Pg Emeroylariffion Abas | Imperial College London (original) (raw)

Papers by Pg Emeroylariffion Abas

Water

For countries with an abundance of rain, there is definite potential to implement a rainwater har... more For countries with an abundance of rain, there is definite potential to implement a rainwater harvesting system for different applications. This paper describes feasibility studies of an open-pond rainwater harvesting system for ablution purposes, analysing the quality of harvested rainwater and formulating a rainwater harvesting model with suitable performance measures. The formulated model can be used to analyse the feasibility of the system in any locality by inputting local meteorological data. Quality analysis has shown that the harvested rainwater can be used safely for ablution purposes, albeit with a slightly acidic pH below 6.5. At a depth of 1.0 m and using the current pond configuration of a local mosque, the reliability of the system is 62.5% (228 days per year), and the amount of water saved is 345 m3, which is 60.7% of the water demand. It has been shown that a pond surface area of 60–70 m2 provides optimum reliability and water saving, and more water savings can be ex...

Sustainability

A home electrical composter has arisen as a popular tool to expedite the lengthy composting proce... more A home electrical composter has arisen as a popular tool to expedite the lengthy composting process. It has been conveniently selected as a compost producer in kitchen households and is especially favoured in urbanized settings. The generated composts from the electrical composter, however, are still found to be immature and would require additional curing. The quality of the compost can be improved by investigating the initial carbon-to-nitrogen ratio (C/N ratio) of kitchen waste. It is, therefore, the aim of this paper to determine the optimum initial C/N ratio by preparing two primary samples: with and without soil. Samples of 10:1, 15:1, 20:1, 25:1, 30:1, and 35:1 C/N ratios were fed into the electrical composter and allowed to cure for 4 weeks. The six main samples were further divided into sub-samples for replications. The phytotoxicity levels and maturity of the produced compost were assessed in terms of the germination index (GI), using a seed germination test. In addition, ...

Inventions

Organic waste management is a major global challenge. It accounts for a significant portion of wa... more Organic waste management is a major global challenge. It accounts for a significant portion of waste that ends up in landfills, where it gradually decomposes and emits methane, a harmful greenhouse gas. Composting is an effective method for potentially solving the problem by converting organic waste into valuable compost. Despite many studies focusing on the composting process, no study has reviewed the technological advancements in the composting fields from the perspective of patents. This review paper begins with background information on the composting process, specifically important factors affecting the process, problems associated with it, and the available technologies to facilitate the process. Different technologies are discussed, ranging from manual to automated methods. Subsequently, 457 patents are selected, classified into different categories, and reviewed in detail, providing a patent technology landscape of composting technology. Automatic composters are more promin...

8TH BRUNEI INTERNATIONAL CONFERENCE ON ENGINEERING AND TECHNOLOGY 2021

The unavailability of public datasets is the main hurdle for speech processing research targeting... more The unavailability of public datasets is the main hurdle for speech processing research targeting under-resourced languages. This paper reports the collection of a speech dataset comprising ten digits from the Kadazan language, which is one of the indigenous southeast Asian languages. Benchmark results for keyword spotting over the dataset using a convolutional neural network, have also been reported, with the benchmark model showing an average classification accuracy of 75.4% across multiple experiments using the dataset. Additionally, the dataset and implementation of the benchmark model have been made public, to facilitate replication and future research in the area of speech processing technologies for the Kadazan language.

8TH BRUNEI INTERNATIONAL CONFERENCE ON ENGINEERING AND TECHNOLOGY 2021

Recognition of human activities is one of the most significant fields in the area of machine inte... more Recognition of human activities is one of the most significant fields in the area of machine intelligence research, with its main objective of distinguishing various human activities. Accurate human posture extraction from images is a crucial step towards this objective. In this paper, laboratory dataset consisting of 2D and 3D images of subjects performing seventeen different activities have been collected, and then used to construct 2D and 3D human postures. Eight different classification models are subsequently used to classify the different activities. Subsequently, it has been shown that Random Forest classification model gives the best performance, in terms of accuracy. The findings also demonstrate that both 2D and 3D postures are capable of achieving significant accuracy score, with very high average accuracies of 99.18% and 99.82%, respectively, using the Random Forest classification model.

