Search

Nowadays, the steady-state somatosensory evoked potential (SSSEP)-based brain-computer interfaces (BCIs) has been developed for improving the quality of daily life for people with physical disabilities. However, due to its poor performance of recognizing selective attention tasks and inattention(rest)-state, the SSSEP-based BCI has not been widely used for practical interfaces. In this paper, we propose a mutual information-based spatio-spectral feature selection method for recognizing selective attention tasks and inattention(rest)-state using dry electrodes considering a real-life application, when vibration stimuli were applied to both index fingers. In our methods, the filter-bank common spatial pattern (FBCSP) was used for extracting spatio-spectral features of the SSSEP. Then, discriminative features were selected using a mutual information-based best individual feature (MIBIF) algorithm. The regularized linear discriminant analysis (RLDA) used as the classifier. The feasibility of the proposed method was demonstrated through eight healthy subjects using the vibration stimuli induced SSSEP with spatially clear and distinguishable patterns for SSSEP-based BCI. From our study, the proposed method showed the best classification accuracy with a kappa value of 0.35±0.17. Furthermore, based on the ANOVA with posthoc tests, the proposed method showed significantly higher accuracy as 57.9% in decoding three classes (p -value

In various injection molding manufacturing plants, there are many difficulties in detecting defective products during production. Since there are limitations in detecting product defects with the human eye, this paper proposes a framework for detecting product defects in a human-free manufacturing environment. We detect product defects using Canny Edge Detection, a powerful edge detector, and provide reliability of products detected using Mask R-CNN, a neural network with excellent speed and accuracy. As the network, the ResNet101 network with the highest accuracy was selected, and the network was used as the backbone network of Mask R-CNN, and the image was resized and sized using LEDs when shooting to detect even small scratches.

In this paper, we propose a process for collecting, storing, monitoring and preprocessing data in Node-RED, an open platform, using IoT devices in a smart factory injection manufacturing environment. In the current manufacturing environment, there is a problem due to the server load caused by big data. In order to solve this problem, studies on various aspects of data processing are being conducted. We propose a data processing process that processes data from IoT devices and delivers them to the server, rather than a method that processes all data collected in a smart factory manufacturing environment on a server.

In recent years, the steady-state somatosensory evoked potential (SSSEP)-based brain-computer interfaces (BCIs) have been developed for assisting people with physical disabilities. However, due to its no well-designed standard tactile vibration stimulator to elicit the SSSEP, the SSSEP-based BCI has not been widely used to real-world applications. In this study, we aim to SSSEP-based BCI with high-frequency vibration stimuli. In our experiments, the stimuli were constructed using the coin vibration motor which controlled by the Arduino. The feasibility of the high-frequency vibration stimuli to elicit SSSEP was validated with SSSEP data from two healthy participants. For data collection, the vibration stimuli were attached on the index finger of the left and right hand. And the participant concentrated on one of the vibration stimuli following the visual commands. For the data analysis, the spatial features were extracted from the acquired EEG signals using the common spatial pattern (CSP) filtering. The regularized linear discriminant analysis (RLDA) also used as a classifier. The experimental results showed that the user's selective attention to one of the high-frequency vibration stimuli was classified with averaged 61.7±9.4%. Based on the experimental results, we can conclude that the SSSEP can elicit by the high-frequency vibration stimuli and it can be used to the BCI.

In this study, we propose a real-time brain-computer interface (BCI)-based control system for lower limb exoskeleton. Voluntarily induced electroencephalogram (EEG) signal during gait and sit motor imagery (MI) was decoded and translate into the exoskeleton as control commands in real-time. While it is hard to tell when the user engages in the conventional BCI control system, the EEG signal via the user's consecutive eyeblink was used as an initiation of MI in the proposed semi-synchronous control system; hence the EEG decoder is able to denote the transition point. This semi-synchronous BCI control system combines the advantages of asynchronous and synchronous BCI systems by understanding the user's MI environment for robust exception handling. The classifier's flow diagram of the presented BCI control system was described as a finite state machine.

