444 research outputs found
Intra-identity PatchSwap: On the Generalizability of Face Presentation Attack Detection
No full textWith the widespread deployment of face recognition systems, face presentation attack detection (PAD) plays an essential role in mitigating their vulnerabilities. Face PAD is employed before the identification system to detect if the presented face is an attack. However, most of the existing face PAD methods tend to overfit on the training data and fail to generalize well on unknown attacks in a real-world scenario. The main reason for such poor generalizability is that existing face PAD datasets are limited in quantity and diversity. Moreover, recent PAD works leverage pixel-wise supervision strategy and show great progress in face PAD performance. Nevertheless, obtaining accurate pixelwise labels is a challenging task. To alleviate these issues, we propose the plug-n-play PatchSwap approach in this thesis. The proposed PatchSwap method maximizes limited data utilization and generates more challenging bonafide/attack samples and partial attacks by swapping patches between training data by a well-designed strategy. Meanwhile, their pixel-wise labels are correspondingly updated. As a result, the augmented training samples contain more complex attack patterns, benefiting robust feature learning. Furthermore, we demonstrated the proposed PatchSwap method combined with three prevailing backbones: ResNet, DenseNet, and MixFaceNet. The extensive experiments were performed on four benchmark datasets under both intra-dataset and cross-dataset scenarios. We also conducted several detailed ablation studies to explore the effect of patch types, selected candidate identity, and the probabilities controlling the swapping process. The experimental results show that the proposed PatchSwap approach achieved significant performance improvement. For example, the ACER value on the most challenging Protocol-4 of Oulu-NPU [38] decreased from 20.51% achieved by DenseNet baseline to 3.41% by DenseNet-PatchSwap
Patch-based Discriminative Learning for Iris Presentation Attack Detection
No full textIris recognition is considered a prominent biometric authentication method. The accuracy, usability and touchless acquisition of iris recognition have led to their wide deployments. However, iris recognition systems are vulnerable to presentation attacks. Presentation attacks, such as printing pictures or wearing cosmetic contact lenses, are used by attackers to mimic the identity of another person or to hide their own identity. Detection of such attacks is essential to ensure the reliability of this biometric system. In recent years, deep learning based iris Presentation Attack Detection (PAD) approaches have already obtained good attack detection performance. However, the existing iris PAD methods still have several shortcomings, such as poor performance under changing environmental conditions and unknown attacks. To address these issues, this work proposes a novel iris PAD method by learning discriminative features from patches. This proposed method is based on the assumption that a unique iris pattern consists of several types of local features, which can also be used for iris PAD. To define the type of local features, we employ a clustering technique to group such local features. Subsequently, models are trained for each group to learn sublet patterns. We conduct detailed experiments to validate our assumption and analyze the effect of the patch size. Furthermore, experimental results, performed on three iris PAD datasets including unknown attacks and cross-dataset scenarios, demonstrate that our method outperforms the state-of-the-art iris PAD methods, indicating higher generalization ability for iris PAD.Die Iriserkennung gilt als eine der wichtigsten biometrischen Authentifizierungsmethoden. Die Genauigkeit, die Benutzerfreundlichkeit und die berührungslose Aufnahme haben zu einer weiten Verbreitung der Iriserkennung geführt. Allerdings sind Iriserkennungssysteme anfällig für Präsentationsangriffe. Präsentationsangriffe, wie z. B. das Ausdrucken von Bildern oder das Tragen kosmetischer Kontaktlinsen, werden von Angreifern genutzt, um die Identität einer anderen Person nachzuahmen oder ihre eigene Identität zu verbergen. Die Erkennung solcher