and cost to airlines and airports

Application overview

• Aircraft Arrival-time prediction: The module combines and enriches existing open source and Airlines’ data with ADB SAFEGATE airport systems output and machine learning capabilities, identifies patterns and predicts landing time and taxi-time under specific conditions e.g. weather, aircraft type, etc. more accurately.

The machine learning models are trained regularly to ensure better accuracy of predictions. The application can feed airport planning and ways allocation tools with an accurate prediction of arrival time at the gate to optimize ways allocation for aircraft before take-off and after landing.

• Resource allocation: This module optimizes resource planning and allocation with predictive estimates based on the Arrival time prediction application and reduces potential delays.

Relying on consolidated and enriched operational data and strong data science capabilities, the module trains and updates machine learning models every few weeks to predict in advance variable (vs. standard) Resource planning timings, given specific constraints and preferences. These hard constraints and soft preferences can be defined by the user. The module also includes computation tools for real-time update of planning and allocation in case of disruption or change of specific constraints.

• Disruption management: This module accelerates recovery time after a disruption or an incident by providing a prescriptive analysis of current operations and recommends scenarios that help airport personnel decide how best to restore operations at each gate.

For every scenario, the models compute the associated risks and KPIs. The potential impacts on cost, time and resources for each action taken are also displayed. Various use cases are combined to provide an end-to-end solution.