Autonomous driving technologies (ADTs) are transformative because they are expected to assume the task of driving. Since travel distance is closely related to the burden of driving, it is expected to affect the ADT preference; however, this effect has not been investigated in the transportation literature. Therefore, this study is designed to investigate the ADT preference and heterogeneity with respect to different travel distances. Hypothetical choice situations were designed to investigate people's ADT preferences over human driving for varying travel distances. Using the stated preference data collected via the experiments, four latent class models were estimated for heterogeneity across classes and travel distances. The estimation results show that three classes, we labeled as competitive, autonomous vehicle oriented, human-driven vehicle oriented, are revealed by the ADT preference, regardless of the travel distance. In addition, in-vehicle activities in hypothetical autonomous driving situations were observed across all the classes and travel distances.

The purpose of this study was to develop sit-to-stand movement intention decoding algorithms based on brain responses to the vibrotactile stimulation. Specifically, we studied a simultaneous hybrid brain-computer interface (BCI) by combining steady-state somatosensory evoked potential (SSSEP) and a motor imagery (MI) task. In our BCI system, a user could generate two possible commands by concentrating on one of two vibration stimuli, which were attached to the left and right hand. The statistical method based on the mutual information between the spatial-temporal patterns was used to detect the user’s intention of sit or stand from the electroencephalography (EEG) signals. The results of our offline experiments demonstrated the feasibility of hybrid MI-SSSEP based BCI system for decoding sit-to-stand movement intention. It is expected that the proposed method and algorithms can be implanted in the future to control the lower limb exoskeleton robot.

In this paper, an innovative modeling approach for Li-ion battery packs is proposed by considering intrinsic cell unbalances and packaging elements. The proposed modeling method shows that the accurate battery pack model can be achieved if the overall influences of intrinsic cell unbalances and packaging elements are taken account. Concurrently, the proposed method takes a practical model structure, resulting in the reduction of computational burden in a battery management system. Furthermore, because the proposed method utilizes cell information without a manufactured battery pack, it can be helpful to design optimal battery packs. The proposed method is verified through simulation and experimental results of the Li-ion battery pack along with the battery cycler. In three test profiles, the mean absolute percentage errors and root mean square errors of the proposed pack model do not exceed 0.5% and 0.07 V, respectively.

Tensor completion is the task of completing multi-aspect data represented as a tensor by accurately predicting missing entries in the tensor. It is mainly solved by tensor factorization methods, and among them, Tucker factorization has attracted considerable interests due to its powerful ability to learn latent factors and even their interactions. Although several Tucker methods have been developed to reduce the memory and computational complexity, the state-of-the-art method still 1) generates redundant computations and 2) cannot factorize a large tensor that exceeds the size of memory. This paper proposes FTcom, a fast and scalable Tucker factorization method for tensor completion. FTcom performs element-wise updates for factor matrices based on coordinate descent, and adopts a novel caching algorithm which stores frequently-required intermediate data. It also uses a tensor file for disk-based data processing and loads only a small part of the tensor at a time into the memory. Experimental results show that FTcom is much faster and more scalable compared to all other competitors. It significantly shortens the training time of Tucker factorization, especially on real-world tensors, and it can be executed on a billion-scale tensor which is bigger than the memory capacity within a single machine.

Unintended acceleration car accident which is hard to be regarded as human errors is examined. We investigated the accident car to determine the cause upon being requested to investigate by the police. From the car black box, we could narrow down the reason of the acceleration. As one of the causes of unintended acceleration we assumed that the restoration failure of the pedal interfered with by improper use of a floor mat and could reconstruct this. We discussed the acceleration values ​​of the accident vehicle and paid attention to the floor mat used in it. The results indicate that the negligible physical interference in the accelerator linkage could cause a devastating accident.

This paper presents an application of an optimal smoothing filter for moving target tracking problem based on measured noise source. In order to measure distance and velocity for the moving target, a beamforming method is applied to use the noise source by using microphone array. Also a Kalman filter and an optimal smoothing algorithm are adopted to improve accuracy of trajectory estimation by using a Singer target model. The simulation is conducted with a missile dynamics to verify performance of the optimal smoothing filter, and a model rocket is used for experiment environment to compare the trajectory estimation results between the beamforming, the Kalman filter, and the smoother. The Kalman filter results show better tracking performance than the beamforming technique, and the estimation results of the optimal smoother outperform the Kalman filter in terms of trajectory accuracy in the experiment results.