Angriffe ist von entscheidender Bedeutung, um die Zuverlässigkeit dieses biometrischen Systems zu gewährleisten. In den letzten Jahren haben auf Deep Learning basierende Ansätze zur Erkennung von Iris-Präsentationsangriffen bereits eine gute Leistung bei der Erkennung von Angriffen erzielen können. Die bestehenden Iris-Präsentationsangriffserkennungs-Methoden weisen jedoch immer noch einige Mängel auf, wie z. B. eine schlechte Leistung bei sich ändernden Umgebungsbedingungen und bei unbekannten Angriffen. Um diese Probleme zu adressieren, wird in dieser Arbeit eine neuartige Methode zur Erkennung von Iris-Präsentationsangriffen vorgeschlagen, die diskriminierende Merkmale aus Patches lernt. Die vorgeschlagene Methode basiert auf der Annahme, dass ein einzigartiges Irismuster eine Kombination aus mehreren Arten von lokalen Merkmalen ist, die auch für die Erkennung von Iris-Präsentationsangriffen verwendet werden können. Um die Arten von lokalen Merkmale zu definieren, wird eine Clustering-Technik angewendet, um die lokalen Merkmale zu gruppieren. Anschließend werden Modelle für die einzelnen Gruppen trainiert, um Teilmuster zu lernen. In detaillierten Experimenten werden unsere Annahmen validiert und die Auswirkungen der unterschiedlichen Größe der extrahierten Flecken analysiert. Darüber hinaus zeigen die experimentellen Ergebnisse, die auf drei Iris-Präsentationsangriffserkennungs-Datensätzen, einschließlich unbekannter Angriffe und datenübergreifender Szenarien, durchgeführt wurden, dass unsere Methode die anderen Iris-Präsentationsangriffserkennungs-Methoden auf dem aktuellen Stand der Technik übertrifft, was auf eine höhere Generalisierungsfähigkeit der Iris-Präsentationsangriffserkennung hinweist
Momentum Contrast for Representative Face Presentation Attack Detection
No full textWith the widespread usage of using face recognition systems, they became vulnerable to presentation attacks encountered by attackers. To tackle this issue, face presentation attack detection (PAD) methods are implemented. However, these methods have several shortcomings including the generalizability of unknown attacks. This thesis targets two main problems that face PAD methods. The first problem that this work target to solve is databases annotation problems. Annotating databases with labels is time-consuming, to solve this problem, a representative learning model (MoCo framework in this thesis) is used as it focuses on unsupervised learning databases. The second problem that this work target is the insufficient PAD data. Most PAD databases are manually collected especially presentation attack samples, thus they are labor-intensive and small-scale. This thesis target this problem by training the model on a face recognition database such as CASIA-Web database which is a very large-scale public facial recognition database, not a PAD database, which is collected randomly in the wild where images are diverse from illumination, sensors, identity. This work proves that using face recognition databases to learn face representation, can be adapted to be used in detecting presentation attacks and the model can benefit from using extra existing face recognition data besides the model becomes more familiar with diverse setups and illuminations within face images. Finally, the classification model suggested by the state-of-art MoCo, is extended by applying pseudo labeling to it, which improved the general results
Bias Exploration and Mitigation in Face Presentation Attack Detection Systems
No full textWith the widespread application of face recognition systems, face presentation attack detection (PAD) plays a critical role to protect the security and credibility of the system. A growing number of researchers have investigated the biases of face recognition systems, and their results demonstrated the existence of demographic and attribute biases in recognition systems. However, the fairness of face PAD system has not attracted much attention. A key problem is the insufficient demographic and attribute annotations of face PAD data. Hence, this thesis first combines six face PAD databases: CASIA-FASD [45], REPLAYATTACK [12], MSU-MFSD [39], HKBU-MARs V1+ [30], OULU-NPU [8], WFFD [27], which consisting of print, replay, 3D mask, wax face attacks. Furthermore, identities from this