8TH BRUNEI INTERNATIONAL CONFERENCE ON ENGINEERING AND TECHNOLOGY 2021

A photonic crystal fiber sensor with dimensions of 2 cladding rings of circular and hexagonal air... more A photonic crystal fiber sensor with dimensions of 2 cladding rings of circular and hexagonal air holes and a single core hole for sulfuric acid sensing is proposed and investigated through a full vector Finite Element Method on COMSOL Multiphysics software. An aqueous sulfuric acid of varied concentrations of 0%, 10%, 20%, 30% and 40% is selected as the test analyte. The sensor established remarkable results of effective refractive index, power fraction, relative sensitivity, confinement loss and chromatic dispersion. At optimum wavelength of 1.3 μm, 0%, 10%, 20%, 30%, 40% sulfuric acid concentrations depict relative sensitivities of 95.35%, 96.16%, 96.71%, 97.19% and 97.59%, respectively, and confinement losses of 5.15×10 −10 dB/m, 1.60×10 −10 dB/m, 4.16×10 −11 dB/m, 1.17×10 −11 dB/m and 3.92×10 −12 dB/m, respectively.

IAES International Journal of Artificial Intelligence (IJ-AI)

Speech dialects refer to linguistic and pronunciation variations in the speech of the same langua... more Speech dialects refer to linguistic and pronunciation variations in the speech of the same language. Automatic dialect classification requires considerable acoustic and linguistic differences between different dialect categories of speech. This paper proposes a classification model composed of a combination of classifiers for the Arabic dialects by utilizing both the acoustic and linguistic features of spontaneous speech. The acoustic classification comprises of an ensemble of classifiers focusing on different frequency ranges within the short-term spectral features, as well as a classifier utilizing the ‘i-vector’, whilst the linguistic classifiers use features extracted by transformer models pre-trained on large Arabic text datasets. It has been shown that the proposed fusion of multiple classifiers achieves a classification accuracy of 82.44% for the identification task of five Arabic dialects. This represents the highest accuracy reported on the dataset, despite the relative sim...

Machine Intelligence Research

Identification of bird species from their sounds has become an important area in biodiversity-rel... more Identification of bird species from their sounds has become an important area in biodiversity-related research due to the relative ease of capturing bird sounds in the commonly challenging habitat. Audio features have a massive impact on the classification task since they are the fundamental elements used to differentiate classes. As such, the extraction of informative properties of the data is a crucial stage of any classification-based application. Therefore, it is vital to identify the most significant feature to represent the actual bird sounds. In this paper, we propose a novel feature that can advance classification accuracy with modified features, which are most suitable for classifying birds from its audio sounds. Modified Gammatone frequency cepstral coefficient (GTCC) features have been extracted with their frequency banks adjusted to suit bird sounds. The features are then used to train and test a support vector machine (SVM) classifier. It has been shown that the modified GTCC features are able to give 86% accuracy with twenty Bornean birds. Furthermore, in this paper, we are proposing a novel probability enhanced entropy (PEE) feature, which, when combined with the modified GTCC features, is able to improve accuracy further to 89.5%. These results are significant as the relatively low-resource intensive SVM with the proposed modified GTCC, and the proposed novel PEE feature can be implemented in a real-time system to assist researchers, scientists, conservationists, and even eco-tourists in identifying bird species in the dense forest.

INDUSTRIAL, MECHANICAL AND ELECTRICAL ENGINEERING

Rainwater harvesting is one of the most promising alternative sources of fresh water, particularl... more Rainwater harvesting is one of the most promising alternative sources of fresh water, particularly in countries with abundance of rainfall all year round, and this is especially important for the agricultural sector, which traditionally requires plenty of water. This work studies the feasibility of implementing a rainwater harvesting system for an existing aquaponics farm in Brunei. A rainwater harvesting model, with system and economic measurement, has been developed, with local meteorological and market data as well as local aquaponics farm used for illustration purpose. Analysis has shown that the rainwater harvesting system has the potential to save up to 127 m 3 of water with a system reliability of up to 67.5% using tank sizes ranging from 0-5 m 3. The system is shown to be feasible with a 3.3 m 3 tank that would require approximately B$2,300 to construct; giving water unit cost at par to the current local water price. These results indicate the potential of rainwater harvesting system for agricultural use in the local context. Although local data have been used, the model can also be adapted for analysis at other localities, and for other applications, such as for domestic or industrial applications.

IAES International Journal of Artificial Intelligence (IJ-AI)

Despite substantial advances in network architecture performance, the susceptibility of adversari... more Despite substantial advances in network architecture performance, the susceptibility of adversarial attacks makes deep learning challenging to implement in safety-critical applications. This paper proposes a data-centric approach to addressing this problem. A nonlocal denoising method with different luminance values has been used to generate adversarial examples from the Modified National Institute of Standards and Technology database (MNIST) and Canadian Institute for Advanced Research (CIFAR-10) data sets. Under perturbation, the method provided absolute accuracy improvements of up to 9.3% in the MNIST data set and 13% in the CIFAR-10 data set. Training using transformed images with higher luminance values increases the robustness of the classifier. We have shown that transfer learning is disadvantageous for adversarial machine learning. The results indicate that simple adversarial examples can improve resilience and make deep learning easier to apply in various applications.

