The environmental noise pollution due to civilian aviation now constitutes the primary obstacle to a sustainable future for air transportation, the continuous growth (+5% every year) of which comes along with increased levels of societal and health concerns worldwide. It is thus critical to mitigate the noise impact entailed by air traffic around airports of major cities – to begin with Hong Kong, whose international airport ranks 8th (resp. 1st) in the world in terms of yearly volume of passengers (resp. 1st cargo), and has recently stretched its capacity (3rd runway opening).

To tackle this issue, HKUST scientists conducted extensive research over the past 5 years, deploying both computational and experimental means (cf. Wu & Redonnet, Transportation Research Part D, 2023 and Wu, Chan & Redonnet, Applied Acoustics, 2024, respectively). This research benefited directly from several UROP projects, some of which were presented in international conferences and/or granted a UROP award :). For instance, one of these UROP projects consisted in acquiring a representative set of aircraft noise samples around HKIA, which was achieved through in-situ audio recordings of aircraft during take-off/landing flight phases. Not only these experimental data proved very valuable for validating the in-house computational tools developed at HKUST, but they also allowed conducting a laboratory psycho-acoustic test, wherein HKUST students ranked the noise signals’ respective level of annoyance according to their subjective perception.

The present UROP project will seek at going a step further, by exploiting the psycho-acoustic test through machine learning. To this end, the UROP student(s) will explore various machine learning strategies and models (e.g. decision tree), to establish the relationship between the psycho-acoustic test’s results and the sound quality metrics characterizing the aircraft noise recordings. The final objective is to uncover how human brains process aircraft noises and ultimately ranks them as more or less annoying.

Successfully conducting this research action will require the UROP participant(s) to acquire and master specific knowledge and skills, whether theoretical (noise physics, psycho-acoustics, artificial intelligence) or practical (manipulating experimental databases, coding/executing machine learning algorithms/protocols). On another hand, this research action shall provide the UROP participant(s) an exciting opportunity to taste the water of what R&D is all about, by tackling a challenging problem within an actual research framework.