combined database are manually annotated with one demographic label and six facial attribute labels. Second, this thesis explores and analyzes demographic bias and additionally facial attribute bias in face PAD methods by using this combined database. To enable the bias study, one hand-crafted feature based model LBP-MLP, and three deep learning based models: ResNet50 [23], DeepPixBis [20], and LMFD-PAD [16], are adopted. In addition to report PAD performance, a modified fairness discrepancy rate (FDR) is introduced to further determine the system fairness. The experimental results point out that deep learning based PAD models trained only on female or male group are unfairer than models trained on the fused data (including female and male). In addition, models trained on fused data and only on occlusion group indicate higher fairness than models training only on non-occlusion data. To further mitigate system bias, a modified version of PatchSwap [3], named cross-identity PatchSwap in this thesis, is introduced to enable patch substitution between identities with different gender and facial attributes. Despite the significantly improved PAD performance achieved by cross-identity PatchSwap, the FDR results also suggest that this approach is able to improve the system fairness for different gender groups when models are trained on fused data and on male data
골격성 III급 부정교합에서 수술 혹은 비수술 교정치료 후 저작 기능 회복 양상
No full textThe converntional treatment for skeletal Class III malocclusion had commonly been accompanied by orthognathic surgey for the improvement of facial profile and occlusion. However, in terms of the cost-effectiveness, risk of surgery and non-surgical treatment techniques have been developed, lots of patients prefer non-surgical orthodontic treatment. The aim of this study was to evaluate differences in masticatory function and recovery pattern of masticatory function following surgical or non-surgical correction of skeletal Class III malocclusion, and correlation between dynamic and static variables. Non-surgical group comprised 9 male patients (mean age 23.75 ± 3.01), surgical group comprised 8 male patients (mean age 26.25 ± 4.27). The variables were recorded immediately after the fixed appliance was removed (T0), 1 month post-treatment (T1), 6 months post-treatment (T2), 12 months post-treatment (T3). Maximum bite force, occlusal contact area were measured with Dental Prescale II system, and mixing ability were measured with Viewgum software. Repeated ANOVA was used to test the variations in maximum bite force, occlusal contact area and mixing ability (20 cycles) in each group during the study. The results are as follows: The maximum bite force and occlusal contact area showed a time-dependent gradual increase in the non-surgical group and the surgical group after treatment (P < 0.001). There was no significant difference in the maintenance period. The mixing ability showed a slow recovery pattern, but there was no statistically significant difference, and there was no significant difference of mixing ability comparison between two groups. The correlation coefficient between occlusal force and occlusal area was 0.944 (P < 0.001) in the non-surgical group and 0.807 (P < 0.05) in the surgical group. There was no significant correlation between mixing ability and maximum bite force, mixing ability and occlusal contact area. According to this study, the mean values of maximum bite force and occlusal contact area were higher in the non-surgical group, but the surgical group recovered faster than the non-surgical group. The above results are considered to be helpful in explaining changes in masticatory function in borderline skeletal Class III malosslusion.
본 연구의 목적은 골격성 III급 부정교합에서 수술 혹은 비수술로 치료한 환자의 교정치료 후 시간에 따른 저작 기능 차이와 회복 양상을 비교하고 동적 평가인 저작 효율과 정적 평가인 교합력 및 교합면적 사이의 상관관 계를 분석하는 것이다. 총 17명의 골격성 III급 부정교합 환자를 대상으로 비수술군 (9명, 평균 나이 : 23.75 ± 3.01), 수술군 (8명, 평균 나이 : 26.25 ± 4.27), 두 군으로 나누 었다. 압력 감지 필름 시스템을 사용하여 교합력, 교합면적을 측정하였고 두가지 색상의 껌을 씹어 색상의 혼합 정도로 저작 효율을 측정 하였다. 고정식 교정장치를 제거 및 고정식 유지 장치 접착 직후, 그 후 1개월, 6개 월, 12개월 되는 시기 각각 측정하였다. 두 군의 측정 값을 비교하고 정적 평가와 동적 평가의 상관관계를 조사하여 다음과 같은 결과를 얻었다. 1. 시간에 따른 변화에서 교합력과 교합면적은 치료 후 비수술군과 수 술군에서 점진적으로 증가하는 양상을 보였지만 (P 0.05). 2. 저작 효율은 시간에 따른 변화에서 느리게 회복되는 양상을 보였으 나 통계적으로 유의한 차이를 보이지 않았고, 군 간 비교시 유의한 차이가 없었다 (P > 0.05). 3. 교합력, 교합면적, 저작 효율 사이의 상관관계를 분석한 결과, 교합 력과 교합면적의 상관계수는 비수술군에서 0.944 (P < 0.001) 수술 군에서 0.807 (P < 0.05) 였고 교합력과 저작 효율, 교합면적과 저 작 효율 간 유의한 상관관계가 존재하지 않았다. 본 연구에 의하면 교정 치료 후 교합력과 교합면적은 각 시기별 평균 적 수치가 비수술군에서 더 높게 나타났지만 수술군이 비수술군보다 회복 이 빨랐다.반면, 저작 효율은 시기에 따른 전반적 향상에도 두 군간의 차 이를 보이지 않았다. 위 결과는 경계성 골격성 III급 부정교합에서 저작 기 능 변화를 설명하는데 도움이 될 수 있을 것으로 사료된다.open박
