RE: Flight Safety Foundation IASS 201929 Oct 2019 16:29
Unfortunately, these eye tracking studies was not using Seeing Machines equipment. Head mounted eye trackers. The analysis is very "mandrolic" and statistical. They would kill to see the automated live view or video of what the pilots were looking at!
https://www.researchgate.net/publication/320222322_Pilot_Flying_and_Pilot_Monitoring's_Aircraft_State_Awareness_During_Go-Around_Execution_in_Aviation_A_Behavioral_and_Eye_Tracking_Study
Eye Tracking
The participants’ eye movements were recorded using two 50 Hz head-mounted Pertech® eye-trackers.The participants’ fixations were mapped on an image of reference, chosen for each participant so that it contained 14 areas of interest (AOIs; see Figure 1). However, to reduce the number of comparisons,the data analysis was focused primarily on the six AOIs that received the majority of visual scans.For each participant, we analyzed two flight phases: final approach starting from 2,500 ft, and go-around (until the stabilization of the aircraft at 2,500 ft). For the purposes of analysis, the data were categorized by aircraft type (2: Boeing, Airbus), pilot role (2: PF, PM), flight phase (2: final approach,go-around), and the six most viewed AOIs. First, mean dwell time on each of the six AOIs wascalculated. Second, to perform complementary basic analyses of eye movements, we determined the occurrence of fixations and saccades through the whole eye-tracking recording. Eye movements under 30°/s and not exceeding 1° in dispersion were considered as fixations (as in Dehais et al.,2015). The computation of an eye metric based on the fixation and saccade ratio was then used to express the balance between the different visual strategies of explore (visual search) and exploit (i.e.,visual processing) (Goldberg & Kotval, 1999). Similar to Dehais et al. (2015), we first selected eye movements that resulted in a Cohen’s dvalue greater than .8, considered a large effect size (Cohen,1992). This led us to identify four categories of eye movements: (a) short fixations of 40 to 80 ms, (b)
medium fixations of 180 to 220 ms, (c) long fixations of 340 to 500 ms, and (d) saccades of 140 to 260 ms. We then constructed a modified ratio defined as R = (short fixations + saccades)/(medium fixations + long fixations). High R values indicate explore behavior (visual search) whereas low Rvalues expressed exploit behavior (processing; Dehais et al., 2015; Goldberg & Kotval, 1999).
Sign In
Follow the stocks
