Abstract:
Cousteau, who invented the self-contained underwater breathing apparatus described the underwater world as the “silent” world. Though, the ocean is lled with ambient sounds and humans are simply not adapted to hearing underwater. This is partly because we lose our ability to localise the direction of incident sound underwater and can become lost and disoriented. We envision a hearing augmentation device that can restore this ability. We looked to sh for inspiration. Their otoliths measure the acoustic particle motion of an acoustic wave, a principle that is very di erent from our ears. The otolith hearing mechanism is similar to an engineering solution known as an inertial underwater acoustic vector sensor (UAVS). An UAVS resolves both the magnitude and directional component of an acoustic wave. It is not a complicated device conceptually but the signal that needs to be measured is small. The input to the system starts from the acoustic boundary condition on the surface of the UAVS. The incident acoustic wave displaces the UAVS, whose motion is measured by its inertial sensor. The inertial sensor converts the mechanical signal into an electrical signal that is nally measured by an electrical circuit. All aspects of the signal chain need to be considered carefully. In this thesis, we explored areas of the signal chain to address knowledge gaps and proposed ways to improve on its overall signal-to-noise ratio (SNR).We focused on capacitive inertial sensors speci cally due to its many favourable characteristics for use in an UAVS. We began by analysing the design of the UAVS housing. Besides its size and buoyancy, there has been little consideration for its other physical parameters. We analysed the e ect of its shape on directivity using a harmonic acoustic nite element model. Then we derived analytically the optimal mass distribution between its housing and its inertial sensor to arrive at the best mechanical SNR. From our analysis, we proposed a set of guidelines for designing the optimal UAVS housing. Then we proposed a novel digital, time mode capacitive sensing circuit for use with capacitive inertial sensors to improve on the electrical SNR of UAVS. This circuit was based on measuring the time di erence resulting from the reactance of capacitors. The theory was presented, followed by a circuit implementation using commercially available parts. A circuit noise density of 3.0aF/»Hz was achieved. Furthermore, the circuit showed great possibility for future improvements. To complete the UAVS signal chain and to test the dynamic performance of the circuit, we designed and built a macro-machined inertial sensor. A systematic analysis of the fringe capacitance was carried out for the capacitive sensor design using an electrical eld simulation. While we initially intended to deploy the sensor as an UAVS, we ran into several issues and was prevented from doing so. These were discussed and improvements were proposed. In summary, we improved the framework on all components of the UAVS signal chain throughout the thesis and hope our ndings will be useful towards future capacitive UAVS development.