IAES International Journal of Artificial Intelligence (IJ-AI), 2022

Transfer learning allows us to exploit knowledge gained from one task to assist in solving anothe... more Transfer learning allows us to exploit knowledge gained from one task to assist in solving another but relevant task. In modern computer vision research, the question is which architecture performs better for a given dataset. In this paper, we compare the performance of 14 pre-trained ImageNet models on the histopathologic cancer detection dataset, where each model has been configured as naive model, feature extractor model, or fine-tuned model. Densenet161 has been shown to have high precision whilst Resnet101 has a high recall. A high precision model is suitable to be used when follow-up examination cost is high, whilst low precision but a high recall/sensitivity model can be used when the cost of follow-up examination is low. Results also show that transfer learning helps to converge a model faster.

Proceedings of the 2020 4th International Conference on Big Data and Internet of Things, 2020

Device identi cation is the process of identifying a device on Internet without using its assigne... more Device identi cation is the process of identifying a device on Internet without using its assigned network or other credentials. e sharp rise of usage in Internet of ings (IoT) devices has imposed new challenges in device identi cation due to a wide variety of devices, protocols and control interfaces. In a network, conventional IoT devices identify each other by utilizing IP or MAC addresses, which are prone to spoo ng. Moreover, IoT devices are low power devices with minimal embedded security solution. To mitigate the issue in IoT devices, ngerprint (DFP) for device identi cation can be used. DFP identi es a device by using implicit identi ers, such as network tra c (or packets), radio signal, which a device used for its communication over the network. ese identi ers are closely related to the device hardware and so ware features. In this paper, we exploit TCP/IP packet header features to create a device ngerprint utilizing device originated network packets. We present a set of three metrics which separate some features from a packet which contribute actively for device identi cation. To evaluate our approach, we used publicly accessible two datasets. We observed the accuracy of device genre classi cation 99.37% and 83.35% of accuracy in the identi cation of an individual device from IoT Sentinel dataset. However, using UNSW dataset device type identi cation accuracy reached up to 97.78%.

Environments, 2021

Water is a quintessential element for the survival of mankind. Its variety of uses means that it ... more Water is a quintessential element for the survival of mankind. Its variety of uses means that it is always in a constant state of demand. The supply of water most primarily comes from large reservoirs of water such as lakes, streams, and the ocean itself. As such, it is good practice to monitor its quality to ensure it is fit for human consumption. Current water quality monitoring is often carried out in traditional labs but is time consuming and prone to inaccuracies. Therefore, this paper aims to investigate the feasibility of implementing an Arduino-based sensor system for water quality monitoring. A simple prototype consisting of a microcontroller and multiple attached sensors was employed to conduct weekly onsite tests at multiple daily intervals. It was found that the system works reliably but is reliant on human assistance and prone to data inaccuracies. The system however, provides a solid foundation for future expansion works of the same category to elevate the system to be...

Photonics, 2021

A simple hexagonal lattice photonic crystal fiber model with liquid-infiltrated core for differen... more A simple hexagonal lattice photonic crystal fiber model with liquid-infiltrated core for different liquids: water, ethanol and benzene, has been proposed. In the proposed structure, three air hole rings are present in the cladding and three equal sized air holes are present in the core. Numerical investigation of the proposed fiber has been performed using full vector finite element method with anisotropic perfectly match layers, to show that the proposed simple structure exhibits high relative sensitivity, high power fraction, relatively high birefringence, low chromatic dispersion, low confinement loss, small effective area, and high nonlinear coefficient. All these properties have been numerically investigated at a wider wavelength regime 0.6–1.8 μm within mostly the IR region. Relative sensitivities of water, ethanol and benzene are obtained at 62.60%, 65.34% and 74.50%, respectively, and the nonlinear coefficients are 69.4 W−1 km−1 for water, 73.8 W−1 km−1 for ethanol and 95.4 ...

Technologies, 2021

The pressure pain threshold (PPT) is a useful tool for evaluating mechanical sensitivity in indiv... more The pressure pain threshold (PPT) is a useful tool for evaluating mechanical sensitivity in individuals suffering from various musculoskeletal disorders. The aim of this study is to investigate PPT at the heel area in order to assist in the design of orthotic shoes for sufferers of heel pain due to a calcaneal spur. The size and location of the calcaneal spur was determined by x-ray images, with PPT data measured around the spur at five points by using algometer FDIX 25. The pain test experiment was conducted by pressing each point to obtain the pain minimum compressive pressure (PMCP) and its location. The information of shoe size, spur location and dimensions, and the PMCP location for each individual is used to obtain the exact point location for applying a softer material to the shoe in-sole, in order to reduce heel pain. The results are significant as it can be used by designers to design appropriate shoe in-soles for individuals suffering from heel pain.