Transmission control protocols have to overcome common problems in wireless networks. TCP employing both packet loss discrimination mechanism and available bandwidth estimation algorithm, known as the good existing solution, shows significant performance enhancement in wireless networks. For instance, TCP New Jersey which exhibits high throughput in wireless networks intends to improve TCP performance by using available bandwidth estimation and congestion warning. Even though it achieves 17% and 85% improvements in terms of goodput over TCP Westwood and TCP Reno, respectively, we further improve it by exploring maximized available bandwidth estimation, handling bit-error-rate error recovery, and effective adjustment of sending rate for retransmission timeout. Hence, we propose TCP NJ+, showing that for up to 5% packet loss rate, it outperforms other TCP variants by 19% to 104% in terms of goodput when the network is in bi-directional background traffic.

A high-performance fundamental frequency estimator inspired by the human auditory system is presented. It uses a cubic-root transform and contrast enhancement on the spectral domain to enhance weak harmonics and suppress noise. The proposed weighted unbiasing process reduces multiple or submultiple pitch errors efficiently. Evaluation is performed under the presence of white noise, babble noise, and telephone channel distortion. The results show that the proposed method performs equivalent to or better than previous state-of-the-art methods in the majority of the noise and channel conditions.

Wireless Sensor Networks(WSN) are used to communicate between Sensor nodes. The ZigBee telecom applications profile specification is announced for a voice communications in the ZigBee networks by the ZigBee Alliance since 2010[1,2], but it has not been considered to interface with the long distance communications. We designed Voice over Sensor Networks(VoSN) considered on long distance communications with each other nodes.

This study presents an application of acoustic source localization using microphone array installed on a drone. Although recent technology of drone enables to record audio and video, there are still acoustic problems such as engine noise, propeller-induced flow noise, microphone directivity, etc. Such unwanted noise shows strong tonal characteristics with broad band noise contents. To acquire usable acoustic signal from array measurements while a drone is flying, we implemented acoustic signal conditioning and array processing. The approach to the problem is to adapt beamforming and TDOA techniques that provide localization capability of multiple sources and the position or direction of arrival of source, respectively. The array configuration and angle resolution are adjusted depending on the application requirements. We limit our attention to stable hovering condition to rule out position and heading error occurred by the drone motion. The experimental results of position estimation under a various noise...

Transport Control Protocol (TCP), the mostly used transport protocol, performs well over wired networks. As much as wireless network is deployed, TCP should be modified to work for both wired and wireless networks. Since TCP is designed for congestion control in wired networks, it cannot clearly detect non-congestion related packet loss from wireless networks. TCP-Westwood presents one solution to this problem for wireless networks, but it provides worse performance over wired networks. To overcome this problem, we propose a new algorithm to increase the slow start threshold in TCP-Westwood. Also, we simulate the proposed algorithm with the NS-2 simulator by comparing it to the original TCP-Westwood for both wireless and wired networks.

The advancement of mobile communications for the last few years has accompanied an increasingly extensive and diverse array of environments for TCP applications, which has led to a host of wireless TCP approaches that may provide more stability and ensure higher performance. In a heterogeneous network, as opposed to conventional wireless networks, a mobile node performs a handoff to another network cells with a different bandwidth and latency. Conventional wireless TCP schemes, however, are not capable of properly addressing sudden changes in round trip time (RTT) and packet loss resulting from handoffs within a heterogeneous network and do experience such problems as the waste of available bandwidth and frequent packet loss. As a solution to these problems, this paper proposes Freeze TCP version 2 (v2), an enhancement of Freeze TCP designed to dynamically obtain the available bandwidth in a new network cell. Comprehensive simulation study is conducted by using a network simulator, ns-2, to compare the proposed scheme to a few other schemes, such as Freeze TCP, DEMO-Vegas, and TCP Vegas, in a heterogeneous network environment with vertical handoffs in terms of throughput per speed of the mobile node and per bit error rate of the wireless link.

TCP has been used as the standard for reliable transmission of application traffic in the internet. Recently, many kinds of wireless TCP schemes, such as TCP Westwood, have been developed with the emergence of wireless networks. Various algorithms have been suggested that are aimed at improving its performance in wireless networks. In this paper, we focus on the spurious timeouts that occur in such networks. E-RTO efficiently treats delay spikes that come from the retransmission of lost packets by the reliable wireless link layer. Also, it efficiently handles delay spikes with segment losses. However, it does not address the mobility of the user in the network. We tried to give a solution to this problem by proposing an BCR algorithm, and we evaluated the performance of BCR through simulation using NS-2. Through the experiment, we showed that the throughput and goodput of BCR are better than those of E-RTO.

