• Journal of Biomedical Engineering Research
• /
• v.21 no.5
• /
• pp.469-475
• /
• 2000

신경신호의 계측은 신경계의 연구에 필수적인 도구로 최근 반도체미세전극기술 등 수십, 수백개의 채널로부터 신경신호를 기록할 수 있는 방법들이 발달함에 따라 많은 수의 뉴런으로부터 신경 신호를 측정하여 컴퓨터로 그 신호를 처리할 수 있는 시스템의 필요성은 더욱 커지고 있다. 본 연구에서는 최대 16채널의 신경신호를 실시간에 측정하여 기록하고, 저장된 신호로부터 활동전위를 검출하며, 단일 뉴런들로부터의 신호를 분류하여 spike train의 형태로 저장한 뒤 여러 뉴런들간의 상관관계를 분석하기 위한 spike train 해석이 가능한 시스템을 개발하였다. 이 시스템은 보통사양의 PC이외에는 단지 신호획득보드만을 포함하여 다채널미세전극으로부터 뉴런의 신호를 측정, 증폭하여 호스트PC로 전송하고 저장하며 이로부터 활동전위를 검출하여 단일뉴런으로부터의 spike train으로 분류할 수 있다. 또한 저장된 spike train들로부터 신경회로망을 이루는 여러뉴런 들간의 관계를 분석하여 기능들이 시스템에 포함되어있다. 개발된 시스템을 사용하여 개구리 감각 신경의 신호를 실시간에 동시기록하여 활동전위을 검출하고 특징추출방법과 principal component analysis를 이용하여 분류한 뒤 spike train 해석을 수행하였다.

• Proceedings of the KOSOMBE Conference
• /
• v.1990 no.11
• /
• pp.114-117
• /
• 1990

EEG signal is analyzed by two methods, analysis by visual inspection of EEG recording sheets and analysis by quantative method. Generally visual inspection method is used in the clinical field. But this method has its limitation because EEG signal is random signal. Therefore it is necessary to analyze EEG signals quantatively to obtain more precise and objective information of neural and brain. In this paper, power spectrum of EEG signal was estimated by AR(AutoRegressive) model in the frequency domain. This process is useful as a preprocessing stage for tomographic brain mapping (TBM) at each frequency, band. As a method for estimating power spectral density of EEG signals, periodogram method, autocorrelation method. covariance method, modified covariance method, and Burg method are tested in this paper.

• Proceedings of the KOSOMBE Conference
• /
• v.1998 no.11
• /
• pp.147-148
• /
• 1998

In this paper, we developed the process for depth-probe type silicon microelectrode arrays. The process consists of four mask steps only. The steps are for defining sites, windows, and for shaping probe using plasma etch from above, and for shaping using wet etch from below, respectively. The probe thickness is controlled by dry etching, not by impurity diffusion. We used gold electrodes with a triple dielectric system consisting of oxide/nitride/oxide. The shank of the probe taper from 200um to tens of urn tip and has 30 um thickness.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.45 no.3
• /
• pp.53-60
• /
• 2008

Recently, various recording technologies are studied for optical data storage. After standardization of BD (Blue-ray Disc) and HD-DVD (High-Definition Digital Versatile Disc), the industry is looking for a suitable technology for next generation optical data storage. Super-RENS (Super-resolution near field structure) technique, which is capable of compatibility with other systems, is one of next optical data storage. In this paper, we analyze the nonlinearity of Super-RENS read-out signal through the bicoherence test, which uses HOS (Higher-Order Statistics) and apply neural networks for nonlinear modeling. The model structure considered in this paper is the NARX (Nonlinear AutoRegressive eXogenous) model. The experiment results indicate that the read-out signals have nonlinear characteristics. In addition, it verified the possibility that neural networks can be utilized for nonlinear modeling of Super-RENS systems.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.45 no.12
• /
• pp.90-96
• /
• 2008

Recently, various recording technologies are studied for optical data storage. After standardization of BD (Blu-ray Disc) and HD-DVD (High-Definition Digital Versatile Disc), the industry is looking for a suitable technology for next generation optical data storage. Super-RENS (Super-Resolution Near Field Structure) technique, which is capable of compatibility with other systems, is one of next optical data storage. In this paper, we proposed a neural network-based nonlinear equalizer (NNEQ) for Super-RENS discs. To mitigate the nonlinear ISI (Inter-Symbol Interference), we applied NARX (Nonlinear AutoRegressive eXogenous) which is a kind of neural networks. Its validity is tested with the RF signal samples obtained from a Super-RENS disc. The performance of the proposed equalizer is superior to the one without equalization and that of the Limit-EQ in terms of BER (Bit Error Rate).

