Can spatial audio support pilots? 3D-audio for future pilot-assistance systems

In modern cockpits most of the information is provided to the pilot visually. This information is typically presented head-down on multiple displays. The advantage of these glass cockpits tends to be impaired by constantly increasing the amount of information presented. Especially during landing however pilots need to maintain a good spatial awareness from visual references in the vicinity of the aircraft. Therefore, the pilot has to deal with multiple concurrent tasks all with dominant impact on the pilot's visual perception. With the limited human visual perception also the cognitive ability of humans might be reached [1, 2]. This introduces new operational burdens and failure modes to the overall human-machine system [3, 4]. To avoid impacts on safety a remarkable step forward would be firstly to reduce the demand on visual perception that would provide additional spare mental capacity to safely monitor and control the aircraft even under high workload phases or system failure conditions. Secondly, to find a way to maintain or even enhance the information flow at the same time. Today, the auditory component conveys no spatial information and is generally used to draw attention to a visual display [5]. Audio research has been sparse in aviation and only covered spatial audio with a set of speakers around the head of the pilot or simple left-right-volume difference in a stereo headset [6]. Several previous studies have suggested a multitude of applications for useful 3D-audio concepts in the cockpit [5, 7, 8]. Humans are able to localize a 3D-audio presentation via headset adequately enough to direct their attention to a specific point inside the cockpit or out-of-the-window. Tracking the head movement becomes one key feature to reduce frontback confusion and inside-the-head localization. With preselected assistance system 3D-audio has the capability to support pilots during critical flight phases potentially even decreasing the overall workload.