Electronics, 2020

Harvesting biomechanical energy is a viable solution to sustainably powering wearable electronics... more Harvesting biomechanical energy is a viable solution to sustainably powering wearable electronics for continuous health monitoring, remote sensing, and motion tracking. A hybrid insole energy harvester (HIEH), capable of harvesting energy from low-frequency walking step motion, to supply power to wearable sensors, has been reported in this paper. The multimodal and multi-degrees-of-freedom low frequency walking energy harvester has a lightweight of 33.2 g and occupies a small volume of 44.1 cm3. Experimentally, the HIEH exhibits six resonant frequencies, corresponding to the resonances of the intermediate square spiral planar spring at 9.7, 41 Hz, 50 Hz, and 55 Hz, the Polyvinylidene fluoride (PVDF) beam-I at 16.5 Hz and PVDF beam-II at 25 Hz. The upper and lower electromagnetic (EM) generators are capable of delivering peak powers of 58 µW and 51 µW under 0.6 g, by EM induction at 9.7 Hz, across optimum load resistances of 13.5 Ω and 16.5 Ω, respectively. Moreover, PVDF-I and PVDF-...

Sustainability, 2019

This paper formulates a rainwater harvesting model, with system and economic measures to determin... more This paper formulates a rainwater harvesting model, with system and economic measures to determine the feasibility of a rainwater harvesting system, which uses water from the mains to complement the system. Although local meteorological and market data were used to demonstrate the model, it can also be easily adapted for analysis of other localities. Analysis has shown that an optimum tank size exists, which minimizes the cost per unit volume of water. Economic performance measures have indicated that rainwater harvesting system is currently infeasible to be implemented in Brunei; with capital cost and water price being shown to be among the prohibiting factors. To improve feasibility, a combination of rebate scheme on capital cost and raising the current water price has been proposed. It has also been shown that the system is more viable for households with high water demand.

Journal of Optical Communications, 2018

A porous core photonic crystal fiber (PCF) for transmitting terahertz waves is reported and chara... more A porous core photonic crystal fiber (PCF) for transmitting terahertz waves is reported and characterized using finite element method. It is shown that by enveloping an octagonal core consisting of only circular air holes in a hexagonal cladding, it is possible to attain low effective material loss that is 73.8% lower than the bulk material absorption loss at 1.0 THz operating frequency. Moreover, a low confinement loss of 7.53×10–5 cm−1 and dispersion profile of 1.0823±0.06 ps/THz/cm within 0.7–1 THz are obtained using carefully selected geometrical design parameters. Other guiding properties such as single-mode operation, bending loss, and effective area are also investigated. The structural design of this porous core PCF is comparatively simple since it contains noncomplex lattices and circular shaped air holes; and therefore, may be implemented using existing fabrication techniques. Due to its auspicious guiding properties, the proposed fiber may be used in single mode terahertz...

The current research in wearable electronics is trending towards miniaturization, portability, in... more The current research in wearable electronics is trending towards miniaturization, portability, integration, and sustainability, with the harvesting of biomechanical energy seen as a promising route to improve the sustainability of these wearable electronics. Efforts have been made to prolong operational life of these harvesters, to overcome energy dissipation, lowering resonant frequency, attaining multi-resonant states as well as widening frequency bandwidth of these biomechanical energy harvesters. Herein, an electromagnetic insole energy harvester (EMIEH), capable of efficiently harvesting lowfrequency biomechanical energy, has been designed, fabricated and experimentally tested. The core component in the device is the vibrating circular spiral spring, holding two magnets as the driving force on the central platform of the circular spiral spring, and just in-line with the upper and lower wound coils. It has been shown that the harvester exhibits higher sensitivity to low-frequency external vibrations than conventional cantilever-based designs, and hence allows low impact energy harvesting such as harvesting energy from walking, running and jogging. The experimentally-tested four resonant frequencies occurred at 8.9 Hz, 28 Hz, 50 Hz, and 51 Hz. At the first resonant frequency of 8.9 Hz under base acceleration of 0.6 g, the lower electromagnetic generator can deliver a peak power of 664.36 mW and an RMS voltage of 170 mV to a matching load resistance of 43.5 X. The upper electromagnetic generator can contribute an RMS voltage of 85 mV, corresponding to the peak power of 175 mW across 41 X under the same experimental condition. Finally, the harvester has been integrated into the shoe and it is able to charge a 100 mF capacitor up to 1 Volt for about 8 minutes foot movement. The result has remarkable significance in the development of wireless body monitoring sensors applications.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Water

