Digital Hearing Protection
More than 33 million workers in North America are exposed daily to excessive level of noise and will, almost certainly, suffer from noise-induced hearing loss, with the personal and societal cost associated. Nonetheless, noise-induced hearing loss is 100% preventable, as long as the appropriate hearing protectors and communication devices are used to reduce the noise dose received by the workers. The research conducted over the last five years through the former Sonomax-ÉTS Industrial Research Chair in In-Ear Technologies led to the development of several prototypes of digital hearing protectors able to protect their wearers while enabling natural communication in noisy environments. Several innovative audio processing algorithms have already been developed, by detecting voice activity in low signal-to-noise ratios [JA9], designing detectors of alarms and warning signals [JA15], using modulation-band filtering to remove noise while keeping speech and warning signals [CP15]. Other “smart” hearing protection algorithms have also been developed, such as the ones for professional musician, with active noise control of the occlusion effect [JA13, CP12], or hypersensitive subjects [CP19]. While promising, these technologies do now require further optimization, as well as an effective implementation on the ARP digital signal processor so that a real-world evaluation of their benefits could be conducted. Several other audio processing algorithms are also to be investigated, such as embedded continuous fit-testing of an in-ear device based on prior developments on field-attenuation measurement [JA18, JA21], in-ear dosimetry [CP5, CP28, CP30], loudness-based uniform attenuation [CP27], level-dependent attenuation [CP12], as well as specialized audio processing algorithms able to combine hearing aid and hearing protection [CP1] to address the issue of the protection of hearing impaired workers [JA12]. Finally, as an initial proof-of-concept showed promising results, research on binaural beamforming processing using 3 outer microphones per ear will also be pursued, taking advantage of the ARP capabilities.
Communication in Noise
Workers need to be able to communicate in noise, and when the acoustic speech signal is too disturbed, in-ear microphone can be used [CP38], and the speech signal can be denoised with adaptive signal processing and its intelligibility improved using spectral extension techniques [CP20, CP21] for use with personal radio communication systems. Natural and robust two-way communication systems have been proposed [CP24] and through a coding of the vocal effort of the speaker and a model of the intended communication distance, dubbed RAVE (Radio-Acoustic Virtual Environments) [CP17] are now being envisioned. Further research is now required to better understand auditory processes involved in speech production under hearing protectors in noise and build the corresponding psychoacoustical model, so that RAVE systems can be validated in real-world situations for industrial and tactical applications.
In-Ear Sensing
Wearable technologies is a generic and hype term that includes more than a dozen market segments, including today’s most obvious “Smart Watches”, “Smart Glasses”, “Fitness and Activity Tracking”. In 2015, the global market of wearable technologies, was estimated at $22.7 billion, a mere 2.3% of the 1$ trillion market potential [Wearable Technology: 2015 Company Profiles, Market Analysis & Forecasts]. Many experts, including the applicant [CP4], believe that the human ear canal is nice place for the collection of several biosignals with superior quality. Using the ARP discrete sensors, the instrumentation for many physical and physiological monitoring applications will be developed, ranging from man-down detection and dead-reckoning using the 9-axis inertial unit, to the monitoring of body temperature and ear canal humidity using the ad-hoc embedded miniaturized sensors, to the measurement of galvanic conductivity using the in-ear electrodes. Biosignals such as heart beats and breathing rate will be extracted from the in-ear microphone signal using time-frequency analysis coupled with the adaptive filtering previously proposed [CP16] for low to medium noise applications. Algorithms have recently been proposed by the applicant [JA4] for the measurement of oto-acoustic emissions (OAE) and can be used to continuously monitor the auditory fatigue of a worker in medium to high noise environments. Finally, the use of in-ear miniaturized active electrodes will enable the capture of various electrophysiological signals, ranging from electrooculography (EOG) for detection of eye gaze, to electrocardiogram (ECG), to electroencephalographic (EEG) information for the measurement of auditory-evoked potentials as recently demonstrated [CP6]. CRITIAS research activities for the in-ear sensing will focus mainly on the development of innovative instrumentation as well as research on advanced audio and biosignal processing algorithms so that robust information could be made available for other researchers or developers for their own intended use.
