Electrical and Electronic Engineering - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 5 of 5
  • Item
    Thumbnail Image
    Efficient algorithms for autonomous agents facing uncertainty
    Selvaratnam, Daniel Devishtan ( 2018)
    This thesis considers the design and mathematical analysis of algorithms enabling autonomous agents to operate reliably in the presence of uncertainty. The algorithms are designed to preserve computational tractability, and to respect communication constraints. Four specific problems are addressed. First, the localisation of a signal source using only random binary measurements. A Bayesian estimation procedure is adopted that discretises the search space to achieve tractability. The effect of this discretisation on convergence is analysed rigorously, as well as the effect of relying on an inexact measurement model. Measurement locations are also optimised with respect to Fisher Information. In the second, the security of general quantized Bayesian estimators is analysed from the perspective of an adversary that sends false measurements to induce a misleading posterior. Fundamental limits on the set of posteriors that can be induced are derived, along with strategies to induce them. The third problem considers the design of control laws for maintaining reliable communication links between agents as they traverse the environment. Robustness to disturbances is established theoretically. Finally, optimisation problems are tackled involving cost functions and constraints that change unpredictably as new information becomes available. Performance bounds are provided for different classes of cost functions, and both first-order and gradient free methods are examined.
  • Item
    Thumbnail Image
    Acoustic beamforming analysis for wearable blind aid applications
    Lim, Wei Shen William ( 2018)
    The World Health Organisation estimates 36 million are blind worldwide; in addition, 217 million have severe or moderate visual impairment. Over the past decades, there has been substantial research in alleviating blindness and visual impairment. However, the blind community has yet to widely accept a single electronic travel aid (ETA) solution; the low cost white cane still remains the most popular device for orientation and mobility. One major limitation of current ETAs is their poor cost-benefit ratio. However, semiconductor advances may have reached a point where previous limitations are now surmountable as miniaturisation, flexible, low-cost and low-power circuits have been key enablers of wearable technology. Sonar has consistently been the preferred modality for single-sensor ETAs. The thesis aims to study performance characteristics of various beamforming aspects in their relation to developing a wearable-sonar system for blind aid applications. The scope of analysis covers 1) Classical Beamfomers 2) Beamforming Augmentation (Geometry, Shading, Adaptive Algorithms) 3) Spherical (3D) Beamsteering and Conformal Arrays
  • Item
    Thumbnail Image
    Noise reduction for cochlear implants
    HERSBACH, ADAM ( 2014)
    Cochlear implant (CI) users generally achieve acceptable speech understanding in quiet conditions, but have difficulty understanding speech in the presence of background noise. In this case, noise reduction processing can be utilised to help improve the situation, and solutions can be distinguished based on the number of microphones used to sample the acoustic environment. Single microphone solutions rely on the statistical properties of speech and noise while multi-microphone solutions can use the spatial characteristics of impinging sound to further separate speech from noise. It is the latter that forms the focus of the current research. A multi-microphone noise reduction algorithm was developed for a CI sound processor that attenuated sound from the rear while passing sound from in front of the listener. The algorithm used two microphones with small physical separation to generate two fixed directional patterns; one facing forward, the other towards the rear. By examining the front-to-back energy ratio, a signal-to-noise ratio (SNR) estimate was obtained, which was used to attenuate noise dominated frequency channels. The algorithm was evaluated acutely with CI listeners, primarily using an adaptive speech reception threshold (SRT) task, although sound quality and acceptable noise level were also studied. The acoustic environment used for evaluation in the laboratory included complex situations. These situations used various numbers of competing talkers or interfering speech weighted noise sources that changed spatial locations during the test. Reverberation was introduced and the algorithm was evaluated in a range of reverberant environments. Microphone sensitivity matching was investigated by introducing controlled levels of mismatch and measuring speech intelligibility performance. The evaluation revealed the algorithm was highly beneficial across a wide range of acoustic situations, outperforming a conventional generalised side-lobe canceller algorithm called Beam. The benefit varied with the spatial configuration of the competing noise and was greatest when the noise was located to the sides and rear of the listener. The benefit in reverberant conditions was maintained. Counter-intuitively, the benefits actually increased in the highest level of reverberation that was evaluated. Microphone mismatch had a detrimental effect on all multi-microphone algorithms that were evaluated, completely negating any multi-microphone benefit when the mismatch was 4 dB or greater. Finally, the algorithm was implemented in a wearable sound processor and CI users evaluated the algorithm outside the laboratory during their normal use of the device. Users were able to vote for their preferred listening program using their processor’s remote control device. The take-home evaluation consolidated the benefits measured acutely in the laboratory and provides critical guidance as to how the algorithm could be integrated into a commercial device.