Development of generative adversarial networks for spatiotemporal fluid flow, atmospheric and flood predictions
No full textAccurate prediction of fluid flow is of vital importance to many applications in engineering and physics. Long-standing challenges in flow simulations are the simultaneous high accuracy and efficiency demanded in numerical computations. Recent advances in machine learning technologies are increasingly of interest for the efficient simulation of nonlinear and complex systems. Deep learning techniques are capable of capturing the physical dynamics without prior knowledge of underlying physical relationships.
However, there remain several challenges in nonlinear fluid flow modelling, i.e., spatiotemporal nonlinear fluid flow modelling, large data-driven fluid flow modelling, real-time forecasting of nonlinear fluid flows beyond the training period, improvement of the forecasting accuracy in long lead-time. To overcome these challenges, in this work, Generative Adversarial Networks (GANs) have been first introduced to spatiotemporal nonlinear fluid flow prediction while data assimilation techniques are used for improving the accuracy of long term forecasting.
The key contributions of this thesis are: firstly, an Artificial Intelligence (AI) fluid model based on a deep convolutional generative adversarial network (DCGAN) has been developed for modelling spatiotemporal flow distributions. Secondly, a hybrid deep adversarial autoencoder (VAE-GAN) model to integrate generative adversarial network (GAN) and variational autoencoder (VAE) has been proposed for large data-driven nonlinear fluid flow prediction. Thirdly, a real-time predictive machine learning model, i.e., artificial neural network (ANN), long short term memory (LSTM), DCGAN, and VAE-GAN has been developed for basin streamflow, urban flooding, and national air pollution forecasting problems respectively. Fourthly, an ensemble Kalman filter (EnKF) for GAN and convolutional LSTM (GAN-ConvLSTM) based forecasting system has been proposed for accurate long lead-time forecasting.
The presented models have been validated by various test cases and the results are in good agreement with high-fidelity models and observations. Promising results have shown that the DCGAN and VAE-GAN models are capable of accurately predicting the spatiotemporal flow features as the flow evolves, with CPU speed-up of several orders of magnitude. In addition, the ANN, LSTM, DCGAN and VAE-GAN models have demonstrated to provide efficient and accurate forecasts for a long lead-time in applications of streamflow, flooding and ozone forecasting in spatial and temporal spaces. Finally, the two-hybrid forecast models (DCGAN-EnKF and ConvLSTM-EnKF) are able to yield long lead-time forecasts of dynamic states and the use of EnKF in ConvLSTM and DCGAN models successfully corrects online model errors and significantly improves the real-time forecasting of dynamic systems.Open Acces
Boosting the Generalizability and Fairness of Presentation Attack Detection
No full textThe vulnerability of biometric recognition to presentation attacks (PAs) has been widely recognized and has attracted increasing attention as it enables attackers to impersonate authentic users. Presentation attack detection (PAD), aiming at automatically catching PAs, is an essential technology to secure biometric systems from PAs such as printed photos and replayed videos. Despite the considerable exploration and remarkable progress in PAD performance, two major issues still constitute a gap in technology. The first is the lack of proper understanding of the fairness of such algorithms over human-related attributes, and the second is the low performance generalizability over variabilities such as unknown attack types and capture environments. These challenges drive the main contributions of this thesis towards analyzing and boosting the fairness and generalizability of PAD.