In this work, we are interested in tracking a rotating sound source using a Kalman filtering technique based on a set of non-linear time difference of arrival (TDOA) measurements. Array of microphones measure acoustic signal emitted from a rotating source and the TDOA estimates are calculated followed by a solution for hyperbolic position fix. The position estimation of sound source based on TDOA is a popular technique in source localization. The method involves calculation of a set of nonlinear equations and poor accuracy of TDOA estimates often results in inaccuracy in location. In this work, the range difference is expressed by a model movement on which a recursive extended Kalman Filter has been developed. The TDOA measurements optimize the estimated values which are reduced as observation in extended Kalman filtering algorithm. Location estimation is updated from TDOA measurements along with the time history data. The Cramer-Rao Lower Bound (CRLB) is derived and simulations are compared. [Work supported by National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (No. 2010-0014978).]

The localization of sound sources in a three dimensional space has been recognized as an important research topic in acoustics. Many acoustic measurement techniques in three-dimensional space have used spherical shape of array to optimize sensor configuration and increase its performance and spatial resolution. In this study, the design of microphone array on polyhedral surfaces is proposed to introduce an easy way of building arrays and to achieve the measurement capability for all direction as a spherical array does. Five symmetrical polyhedrons are used and the simulation results are compared. MEMS microphones are used in arrays and the tests are made in an anechoic environment to validate the performance of arrays. MEMS sensor makes it easy to build a three-dimensional array and enables to implement large number of sensors in a small area. Sensors are built on a printed circuit board (PCB) which becomes a sub-array of polyhedrons. The array output is processed using conventional beamforming method to ...

Recent development of signal processing technologies has encouraged the aero‐acousticians to implement the state‐of‐the‐art sensors and data acquisition devices. They sometimes cost too much; however, especially for the academic arenas, the efforts of reducing the expense of integrating a system and increasing performance of phased array have initiated the idea of designing the synthetic aperture microphone array. The application of beamforming technique in localizing rotor blade noise and the use of synthetic aperture array in wind tunnel testing are presented. To apply the synthetic aperture beamforming on rotating objects, three steps should be considered. The first step is to restore the original acoustic waveform radiated from a rotating sound source. The second step is to synthesize independent array measurement data by correcting phase information of the received acoustic signals. Finally, the beamformer calculates the power on each point of grid in the space. The paper describes the details of dep...

Background Nonnegative matrix factorization (NMF) is a feature extraction method that has the property of intuitive part-based representation of the original features. This unique ability makes NMF a potentially promising method for biological sequence analysis. Here, we apply NMF to fold recognition and remote homolog detection problems. Recent studies have shown that combining support vector machines (SVM) with profile-profile alignments improves performance of fold recognition and remote homolog detection remarkably. However, it is not clear which parts of sequences are essential for the performance improvement. Results The performance of fold recognition and remote homolog detection using NMF features is compared to that of the unmodified profile-profile alignment (PPA) features by estimating Receiver Operating Characteristic (ROC) scores. The overall performance is noticeably improved. For fold recognition at the fold level, SVM with NMF features recognize 30% of homolog proteins at > 0.99 ROC scores, while original PPA feature, HHsearch, and PSI-BLAST recognize almost none. For detecting remote homologs that are related at the superfamily level, NMF features also achieve higher performance than the original PPA features. At > 0.90 ROC50 scores, 25% of proteins with NMF features correctly detects remotely related proteins, whereas using original PPA features only 1% of proteins detect remote homologs. In addition, we investigate the effect of number of positive training examples and the number of basis vectors on performance improvement. We also analyze the ability of NMF to extract essential features by comparing NMF basis vectors with functionally important sites and structurally conserved regions of proteins. The results show that NMF basis vectors have significant overlap with functional sites from PROSITE and with structurally conserved regions from the multiple structural alignments generated by MUSTANG. The correlation between NMF basis vectors and biologically essential parts of proteins supports our conjecture that NMF basis vectors can explicitly represent important sites of proteins. Conclusion The present work demonstrates that applying NMF to profile-profile alignments can reveal essential features of proteins and that these features significantly improve the performance of fold recognition and remote homolog detection.