Published in Aviation Business, April 2018
Republished by permission of Aviation Business
Current research and development in big data and artificial intelligence (AI) is producing considerable excitement in the aviation and aerospace industries. This is early days for both fields, so “potential” is the operative word. But a peek over the horizon confirms that big data and AI are transforming aviation and aerospace.
For much of the second half of the twentieth century, commercial transport aircraft produced data for their airlines, the most common being aircraft position, crew activity, weather systems, fuel consumption, and engine system performance.
Today, however, we are entering a new age of aircraft sensors and processors that will put the “big” in aviation data gleaned from throughout the aviation ecosystem—areas such as connectivity, operations, predictive maintenance, and more. It’s no exaggeration to suggest that applications for aviation big data are going to explode across the industry.
Here’s one example of how the aviation industry is enabling and leveraging big data. Rockwell Collins’ solution for the Airbus flight operations and maintenance exchanger (FOMAX) is a compact digital connectivity unit. As an enabling component of Airbus’ Skywise Predictive Maintenance services, FOMAX will ensure each A320 family aircraft captures more than 24,000 health parameters compared to 400 previously.
As impressive as this is, the aviation industry’s use of big data to further enable predictive maintenance will expand considerably in the next decade. Bigger data sets will enable OEMs to identify long-term trends. These trends will ensure the maintenance of complex aircraft systems with far more accuracy and timeliness than is available today. Thanks to big data, aircraft flight delays due to maintenance issues may become another relic of “the bad-old, good-old days.”
As with big data, exciting things are afoot in artificial intelligence. For many people, the first application that comes to mind is AI on the flight deck as kind of a cybernetic copilot. While this may be our future, it isn’t a near future in commercial aviation. There are other applications of AI that we’ll see first. To help enable these applications are two areas of aerospace research: efficient verification processes and cyber-resilient connectivity.
Efficient Verification Processes
Safety is paramount in aviation. As a highly-regulated industry it has been successful in making commercial aircraft travel one-hundred-times safer than traveling by automobile. This achievement, however, is the result of a time-consuming certification process with many stakeholders. In one notorious example, it took seven years to get an integrated avionics system approved by five international authorities. Imagine the verification/certification processes that will be necessary for AI systems, which cannot be certified through traditional FAA processes. Before AI with commercial applications can reach its potential, new, more efficient verification processes must be developed to provide the expected level of trust in AI.
Connectivity is the key to a reimagined commercial aviation ecosystem that’s driven by processors, sensors and smart aviation subsystems that are able to leverage them. Connectivity with the depth and breadth of an ecosystem requires that crucial attention be paid to cybersecurity and the development of cyber-resilient systems. Again, it’s no exaggeration to say that designing cyber-resilient systems for AI will be exponentially more complex than it is for current avionics systems. But making the complex less complex is what drives engineers. Rockwell Collins’ Advanced Technology Center is working with the Defense Advanced Research Projects Agency (DARPA) on Cyber Assured Systems Engineering (CASE). The goal of CASE, as explained on the DARPA website, is to develop the necessary design, analysis and verification tools to enable systems engineers to design-in cyber resilience. The knowledge gained here could find its way to AI.
Big data analytics and AI will impact the aviation and aerospace industries in ways big and small within the next decade. Simply put, a new generation of products and programs will require a complex understanding of the environmental parameters in which they work.