For countries with an abundance of rain, there is definite potential to implement a rainwater har... more For countries with an abundance of rain, there is definite potential to implement a rainwater harvesting system for different applications. This paper describes feasibility studies of an open-pond rainwater harvesting system for ablution purposes, analysing the quality of harvested rainwater and formulating a rainwater harvesting model with suitable performance measures. The formulated model can be used to analyse the feasibility of the system in any locality by inputting local meteorological data. Quality analysis has shown that the harvested rainwater can be used safely for ablution purposes, albeit with a slightly acidic pH below 6.5. At a depth of 1.0 m and using the current pond configuration of a local mosque, the reliability of the system is 62.5% (228 days per year), and the amount of water saved is 345 m3, which is 60.7% of the water demand. It has been shown that a pond surface area of 60–70 m2 provides optimum reliability and water saving, and more water savings can be ex...

Sustainability

A home electrical composter has arisen as a popular tool to expedite the lengthy composting proce... more A home electrical composter has arisen as a popular tool to expedite the lengthy composting process. It has been conveniently selected as a compost producer in kitchen households and is especially favoured in urbanized settings. The generated composts from the electrical composter, however, are still found to be immature and would require additional curing. The quality of the compost can be improved by investigating the initial carbon-to-nitrogen ratio (C/N ratio) of kitchen waste. It is, therefore, the aim of this paper to determine the optimum initial C/N ratio by preparing two primary samples: with and without soil. Samples of 10:1, 15:1, 20:1, 25:1, 30:1, and 35:1 C/N ratios were fed into the electrical composter and allowed to cure for 4 weeks. The six main samples were further divided into sub-samples for replications. The phytotoxicity levels and maturity of the produced compost were assessed in terms of the germination index (GI), using a seed germination test. In addition, ...

Inventions

Organic waste management is a major global challenge. It accounts for a significant portion of wa... more Organic waste management is a major global challenge. It accounts for a significant portion of waste that ends up in landfills, where it gradually decomposes and emits methane, a harmful greenhouse gas. Composting is an effective method for potentially solving the problem by converting organic waste into valuable compost. Despite many studies focusing on the composting process, no study has reviewed the technological advancements in the composting fields from the perspective of patents. This review paper begins with background information on the composting process, specifically important factors affecting the process, problems associated with it, and the available technologies to facilitate the process. Different technologies are discussed, ranging from manual to automated methods. Subsequently, 457 patents are selected, classified into different categories, and reviewed in detail, providing a patent technology landscape of composting technology. Automatic composters are more promin...

8TH BRUNEI INTERNATIONAL CONFERENCE ON ENGINEERING AND TECHNOLOGY 2021

The unavailability of public datasets is the main hurdle for speech processing research targeting... more The unavailability of public datasets is the main hurdle for speech processing research targeting under-resourced languages. This paper reports the collection of a speech dataset comprising ten digits from the Kadazan language, which is one of the indigenous southeast Asian languages. Benchmark results for keyword spotting over the dataset using a convolutional neural network, have also been reported, with the benchmark model showing an average classification accuracy of 75.4% across multiple experiments using the dataset. Additionally, the dataset and implementation of the benchmark model have been made public, to facilitate replication and future research in the area of speech processing technologies for the Kadazan language.

8TH BRUNEI INTERNATIONAL CONFERENCE ON ENGINEERING AND TECHNOLOGY 2021

Recognition of human activities is one of the most significant fields in the area of machine inte... more Recognition of human activities is one of the most significant fields in the area of machine intelligence research, with its main objective of distinguishing various human activities. Accurate human posture extraction from images is a crucial step towards this objective. In this paper, laboratory dataset consisting of 2D and 3D images of subjects performing seventeen different activities have been collected, and then used to construct 2D and 3D human postures. Eight different classification models are subsequently used to classify the different activities. Subsequently, it has been shown that Random Forest classification model gives the best performance, in terms of accuracy. The findings also demonstrate that both 2D and 3D postures are capable of achieving significant accuracy score, with very high average accuracies of 99.18% and 99.82%, respectively, using the Random Forest classification model.