PAD fairness over different human attributes is extremely understudied. Such underexploration is mainly due to the lack of suitable data. Towards enabling the fairness assessment and enhancement in face PAD, this thesis first introduces a combined attribute annotated PAD dataset, including both demographic and non-demographic attribute labels. Meanwhile, this thesis presents a new metric, accuracy balanced fairness, to simultaneously represent both the PAD fairness and the absolute PAD performance. Then, a comprehensive analysis of fairness in face PAD is conducted to study its relation to the nature of training data and the methodology of decision threshold selection. Guided by the outcomes of these analyses, a data augmentation method, namely FairSWAP, is successfully proposed to enhance the fairness of face PAD.
In addition to the PAD generalizability over human-related attributes, seen as fairness, another emerging challenge that encountered face PAD during the COVID-19 pandemic is the PAD generalizability to subjects wearing facial masks. To address this issue, this thesis first provides a collaborative real mask attack dataset involving the conventional unmasked bona fide and attacks, masked bona fide sample, novel attacks with faces wearing masks, and attacks with real masks placed on spoof faces. This thesis performs a set of extensive experiments to investigate the impact of masked faces on recognition vulnerability and PAD behaviour. Observing the degradation of PAD performance caused by the facial masks, this thesis presents a solution to target this issue by refining the partial attack supervision and the regional weighted inference.
The third part of this thesis targets the more conventional PAD generalizability issues, such as variabilities in attack creation and capture scenarios. Aiming to boost the generalizability of face PAD, this thesis proposes to leverage the information from the frequency domain in an optimized manner, assisting the information in the spatial domain to learn a more generalized representation under intra-dataset and cross-dataset settings. With a focus on enhancing the generalizability of iris PAD, this thesis proposes a micro-stripe analyses solution that leverages the benefit of the spatially aware processing of well-defined regions in the iris and its border with the sclera. This thesis further introduces a novel attention-based deep pixel-wise binary supervision method, A-PBS, for iris PAD. This solution aims to capture the fine-grained pixel/patch-level attack clues and automatically locate regions that contribute the most to an accurate PAD decision. The generalizability of the proposed iris PAD solutions is demonstrated under real-world cross-testing cases, including cross-attack, cross-dataset, and cross-spectrum settings.
