(AI) Computer vision, a technique that uses cameras and ML to monitor complex ground servicing activities, detect safety issues or sound alarms when a service is taking longer than expected.
Data Flows
Title
Devices
Icon
Description
Various types of cameras at all critical locations
Real-time image data to be collected as well as other sensor data (weather / temp / etc.)
Title
Connectivity
Icon
Description
Images transmitted to edge and enterprise storage
Develop MV models driven by analytics needs
Title
Edge Compute
Icon
Description
AI (MV) to analyse images in real-time for direct action to be taken
Assume both real-time and non-real-time needs
Title
Cloud Compute & Storage
Icon
Description
All data is stored in enterprise data storage, where data ownership belongs to the company
Title
Applications & Services
Icon
Description
Non-critical MV / ML models execution
Automated processes by default
Title
Inform Decision Makers
Icon
Description
Exploit XR to display passenger analytics
Title
Support Decision Making
Icon
Description
End of process
Application Logic
Description
Personal data privacy and protection is to be considered when collecting, analysing and viewing data.
Assume that real-time is the norm which means that direct intervention is possible.
Collect images from different angles as they will have a different purpose; Also collect from multiple sources (airline / refuel company / etc.).
Security issues need to be considered as information would be of interest to many parties.
Description
All AI (ML and MV) models will be developed through a joint effort of SME and data scientists.
Development of these models will be iterative (will require time) to achieve an appropriate quality level.
There will be multiple MV models as ground operations will be tracked for various reasons.
Focus will be on MV, however ML is required for analytics of other data types, e.g. for weather forecast info as input to planned timings for ground operations.
Description
MV and ML models have been developed.
Most of the time series data are likely not time critical, therefore cloud based.
All data will be stored in Enterprise data storage, provided as needed, for AI applications.
It is important as AI is able to predict the effort and time it takes to turnaround the plane; For this access to all data and of course learnings for the future.
Expected benefits
5G enables sensor and cameras to be installed and will enable real-time perspective
View conflicts before occurrences and resolve them before they happen
Ability to view safety behaviour and compliances
Key value created
Reduced turnaround times for planes on ground, contributing direct value