Headsets in the light aircraft cockpit: Speech intelligibility, PELs, and flight performance

摘要

Data from a flight simulation experiment revealed differences among contemporary ActiveNoise Reduction (ANR) and passive headsets on a variety of human factors metrics that directlyreflect piloting performance and safety in flight. Perhaps more importantly, these data point tothe need for further analyses and experimental trials to isolate the specific headset designparameters which are responsible for the significant performance effects found herein. Thisexperiment employed a high workload, realistic, simulated flight scenario over 3.5 hours with aheavy communications component, and performance measurement during the experiment wasperformed using a multivariate approach. Measures of speech intelligibility, protected exposurelevel (PEL), pilotsu2019 perceived workload, and flight control performance were collected for 3ANR headsets (Bose Aviation X, Sennheiser HMEC 300, LightSPEED Thirty 3G) and onepassive headset (David Clark H10-13S). One headset was used per each of 4 matched workloadflights, instrument-rated pilots served as subjects, and the FAA-certified iGATEU00100001 simulatorimposed 95 dBA Cessna 172 cockpit noise. Speech intelligibility results showed that at aSpeech Transmission Index (STI) value of 0.30, pilots required more ATC commandtransmissions to respond accurately when wearing the passive headset than when wearing anyof the 3 ANR headsets, while at STI values of 0.50 and 0.70 there were no headset differences.The PEL measures, obtained using measurement microphones under the subjectsu2019 headsets inthe cockpit noise and with earphone communications input, indicated that the exposure levelneeded to attain 80% speech intelligibility for the passive headset was 93.1 dBu2014a valueapproximately 10 dB greater than the PEL for any of the 3 ANR headsets. Pilotsu2019 ratings ofworkload were slightly but significantly lower under the ANR headsets than the passiveheadset, though there were no differences among the 3 ANR devices. Three of 4 measuresreflecting flight control performance favored one ANR headset as compared to the other 3devices. Recommendations for follow-on work with implications for aviation headset designare presented.