Masters Degrees (Electrical and Electronic Engineering)
Permanent URI for this collection
Browse
Browsing Masters Degrees (Electrical and Electronic Engineering) by Subject "Accelerometers"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
- ItemDevelopment and evaluation of an animal-borne sensor prototype with integrated kinetic energy harvesting for rhinoceros conservation(Stellenbosch : Stellenbosch University, 2023-03) Robinson, Simon; Niesler, Thomas; Wolhuter, Riaan; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: This project aimed to evaluate the potential of a previously developed kinetic energy harvester as a means to improve the battery life and prolong the deployment of an animal-borne sensor. It builds on animal-borne sensor hardware which was developed previously and forms a part of the ongoing RhinoNET research project. This project has already lead to the development of a GPS enabled animal-borne sensor with accelerometer-based on-board behaviour classification, and which can communicate over a LORA radio network. This project extends this work by extending the existing hardware and integrating a kinetic energy harvester and associated energy management circuitry into the sensor. Energy management software for this updated animal-borne sensor is included, in order to allow the minimisation of its energy usage. This updated sensor and software are tested by practical measurement of the energy output of the kinetic energy harvester and the energy consumption of the animal-borne sensor. Because physical measurements on rhinoceros was not possible, this evaluation was achieved first t hrough s imulation. A rhinoceros behaviour model was designed based on previously gathered behaviour data and used to simulate the operation of the entire animal-borne sensor when subjected to rhinoceros activity. This included the modelling of the energy management software of the sensor, the energy consumption of the sensor, and the energy generation of the kinetic energy harvester. Subsequently, the energy generation of the kinetic energy harvester was measured on a human volunteer. Through this combination of practical measurement and simulation, it was estimated that, when attached to a rhinoceros, the kinetic energy harvester would generate on average 14J of energy per day, which is enough to power the sensor and transmit a GPS location update four times. Alternatively, it can increase the battery life of the sensor by between 3 and 60 days, depending on the frequency of GPS updates. The simulation also showed that when compared to a naïve approach, the energy management software could reduce the energy usage of the animal-borne sensor from 1548.09J to 507.2J per day. It was observed that the energy output waveform of the energy harvester is not ideal for charging a lithium-polymer battery, as used in the prototype. Therefore, future work should include consideration of other energy storage technologies. If these could be found, the integration of the energy harvester would in principle allow indefinite operation of the animal-borne sensor.
- ItemDiscriminating coughs in a multi-bed ward environment(Stellenbosch : Stellenbosch University, 2021-12) Leng, Corwynne; Niesler, Thomas; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: Automatic cough detection algorithms play a key role in cough monitoring systems. These systems assess a patient’s recovery by monitoring the frequency and other characteristics of their coughs during treatment. An audio-based cough detection system evaluates features extracted from an audio signal and classifies these as either a cough or non-cough. However, these systems can struggle to distinguish between coughs of different individuals, such as a ward with more than one patient. This study designs and tests a cough detection algorithm that can reliably detect cough sounds in such an environment. We hypothesise that including the vibrations of the patient’s bed with the audio features will allow the classifier to differentiate between coughs originating from different beds in a hospital ward. Two datasets were compiled, containing simultaneously captured audio and accelerometer signals recorded by devices attached to the frame of each bed in a ward. These datasets were manually annotated with labels representing a cough from the bed in question, a cough from another source, and noise. The first dataset was used to train the classifier while the second was used to measure its performance. We extracted audio and accelerometer feature vectors using methods that have proved effective in automatic speech recognition systems, such as mel filter bank energies and mel-frequency cepstral coefficients. For the accelerometer signals, we included time-domain features. These feature vectors were presented to several classifiers, including convolutional neural networks and deep neural networks. Classifier training employed nested cross-validation to compensate for the small size of the training dataset and to allow for robust hyperparameter optimisation. The best classifier using audio and accelerometer features achieved an area under the receiver operating characteristic curve (AUC) score of 0.9842 while the best classifier using only audio features achieved an AUC score of 0.9334. Therefore, the inclusion of accelerometer features increased the AUC score by 0.0508 and improves the classifier’s ability to reject coughs from other sources by 10.72%. Additionally, the accelerometer features reduce the false detection of coughs from other sources and noise from 21.08% to 6.83%, while maintaining a sensitivity of 95%. We conclude that including the accelerometer signals of the patient’s bed with the audio features allows the classifier to better reject coughing sounds from sources other than the patient being monitored.