8TH BRUNEI INTERNATIONAL CONFERENCE ON ENGINEERING AND TECHNOLOGY 2021

A photonic crystal fiber sensor with dimensions of 2 cladding rings of circular and hexagonal air... more A photonic crystal fiber sensor with dimensions of 2 cladding rings of circular and hexagonal air holes and a single core hole for sulfuric acid sensing is proposed and investigated through a full vector Finite Element Method on COMSOL Multiphysics software. An aqueous sulfuric acid of varied concentrations of 0%, 10%, 20%, 30% and 40% is selected as the test analyte. The sensor established remarkable results of effective refractive index, power fraction, relative sensitivity, confinement loss and chromatic dispersion. At optimum wavelength of 1.3 μm, 0%, 10%, 20%, 30%, 40% sulfuric acid concentrations depict relative sensitivities of 95.35%, 96.16%, 96.71%, 97.19% and 97.59%, respectively, and confinement losses of 5.15×10 −10 dB/m, 1.60×10 −10 dB/m, 4.16×10 −11 dB/m, 1.17×10 −11 dB/m and 3.92×10 −12 dB/m, respectively.

IAES International Journal of Artificial Intelligence (IJ-AI)

Speech dialects refer to linguistic and pronunciation variations in the speech of the same langua... more Speech dialects refer to linguistic and pronunciation variations in the speech of the same language. Automatic dialect classification requires considerable acoustic and linguistic differences between different dialect categories of speech. This paper proposes a classification model composed of a combination of classifiers for the Arabic dialects by utilizing both the acoustic and linguistic features of spontaneous speech. The acoustic classification comprises of an ensemble of classifiers focusing on different frequency ranges within the short-term spectral features, as well as a classifier utilizing the ‘i-vector’, whilst the linguistic classifiers use features extracted by transformer models pre-trained on large Arabic text datasets. It has been shown that the proposed fusion of multiple classifiers achieves a classification accuracy of 82.44% for the identification task of five Arabic dialects. This represents the highest accuracy reported on the dataset, despite the relative sim...

Machine Intelligence Research

Identification of bird species from their sounds has become an important area in biodiversity-rel... more Identification of bird species from their sounds has become an important area in biodiversity-related research due to the relative ease of capturing bird sounds in the commonly challenging habitat. Audio features have a massive impact on the classification task since they are the fundamental elements used to differentiate classes. As such, the extraction of informative properties of the data is a crucial stage of any classification-based application. Therefore, it is vital to identify the most significant feature to represent the actual bird sounds. In this paper, we propose a novel feature that can advance classification accuracy with modified features, which are most suitable for classifying birds from its audio sounds. Modified Gammatone frequency cepstral coefficient (GTCC) features have been extracted with their frequency banks adjusted to suit bird sounds. The features are then used to train and test a support vector machine (SVM) classifier. It has been shown that the modified GTCC features are able to give 86% accuracy with twenty Bornean birds. Furthermore, in this paper, we are proposing a novel probability enhanced entropy (PEE) feature, which, when combined with the modified GTCC features, is able to improve accuracy further to 89.5%. These results are significant as the relatively low-resource intensive SVM with the proposed modified GTCC, and the proposed novel PEE feature can be implemented in a real-time system to assist researchers, scientists, conservationists, and even eco-tourists in identifying bird species in the dense forest.

INDUSTRIAL, MECHANICAL AND ELECTRICAL ENGINEERING

Rainwater harvesting is one of the most promising alternative sources of fresh water, particularl... more Rainwater harvesting is one of the most promising alternative sources of fresh water, particularly in countries with abundance of rainfall all year round, and this is especially important for the agricultural sector, which traditionally requires plenty of water. This work studies the feasibility of implementing a rainwater harvesting system for an existing aquaponics farm in Brunei. A rainwater harvesting model, with system and economic measurement, has been developed, with local meteorological and market data as well as local aquaponics farm used for illustration purpose. Analysis has shown that the rainwater harvesting system has the potential to save up to 127 m 3 of water with a system reliability of up to 67.5% using tank sizes ranging from 0-5 m 3. The system is shown to be feasible with a 3.3 m 3 tank that would require approximately B$2,300 to construct; giving water unit cost at par to the current local water price. These results indicate the potential of rainwater harvesting system for agricultural use in the local context. Although local data have been used, the model can also be adapted for analysis at other localities, and for other applications, such as for domestic or industrial applications.

IAES International Journal of Artificial Intelligence (IJ-AI)

Despite substantial advances in network architecture performance, the susceptibility of adversari... more Despite substantial advances in network architecture performance, the susceptibility of adversarial attacks makes deep learning challenging to implement in safety-critical applications. This paper proposes a data-centric approach to addressing this problem. A nonlocal denoising method with different luminance values has been used to generate adversarial examples from the Modified National Institute of Standards and Technology database (MNIST) and Canadian Institute for Advanced Research (CIFAR-10) data sets. Under perturbation, the method provided absolute accuracy improvements of up to 9.3% in the MNIST data set and 13% in the CIFAR-10 data set. Training using transformed images with higher luminance values increases the robustness of the classifier. We have shown that transfer learning is disadvantageous for adversarial machine learning. The results indicate that simple adversarial examples can improve resilience and make deep learning easier to apply in various applications.

