Air New Zealand taps AI in airside operations

Air New Zealand has teamed up with Auckland Airport to test the use of computer vision to improve aircraft turnaround time and airside operations.

The artificial intelligence (AI) technology, supplied by Swiss AI software startup Assaia, uses cameras installed at ramps and aerobridges to capture footage of some 50 turnaround and ground handling activities, such as cabin cleaning and refuelling.

The footage is then processed by Assaia’s software, which generates real-time insights and predictions to help airlines better manage their airside operations.

Marianita Willis, Air New Zealand’s programme manager for operational performance, said the Assaia system will provide additional data points that will enable the carrier to get its staff and equipment in place at each point in the turnaround process.

“For each flight there is a precise timeline associated with each activity to get the plane turned – understanding what is happening at each of these points enables us to proactively manage them,” she said.

Besides Air New Zealand, London’s Gatwick Airport has also deployed Assaia’s technology to improve airside operations, which are a good example of complex systems that AI was conceived to optimise, according to the airport’s head of IT commercial and innovation, Abhi Chacko.

“Considering the likely growth in air traffic coming our way, the earlier we start using AI to improve our operations, the more successful we will be as an airport,” he said.

Jean-Paul Isson, chief data science and AI officer at Sita, said AI has great potential in helping airlines shorten their turnaround time, which has been a key business challenge for them.

“The turnaround process can adversely impact on-time performance of airlines, airports and the overall passenger experience – often at great cost, both monetary and reputational,” he said in a blog post.

“Today, the process is managed by a turnaround manager who follows a rules-based list of activities that need to take place and ensures that all these activities are performed.”

Isson added that Sita is also prototyping an AI system that leverages computer vision to capture and explain the turnaround process, from on-block to off-block, and to provide real-time recommendations to proactively manage disruption.

“We’re using machine learning first to understand and explain video captions of what’s happening around the aircraft during turnaround. This will complement existing features to better explain turnaround activities and anticipate disruption during the process.

“We’re also using advanced optimisation algorithms to provide the best sequence of activities that should take place under some constraints to effectively manage turnaround and disruption,” he said.