Technology seems to have found a way to predict the erratic flight path of an inexperienced pilot in real time.
According to a recent report, the trajectory of any object can now be predicted by an algorithm developed by researchers from the Queensland University of Technology (QUT) through harnessing data analytics.
Their algorithm can make the predictions faster and more accurately than existing approaches.
The University’s Professor, who leads the Vision and Signal Processing research discipline in the Science and Engineering Faculty, claims that they can predict it as long as it has trajectory.
This tool can be very useful in a Defence environment as it can provide greater situational awareness of both owned and enemy assets and airspace.
Their algorithm can be applied to airspace, military bases, public transport or shopping centres. It can be applied anywhere movement needs to be analysed.
Deep neural networks and memory networks are two machine learning techniques that were combined by the unique algorithm in order to analyse and predict trajectories in real-time.
In essence, it was built to measure a trajectory in and predict a trajectory out.
As it is taking in the trajectory of the target object, it is also taking in the trajectories of neighbouring objects.
This is done to create an awareness of what is around the target and how those objects are moving.
In addition, it draws on memory networks of stored historical trajectories for the same location. This is an attempt to emulate how the human memory works.
Those two sets of data are then analysed by another subnetwork that determines where the target will go next.
The researchers trained the algorithm by using disparate big data sets in order to guarantee its robustness.
The data sets include air traffic control data from the Brisbane Airport, radar and camera data from pedestrian traffic at the University as well as pedestrian trajectory databases from Edinburgh and New York.
It can crunch about 1000 predictions in a couple of seconds.
With the data from Brisbane Airport during a 2015 severe weather event, the team was able to test how well their algorithm coped in such a dynamic situation.
Its predictions were very accurate because it factored how previous pilots behaved in similar conditions to predict what the target pilot is likely to do next.
This algorithm can be very useful potentially in civilian airspace as it could aid in managing drones, where there is an increasingly crowded and constrained airspace.
The research was funded in part by a A$ 100,000 Defence Science and Technology Group grant.
Hopefully, the project will be extended in the future so that the team can investigate how the algorithm could be used to optimise flight paths and travel routes.