IAES International Journal of Artificial Intelligence (IJ-AI), 2022

Transfer learning allows us to exploit knowledge gained from one task to assist in solving anothe... more Transfer learning allows us to exploit knowledge gained from one task to assist in solving another but relevant task. In modern computer vision research, the question is which architecture performs better for a given dataset. In this paper, we compare the performance of 14 pre-trained ImageNet models on the histopathologic cancer detection dataset, where each model has been configured as naive model, feature extractor model, or fine-tuned model. Densenet161 has been shown to have high precision whilst Resnet101 has a high recall. A high precision model is suitable to be used when follow-up examination cost is high, whilst low precision but a high recall/sensitivity model can be used when the cost of follow-up examination is low. Results also show that transfer learning helps to converge a model faster.

Proceedings of the 2020 4th International Conference on Big Data and Internet of Things, 2020

Device identi cation is the process of identifying a device on Internet without using its assigne... more Device identi cation is the process of identifying a device on Internet without using its assigned network or other credentials. e sharp rise of usage in Internet of ings (IoT) devices has imposed new challenges in device identi cation due to a wide variety of devices, protocols and control interfaces. In a network, conventional IoT devices identify each other by utilizing IP or MAC addresses, which are prone to spoo ng. Moreover, IoT devices are low power devices with minimal embedded security solution. To mitigate the issue in IoT devices, ngerprint (DFP) for device identi cation can be used. DFP identi es a device by using implicit identi ers, such as network tra c (or packets), radio signal, which a device used for its communication over the network. ese identi ers are closely related to the device hardware and so ware features. In this paper, we exploit TCP/IP packet header features to create a device ngerprint utilizing device originated network packets. We present a set of three metrics which separate some features from a packet which contribute actively for device identi cation. To evaluate our approach, we used publicly accessible two datasets. We observed the accuracy of device genre classi cation 99.37% and 83.35% of accuracy in the identi cation of an individual device from IoT Sentinel dataset. However, using UNSW dataset device type identi cation accuracy reached up to 97.78%.

Environments, 2021

Water is a quintessential element for the survival of mankind. Its variety of uses means that it ... more Water is a quintessential element for the survival of mankind. Its variety of uses means that it is always in a constant state of demand. The supply of water most primarily comes from large reservoirs of water such as lakes, streams, and the ocean itself. As such, it is good practice to monitor its quality to ensure it is fit for human consumption. Current water quality monitoring is often carried out in traditional labs but is time consuming and prone to inaccuracies. Therefore, this paper aims to investigate the feasibility of implementing an Arduino-based sensor system for water quality monitoring. A simple prototype consisting of a microcontroller and multiple attached sensors was employed to conduct weekly onsite tests at multiple daily intervals. It was found that the system works reliably but is reliant on human assistance and prone to data inaccuracies. The system however, provides a solid foundation for future expansion works of the same category to elevate the system to be...

Photonics, 2021

A simple hexagonal lattice photonic crystal fiber model with liquid-infiltrated core for differen... more A simple hexagonal lattice photonic crystal fiber model with liquid-infiltrated core for different liquids: water, ethanol and benzene, has been proposed. In the proposed structure, three air hole rings are present in the cladding and three equal sized air holes are present in the core. Numerical investigation of the proposed fiber has been performed using full vector finite element method with anisotropic perfectly match layers, to show that the proposed simple structure exhibits high relative sensitivity, high power fraction, relatively high birefringence, low chromatic dispersion, low confinement loss, small effective area, and high nonlinear coefficient. All these properties have been numerically investigated at a wider wavelength regime 0.6–1.8 μm within mostly the IR region. Relative sensitivities of water, ethanol and benzene are obtained at 62.60%, 65.34% and 74.50%, respectively, and the nonlinear coefficients are 69.4 W−1 km−1 for water, 73.8 W−1 km−1 for ethanol and 95.4 ...

Technologies, 2021

The pressure pain threshold (PPT) is a useful tool for evaluating mechanical sensitivity in indiv... more The pressure pain threshold (PPT) is a useful tool for evaluating mechanical sensitivity in individuals suffering from various musculoskeletal disorders. The aim of this study is to investigate PPT at the heel area in order to assist in the design of orthotic shoes for sufferers of heel pain due to a calcaneal spur. The size and location of the calcaneal spur was determined by x-ray images, with PPT data measured around the spur at five points by using algometer FDIX 25. The pain test experiment was conducted by pressing each point to obtain the pain minimum compressive pressure (PMCP) and its location. The information of shoe size, spur location and dimensions, and the PMCP location for each individual is used to obtain the exact point location for applying a softer material to the shoe in-sole, in order to reduce heel pain. The results are significant as it can be used by designers to design appropriate shoe in-soles for individuals suffering from heel pain.