  • Item
    Thumbnail Image
    Design and signal processing for CMOS automotive radar
    Li, John Zhong-Chen ( 2014)
    There is an increasing use of radar for sensing the environment in automotive applications to provide data for applications such as collision avoidance and adaptive cruise control systems. In this thesis, the waveform design, signal processing and architecture of automotive radars are explored. A particular emphasis is placed on reducing the implementation cost to enable widespread adoption of safety systems. While an automotive radar is unlikely to experience intentional jamming, the anticipated increase in density of radars with falling cost and improved availability is expected to lead to more interference as more users begin to share the available band. It is thus important that the performance of the system is understood in the presence of interference. This thesis provides some insight into the severity of the problem and some strategies for mitigating the impact. The requirements of automotive radar typically mandate the use of full-amplitude continuous-wave radar transmitters which occupy as much bandwidth as possible to meet the constraints on transmitted power for detecting targets at long range. Thus, some of the usual interference mitigation techniques in mobile communications such as power control or frequency or time division are not readily applicable. Instead, the available spectrum must be shared with a code-division multiple access scheme. It is important to have an understanding of the availability of safety applications under interference. Thus, a simple but typical automotive interference environment is modelled under the assumption that a multiple access scheme is employed. The model is used to determine empirical guides on the number of effective channels required under various scenarios, as well as to understand the average and worst case times that users will be blocked. Two classes of frequency modulated waveforms are investigated in this thesis. First, Linear Frequency Modulated Continuous Wave (LFM-CW), which has emerged as a popular choice in automotive radar because it readily admits a simple transceiver architecture. The return signals are usually processed using FFTs. However, mitigating interference with such waveforms requires the sacrifice of some estimation performance. A second scheme of using Random Stepped Frequency (RSF) waveforms scheduled in conjunction with single tone continuous wave (CW) waveforms is also presented. A common transceiver architecture can be used but the estimation performance of the RSF waveform can be improved with more sophisticated signal processing beyond the typical limits imposed by the FFT processing typically used with LFM-CW waveforms, while still maintaining low sampling and data rates. The performance of the algorithm when the system is in a non-optimal state due to interference or hardware limitations is demonstrated to degrade gracefully. The waveform also lends itself to interference mitigation as in each time slot each user is only concerned with a small part of the full band. The increased flexibility of RSF waveforms will thus allow radars to operate cognitively by rearranging their transmitted sequence when spurious returns from other users are detected.
  • Item
    Thumbnail Image
    Topics in resource optimization in wireless networks with limited feedback
    He, Yuan Yuan ( 2011)
    This thesis focuses on the design of various optimal resource allocation algorithms for wireless communication networks with imperfect channel state information (CSI) available at the transmitter obtained via a finite-rate feedback channel from the receiver (the so-called limited feedback technique). We first look at an M parallel block-Nakagami-fading channels where we seek to design an optimal power allocation scheme that minimize the outage probability under a long term average transmit power (ATP) constraint with quantized CSI. A simultaneous perturbation stochastic approximation algorithm (SPSA) based simulation-optimization method is applied to obtain a locally optimal power codebook. As this method is computationally intensive and time-consuming, we then derive a number of reduced-complexitysuboptimal finite-rate power codebook design algorithms. For the large number of parallel channels case, a Gaussian approximation based low-complexity power allocation strategy is provided. We then consider a secondary user (SU) transmit power control problem in a spectrum sharing cognitive radio networks scenario with quantized channel feedback for optimizing relevant performance measures such as secondary ergodic capacity or outage probability under interference power constraints (which can be restricted either by an average (AIP) or a peak (PIP) constraint) at primary user (PU) receivers to protect the PU, and an average transmit power (ATP) constraint on SU. Firstly, we study the problem of ergodic capacity maximization over M parallel channels (each of which is licensed to a distinct PU) of SU subject to an ATP constraint at SU and M individual AIP constraints on each PU with quantized feedback of the joint channel space of SU transmitter to SU receiver and SU transmitter to PU receivers. We develop a “modified generalized Lloyds-type algorithm (GLA)” for finding a locally optimal quantized power codebook. An approximate but computationally efficient quantized power allocation algorithm is then derived for the case of large number of feedback bits. It is seen that only 3-4 bits of feedback per channel band achieves SU ergodic capacity close to the full CSI based performance. We also extend these results to the noisy limited feedback case. We then consider the problem of throughput maximization of SU with a finite rate power codebook under an ATP constraint at SU and N individual PIP constraints on each PU receiver. With quantized channel (from the SU-TX to each PU-RX) knowledge, three different quantized power allocation schemes are proposed corresponding to three distinct forms of CSI obtained regarding the channel from SU-TX to SU-RX link at SU-TX : full CSI, noisy estimated CSI and quantized CSI. Finally, we consider the joint optimization of the quantization regions and the transmission power codebook such that the outage probability of the SU is minimized while an ATP constraint at the SU and an AIP constraint on the PU are met. Explicit expressions for asymptotic behavior of the SU outage probability at high rate quantization are also developed.