- BB
Authors: Zhang, H.; Advisor: -; Participants: Fu, Z.; Shu, K. (2019) - With the development of Internet of Things (IoT) technology and various sensing technologies, some newer ways of perceiving people and the environment have emerged. Commercial wearable sensing devices integrate a variety of sensors that can play a signi cant role in motion capture and behavioral analysis. This paper proposes a solution for recognizing human motion in ping-pong using a commercial smart watch. We developed a data acquisition system based on the IoT architecture to obtain data relating to areas such as acceleration, angular velocity, and magnetic induction of the watch. Based on the features of the extracted data, experiments were performed using major machine learning classi cation algorithms including k-nearest neighbor, support vector machine, Naive Bayes, logistic ...
|
- BB
Authors: Xu,X.; Advisor: -; Participants: Zhang, Y.; Tang, M.; Gu, H.; Yan, S.; Yang, J. (2019) - Corresponding to the continual development of human-computer interaction technology, the use of emotional computing (EC) is gradually emerging in the Internet of Things (IoT). Emotion recognition is considered a highly valuable aspect of EC. Numerous studies have examined emotion recognition based on electroencephalogram (EEG) signals, but the recognition rate is unreliable. In this paper, a feature extraction method is proposed that is based on double tree complex wavelet transform (DTCWT) and machine learning. The emotions of 16 subjects are induced under video stimulation, and the original signal is acquired using a Neuroscan device. Both EEG and electromyography (EMG) signal are then eliminated by band-pass ltering, and the reconstructed signal of each frequency band is obtaine...
|
- BB
Authors: Alves, J. M.; Advisor: -; Participants: Honório, L. M.; Capretz, M. A. M. (2019) - Internet of Things (IoT) applications generate vast amounts of real-time data. Temporal analysis of these data series to discover behavioural patterns may lead to quali ed knowledge affecting a broad range of industries. Hence, the use of machine learning (ML) algorithms over IoT data has the potential to improve safety, economy, and performance in critical processes. However, creating ML work ows at scale is a challenging task that depends upon both production and specialized skills. Such tasks require investigation, understanding, selection, and implementation of speci c ML work ows, which often lead to bottlenecks, production issues, and code management complexity and even then may not have a nal desirable outcome. This paper proposes the Machine Learning Framework for IoT data ...
|
- BB
Authors: Kumar, S.; Advisor: -; Participants: Singh, K.; Kaiwartya, O.; Cao, Y.; Zhou, H. (2019) - Recently, Internet of vehicles (IoV) has witnessed signi cant research and development attention in both academia and industries due to the potential towards addressing traf c incidences and supporting green mobility. With the growing vehicular network density, jamming signal centric security issues have become challenging task for IoV network designers and traf c applications developers. Global positioning system (GPS) and roadside unit (RSU) centric related literature on location-based security approaches lacks signal characteristics consideration for identifying vehicular network intruders or jammers. In this context, this paper proposes a machine learning oriented as Delimitated Anti Jamming protocol for vehicular traf c environments. It focuses on jamming vehicle's discriminate...
|
- BB
Authors: Kasongo, S. M.; Advisor: -; Participants: Sun, Y. (2019) - In recent years, the increased use of wireless networks for the transmission of large volumes of information has generated a myriad of security threats and privacy concerns; consequently, there has been the development of a number of preventive and protective measures including intrusion detection systems (IDS). Intrusion detection mechanisms play a pivotal role in securing computer and network systems; however, for various IDS, the performance remains a major issue. Moreover, the accuracy of existing methodologies for IDS using machine learning is heavily affected when the feature space grows. In this paper, we propose a IDS based on deep learning using feed forward deep neural networks (FFDNNs) coupled with a lter-based feature selection algorithm. The FFDNN-IDS is evaluated usin...
|
- BB
Authors: Liu, Y.; Advisor: -; Participants: He, C.; Li, X.; Zhang, C.; Tian, C. (2019) - In this paper, we investigate the great potential of the combination of machine learning technology and wireless communications. Currently, many researchers have proposed various optimization algorithms on resource allocation for distributed antenna systems (DASs). However, the existing methods are mostly hard to implement because of their high computational complexity. In this paper, a new system model for machine learning is considered for the scenario of DAS, which is more practical with its low computational complexity. We utilize the k-nearest neighbor (k-NN) algorithm based on the database of a traditional sub-gradient iterative method to get a power allocation scheme for DAS. The simulation results show that our k-NN algorithm can also obtain the power distribution scheme whi...
|
- BB
Authors: Koohestani, A.; Advisor: -; Participants: Abdar, M.; Khosravi, A.; Nahavandi, S.; Koohestani, M. (2019) - The level of consciousness and the concentration of drivers while driving play a vital role for reducing the number of accidents. In recent decade, in-vehicle infotainment (IVI) [or in-car entertainment (ICE)] is one of the main reasons that lead to degradation of drivers performance and losing awareness. However, the impacts of some other reasons, such as drowsiness and driving fatigue, are entirely important as well. Hence, early detection of such performance degradation using different methods is a very hot research domain. To this end, the data set is collected using two different simulated driving scenarios: normal and loaded drive (17 elderly and 51 young/35 male and 33 female). This paper, therefore, concentrates on driving performance analysis using various machine learning ...
|
- BB
Authors: Strodthoff, N.; Advisor: -; Participants: Göktepe,B.; Schierl, T.; Hellge, C.; Samek, W. (2019) - We investigate Early Hybrid Automatic Repeat reQuest (E-HARQ) feedback schemes enhanced by machine
learning techniques as a path towards ultra-reliable and lowlatency communication (URLLC). To this end, we propose machine learning methods to predict the outcome of the decoding process ahead of the end of the transmission. We discuss different input features and classification algorithms ranging from traditional methods to newly developed supervised autoencoders. These methods are evaluated based on their prospects of complying with the URLLC requirements of effective block error rates below 10 at small latency overheads. We provide
realistic performance estimates in a system model incorporating scheduling effects to demonstrate the feasibility of E-HARQ across different signal-to-n...
|
- BB
Authors: Decaro, C.; Advisor: -; Participants: Montanari, G.B.; Molinari, R.; Gilberti, A.; Bagnoli, D.; Bianconi, M.; Bellanca, G. (2019) - Objective: This paper shows the application of machine learning techniques to predict hematic parameters using blood visible spectra during ex-vivo treatments. Methods: A spectroscopic setup was prepared for acquisition of blood absorbance spectrum and tested in an operational environment. This setup is non invasive and can be applied during dialysis sessions. A support vector machine and an arti cial neural network, trained with a dataset of spectra, have been implemented for the prediction of hematocrit and oxygen saturation. Results & Conclusion: Results of different machine learning algorithms are compared, showing that support vector machine is the best technique for the prediction of hematocrit and oxygen saturation.
|
- BB
Authors: Sun, P.; Advisor: -; Participants: Wang, D.; Mok, V. C.; Shi, L. (2019) - Radiomics-based researches have shown predictive abilities with machine-learning approaches. However, it is still unknown whether different radiomics strategies affect the prediction performance. The aim of this study was to compare the prediction performance of frequently utilized radiomics feature selection and classi cation methods in glioma grading. Quantitative radiomics features were extracted from tumor regions in 210 Glioblastoma (GBM) and 75 low-grade glioma (LGG) MRI subjects. Then, the diagnostic performance of sixteen feature selection and fteen classi cation methods were evaluated by using two different test modes: ten-fold cross-validation and percentage split. Balanced accuracy and area under the curve (AUC) of the receiver operating characteristic were used to evalu...
|