Electronics, 2020

Harvesting biomechanical energy is a viable solution to sustainably powering wearable electronics... more Harvesting biomechanical energy is a viable solution to sustainably powering wearable electronics for continuous health monitoring, remote sensing, and motion tracking. A hybrid insole energy harvester (HIEH), capable of harvesting energy from low-frequency walking step motion, to supply power to wearable sensors, has been reported in this paper. The multimodal and multi-degrees-of-freedom low frequency walking energy harvester has a lightweight of 33.2 g and occupies a small volume of 44.1 cm3. Experimentally, the HIEH exhibits six resonant frequencies, corresponding to the resonances of the intermediate square spiral planar spring at 9.7, 41 Hz, 50 Hz, and 55 Hz, the Polyvinylidene fluoride (PVDF) beam-I at 16.5 Hz and PVDF beam-II at 25 Hz. The upper and lower electromagnetic (EM) generators are capable of delivering peak powers of 58 µW and 51 µW under 0.6 g, by EM induction at 9.7 Hz, across optimum load resistances of 13.5 Ω and 16.5 Ω, respectively. Moreover, PVDF-I and PVDF-...

Sustainability, 2019

This paper formulates a rainwater harvesting model, with system and economic measures to determin... more This paper formulates a rainwater harvesting model, with system and economic measures to determine the feasibility of a rainwater harvesting system, which uses water from the mains to complement the system. Although local meteorological and market data were used to demonstrate the model, it can also be easily adapted for analysis of other localities. Analysis has shown that an optimum tank size exists, which minimizes the cost per unit volume of water. Economic performance measures have indicated that rainwater harvesting system is currently infeasible to be implemented in Brunei; with capital cost and water price being shown to be among the prohibiting factors. To improve feasibility, a combination of rebate scheme on capital cost and raising the current water price has been proposed. It has also been shown that the system is more viable for households with high water demand.

Journal of Optical Communications, 2018

A porous core photonic crystal fiber (PCF) for transmitting terahertz waves is reported and chara... more A porous core photonic crystal fiber (PCF) for transmitting terahertz waves is reported and characterized using finite element method. It is shown that by enveloping an octagonal core consisting of only circular air holes in a hexagonal cladding, it is possible to attain low effective material loss that is 73.8% lower than the bulk material absorption loss at 1.0 THz operating frequency. Moreover, a low confinement loss of 7.53×10–5 cm−1 and dispersion profile of 1.0823±0.06 ps/THz/cm within 0.7–1 THz are obtained using carefully selected geometrical design parameters. Other guiding properties such as single-mode operation, bending loss, and effective area are also investigated. The structural design of this porous core PCF is comparatively simple since it contains noncomplex lattices and circular shaped air holes; and therefore, may be implemented using existing fabrication techniques. Due to its auspicious guiding properties, the proposed fiber may be used in single mode terahertz...

The current research in wearable electronics is trending towards miniaturization, portability, in... more The current research in wearable electronics is trending towards miniaturization, portability, integration, and sustainability, with the harvesting of biomechanical energy seen as a promising route to improve the sustainability of these wearable electronics. Efforts have been made to prolong operational life of these harvesters, to overcome energy dissipation, lowering resonant frequency, attaining multi-resonant states as well as widening frequency bandwidth of these biomechanical energy harvesters. Herein, an electromagnetic insole energy harvester (EMIEH), capable of efficiently harvesting lowfrequency biomechanical energy, has been designed, fabricated and experimentally tested. The core component in the device is the vibrating circular spiral spring, holding two magnets as the driving force on the central platform of the circular spiral spring, and just in-line with the upper and lower wound coils. It has been shown that the harvester exhibits higher sensitivity to low-frequency external vibrations than conventional cantilever-based designs, and hence allows low impact energy harvesting such as harvesting energy from walking, running and jogging. The experimentally-tested four resonant frequencies occurred at 8.9 Hz, 28 Hz, 50 Hz, and 51 Hz. At the first resonant frequency of 8.9 Hz under base acceleration of 0.6 g, the lower electromagnetic generator can deliver a peak power of 664.36 mW and an RMS voltage of 170 mV to a matching load resistance of 43.5 X. The upper electromagnetic generator can contribute an RMS voltage of 85 mV, corresponding to the peak power of 175 mW across 41 X under the same experimental condition. Finally, the harvester has been integrated into the shoe and it is able to charge a 100 mF capacitor up to 1 Volt for about 8 minutes foot movement. The result has remarkable significance in the development of wireless body monitoring sensors applications.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.