To summarize, this thesis first provides a much-needed comprehensive analysis of fairness in PAD, leading to a well-founded and integrable fairness enhancement solution. Then, it presents detailed investigations of the masked face PAD challenge along with a technical solution towards improving the masked face PAD performance. The thesis then presents a set of novel contributions to boost the generalizability of face and iris PAD techniques. This thesis thus yields practically-aware advancements in understanding and mitigating vulnerabilities of biometric systems and lays the groundwork for future research into developing and deploying generalized PAD systems
Safety and Efficacy of Cryoballoon Pulmonary Vein Isolation and Left Atrial Appendage Closure Combined Procedure and Half-Dose Rivaroxaban After Operation in Elderly Patients with Atrial Fibrillation
No full textXiaogang Zhang,* Zhongying Xing,* Chao Fang, Meiling Yang, Jun Luo, Zhongping Ning Department of Cardiology, Shanghai Pudong New Area Zhoupu Hospital (Shanghai Health Medical College Affiliated Zhoupu Hospital), Shanghai, 201318, People’s Republic of China*These authors contributed equally to this workCorrespondence: Zhongping Ning, Department of Cardiology, Shanghai Pudong New Area Zhoupu Hospital (Shanghai Health Medical College Affiliated Zhoupu Hospital), No. 1500 Zhouyuan Road, Pudong New District, Shanghai, 201318, People’s Republic of China, Tel +86-021-68135590, Email [email protected]: To investigate the safety and effectiveness of cryo-balloon pulmonary vein isolation (PVI) and left atrial appendage closure (LAAC) combined procedure and half-dose rivaroxaban after operation in elderly patients with atrial fibrillation (AF).Patients and Methods: A total of 203 AF patients presented for cryo-balloon PVI, and LAAC combined procedure was included from 2019 to 2021. Postoperative patients were anticoagulated with rivaroxaban with/without clopidogrel for 60 days, with oral rivaroxaban of 10 mg in the elderly group and 20 mg in the non-elderly group. Patients with AF ≥ 80 and 0.05).Conclusion: This study suggests that cryo-balloon PVI and LAAC combined procedure and half-dose rivaroxaban after the operation is safe and effective in treating elderly patients with AF.Keywords: pulmonary vein isolation, left atrial appendage closure, cryo-balloon, rivaroxaban, atrial fibrillation, elderly patient
Phylogenetic relationships and estimation of divergence times among Sisoridae catfishes
No full textNineteen taxa representing 10 genera of Sisoridae were subjected to phylogenetic analyses of sequence data for the nuclear genes Plagl2 and ADNP and the mitochondrial gene cytochrome . The three data sets were analyzed separately and combined into a single data set to reconstruct phylogenetic relationships among Chinese sisorids. Both Chinese Sisoridae as a whole and the glyptosternoid taxa formed monophyletic groups. The genus is likely to be the earliest diverging extant genus among the Chinese Sisoridae. The four species included in the study formed a monophyletic group. was indicated to be earliest diverging glyptosternoid, followed by and . Our data supported the conclusion that and both formed a monophyletic group. On the basis of the fossil record and the results of a molecular dating analysis, we estimated that the Sisoridae diverged in the late Miocene about 12.2 Mya. The glyptosternoid clade was indicated to have diverged, also in the late Miocene, about 10.7 Mya, and the more specialized glyptosternoid genera, such as , originated in the Pleistocene (within 1.9 Mya). The speciation of glyptosternoid fishes is hypothesized to be closely related with the uplift of the Qinghai-Tibet Plateau.Nineteen taxa representing 10 genera of Sisoridae were subjected to phylogenetic analyses of sequence data for the nuclear genes Plagl2 and ADNP and the mitochondrial gene cytochrome . The three data sets were analyzed separately and combined into a single data set to reconstruct phylogenetic relationships among Chinese sisorids. Both Chinese Sisoridae as a whole and the glyptosternoid taxa formed monophyletic groups. The genus is likely to be the earliest diverging extant genus among the Chinese Sisoridae. The four species included in the study formed a monophyletic group. was indicated to be earliest diverging glyptosternoid, followed by and . Our data supported the conclusion that and both formed a monophyletic group. On the basis of the fossil record and the results of a molecular dating analysis, we estimated that the Sisoridae diverged in the late Miocene about 12.2 Mya. The glyptosternoid clade was indicated to have diverged, also in the late Miocene, about 10.7 Mya, and the more specialized glyptosternoid genera, such as , originated in the Pleistocene (within 1.9 Mya). The speciation of glyptosternoid fishes is hypothesized to be closely related with the uplift of the Qinghai-Tibet Plateau
Enhanced piezoelectric energy harvesting powered wireless sensor nodes using passive interfaces and power management approach
No full textLow-frequency vibrations typically occur in many practical structures and
systems when in use, for example, in aerospaces and industrial machines.
