Improve ATM data quality using sensor streaming data

Jorge Martin

2018-11-15 11:31:32
Reading Time: 3 minutes

Aircraft today are loaded with hundreds of sensors that monitor the status of the aircraft and its context several times per second. This data is stored continuously during the flight and then synchronized to a data lake either wirelessly, as soon as the plane receives signal from the airline IT infrastructure on ground, or manually, by exchanging hot-swap storage drives from the aircraft’s QaS. Several actors can use that information in real time: the Airline operator itself, the maintenance operators and even the ATCs. All said actors can benefit from this the opportunity to improve the quality of their operations by augmenting their own datasets. One example of this would be by adding new mechanisms such as DCI to improve the overall ATM performance by fusing this data with other public/private streams of data already available.

 

Why it is important now?

First, satellite connections are getting better as more dedicated satellites are being deployed in near orbits to provide high speed connections to remote locations. Today, in-flight WIFI connections are becoming the norm as a result of this. Why don’t we reuse this capability to make sensor data available in real time?

Connected sensor gateways are becoming more reliable and powerful thanks to the growing smart home appliance market. The communication protocols between sensors and gateways have been enhanced and, thanks to the level of miniaturization we have achieved, more powerful computation can be performed on the same size and cheaper. Now, more and more sensor networks delegate some computation tasks to where the data is measured and collected (edge computing) in order to reduce the latency of data and reduce the data transfer rate requirements. As a result, upgrading sensor equipment to connect sensors to private networks is really becoming a cost-efficient improvement for aircraft fleets.

Also, new IoT protocols, tools and services are continuously emerging. For example:

  • IoT platform management portal to manage both the gateways and the processing pipeline using the same interface, ensuring the whole infrastructure is secure, reliable and up-to-date.
  • Gateway real-time operative systems to support edge computing using existing software libraries and workflows.
  • Message brokers to route sensor data messages from the gateway to the existing data processing connectors
  • Streaming data connectors to deal with the complexity of streaming data to avoid being overloaded with too much data to be analysed, but also providing the data as ordered as possible.
  • Data processing libraries and schedulers to generate valuable insights from data using cloud computing clusters.

Does Data fusion really matters in the quality of data?

The different stakeholders involved in the ATM system use data from different sources with different cardinalities and levels of quality. There are different initiatives to improve the overall performance of the ATM processes through the application of new collaborative mechanisms, but some of them are blocked as not enough data cardinality or quality is available since most of them are stored on separate silos. In the context of Safeclouds.eu, we are deploying DataBeacon, a secure infrastructure to fuse data from different data providers, ensuring data is confidentially stored and managed and also enabling selected data scientists to perform complex analytics that improves safety in the ATM environment, with promising results. Such safety improvements could have great impacts on passenger satisfaction ratio and could increase passenger demand.

Moreover, as well as improving the quality of the different indicators used during the operational phase, sensor data can enable new operational tasks to be applied in the system, such as scheduling preventive maintenance tasks automatically as soon as the values measured by sensors go out of bounds. Some maintenance facilities are including this sort of intelligent maintenance in their service portfolio for several aircraft types and benefit from noticeable reductions in turnaround time and aircraft availability. If maintenance facilities are already using this sort of data, they could possibly use the streaming data generated as a new source of the infrastructure that monitors the status of the ATM system during the operational phase, reducing both the deployment times and costs.

As mentioned, using streaming sensor data during the operational phase could bring important benefits for both stakeholders and passengers. Also, the technology available today already supports this and there are experienced professionals working on it. We will provide more examples and tools related to IoT in aviation in the following posts. Stay tuned!

Author: Jorge Martin

© datascience.aero