Piezoelectric materials feature compactness, lightweight, high integration
potential, and permit to transduce mechanical energy from vibrations into
electrical energy. Because of their properties, piezoelectric materials have been
receiving growing interest during the last decades as potential vibration-
harvested energy generators for the proliferating number of embeddable
wireless sensor systems in applications such as structural health monitoring
(SHM). The basic idea behind piezoelectric energy harvesting (PEH) powered
architectures, or energy harvesting (EH) more in general, is to develop truly “fit
and forget” solutions that allow reducing physical installations and burdens to
maintenance over battery-powered systems. However, due to the low
mechanical energy available under low-frequency conditions and the relatively
high power consumption of wireless sensor nodes, PEH from low-frequency
vibrations is a challenge that needs to be addressed for the majority of the
practical cases. Simply saying, the energy harvested from low-frequency
vibrations is not high enough to power wireless sensor nodes or the power
consumption of the wireless sensor nodes is higher than the harvested energy.
This represents a main barrier to the widespread use of PEH technology at the
current state of the development, despite the advantages it may offer.
The main contribution of this research work concerns the proposal of a novel
EH circuitry, which is based on a whole-system approach, in order to develop
enhanced PEH powered wireless sensor nodes, hence to compensate the
existing mismatch between harvested and demanded energy. By whole-system
approach, it is meant that this work develops an integrated system-of-systems
rather than a single EH unit, thus getting closer to the industrial need of a ready-
to-use energy-autonomous solution for wireless sensor applications such as
SHM. To achieve so, this work introduces:
Novel passive interfaces in connection with the piezoelectric harvester
that permit to extract more energy from it (i.e., a complex conjugate
impedance matching (CCIM) interface, which uses a PC permalloy
toroidal coil to achieve a large inductive reactance with a centimetre-
scaled size at low frequency; and interfaces for resonant PEH
applications, which exploit the harvester‟s displacement to achieve a
mechanical amplification of the input force, a magnetic and a mechanical
activation of a synchronised switching harvesting on inductor (SSHI)
mechanism).
A novel power management approach, which permits to minimise the
power consumption for conditioning the transduced signal and optimises
the flow of the harvested energy towards a custom-developed wireless
sensor communication node (WSCN) through a dedicated energy-aware
interface (EAI); where the EAI is based on a voltage sensing device
across a capacitive energy storage.
Theoretical and experimental analyses of the developed systems are carried
in connection with resistive loads and the WSCN under excitations of low
frequency and strain/acceleration levels typical of two potential energy-
autonomous applications, that are: 1) wireless condition monitoring of
commercial aircraft wings through non-resonant PEH based on Macro-Fibre
Composite (MFC) material bonded to aluminium and composite substrates; and
wireless condition monitoring of large industrial machinery through resonant
PEH based on a cantilever structure.
shown that under similar testing conditions the developed systems feature a
performance in comparison with other architectures reported in the
literature or currently available on the market. Power levels up to 12.16 mW and
116.6 µW were respectively measured across an optimal resistive load of 66
277 kΩ for an implemented non-resonant MFC energy harvester on
aluminium substrate and a resonant cantilever-based structure when no
interfaces were added into the circuits. When the WSCN was connected to the
harvesters in place of the resistive loads, data transmissions as fast as 0.4 and
s were also respectively measured. By use of the implemented passive
interfaces, a maximum power enhancement of around 95% and 452% was
achieved in the two tested cases and faster data transmissions obtained with a
maximum percentage improvement around 36% and 73%, respectively. By the
use of the EAI in connection with the WSCN, results have also shown that the
overall system‟s power consumption is as low as a few microwatts during non-
active modes of operation (i.e., before the WSCN starts data acquisition and
transmission to a base station).
Through the introduction of the developed interfaces, this research work takes a
whole-system approach and brings about the capability to continuously power
wireless sensor nodes entirely from vibration-harvested energy in time intervals
of a few seconds or fractions of a second once they have been firstly activated.
Therefore, such an approach has potential to be used for real-world energy-
autonomous applications of SHM
- …