• Proceedings of the KIEE Conference
• /
• 2002.11a
• /
• pp.257-259
• /
• 2002

In investigating the characteristics of a neural network, the use of planar microelectrode array shows several advantages over normal intracellular recording[1]. A transparent indium tin oxide(ITO) multi-electrode array(MEA) was fabricated and its top surface was insulated with photodefinable polyimide(HD-8001) except the exposed area for interfacing between the ITO electrodes and the neuronal cells. The exposed ITO electrodes were platinized in order to reduce the impedance between the electrodes and electrolyte. The one-minute platinization with $0.99nA/{\mu}m^2$ current density reduced the average impedance of the electrodes from $2.5M\Omega\;to\;90k\Omega$ at 1kHz in normal ringer solution. Cardiac cells were cultured on this MEA as a pilot study before neuron culture. The signals detected by the platinized electrodes had larger amplitudes and improved signal to noise ratio(SNR) compared to non-platinized electrodes. It is clear that microelectrodes need to have lower impedance to make reliable extracellular recordings, and thus platinization is essential part of MEA fabrication. Burst spike of cultured olfactory bulb was also detected with the MEA having platinized electrodes.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.36S no.1
• /
• pp.123-132
• /
• 1999

In this paper, a quadratic sigmoid neural equalizer(QSNE) is proposed to improve the performance of conventional neural equalizer in terms of bit error probability by using a quadratic sigmoid function as the activation function of neural networks. Conventional neural equalizers which have been used to compensate for nonlinear distortions adopt the sigmoid function. In the case of sigmoid neural equalizer, each neuron has one linear decision boundary. So many neurons are required when the neural equalizer has to separate complicated structure. But in case of the proposed QSNF and quadratic sigmoid neural decision feedback equalizer(QSNDFE), each neuron separates decision region with two parallel lines. Therefore, QSNE and QSNDFE have better performance and simpler structure than the conventional neural equalizers in terms of bit error probability. When the proposed QSNDFE is applied to communication systems and digital magnetic recording systems, it is an improvement of approximately 1.5dB~8.3dB in signal to moise ratio(SNR) over the conventional decision feedback equalizer(DEF) and neural decision feedback equalizer(NDFE). As intersymbol interference(ISI) and nonlinear distortions become severer, QSNDFE shows astounding SNR shows astounding SNR performance gain over the conventional equalizers in the same bit error probability.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39C no.9
• /
• pp.820-827
• /
• 2014

One of the basic problems in Wireless Sensor Networks (WSNs) is the localization of the sensor nodes based on the known location of numerous anchor nodes. WSNs generally consist of a large number of sensor nodes and recording the location of each sensor nodes becomes a difficult task. On the other hand, based on the application environment, the nodes may be subject to mobility and their location changes with time. Therefore, a scheme that will autonomously estimate or calculate the position of the sensor nodes is desirable. This paper presents an intelligent localization scheme, which is an artificial neural network (ANN) based localization scheme used to estimate the position of the unknown nodes. In the proposed method, three anchors nodes are used. The mobile or deployed sensor nodes request a beacon from the anchor nodes and utilizes the received signal strength indicator (RSSI) of the beacons received. The RSSI values vary depending on the distance between the mobile and the anchor nodes. The three RSSI values are used as the input to the ANN in order to estimate the location of the sensor nodes. A feed-forward artificial neural network with back propagation method for training has been employed. An average Euclidian distance error of 0.70 m has been achieved using a ANN having 3 inputs, two hidden layers, and two outputs (x and y coordinates of the position).

• Journal of Biomedical Engineering Research
• /
• v.22 no.2
• /
• pp.145-150
• /
• 2001

대뇌 피질에 삽입하여 깊이에 따라 신경 신호를 기록하기 위한 탐침형 반도체 미세전극 어레이(depth-type silicon microelectrode array, 일명 SNU probe)를 제작하였다. 붕소를 확산시켜 생성된 고농도 p-type doping된 p＋ 영역을 습식식각 정지점으로 사용하는 기존의 방법과 달리 실리콘 웨이퍼의 앞면을 건식식각하여 원하는 탐침 두께만큼의 깊이로 트렌치(trench)를 형성한 후 뒷면을 습식식각하는 방법으로 탐침 형태의 미세 구조를 만들었다. 제작된 반도체 미세전극 어레이의 탐침 두께는 30 $\mu\textrm{m}$이며 실리콘 건식식각을 위한 마스크로 6 $\mu\textrm{m}$ 두께의 LTO(low temperature oxide)를 사용하였다. 탐침의 두께는 개발된 본 공정을 이용해서 5~90 $\mu\textrm{m}$ 범위까지 쉽게 조절할 수 있었다. 탐침의 두께를 보다 쉽게 조절할 수 있게 됨에 따라 여러 신경조직에 필요한 다양한 구조의 반도체 미세전극 어레이를 개발할 수 있게 되었다. 본 공정을 이용해서 개발된 4채널 SUN probe를 사용하여 흰쥐의 제1차 체감각 피질에서 4채널 신경 신호를 동시에 기록하였으며, 전기적 특성검사에서 기존의 탐침형 반도체 미세전극, 텅스텐 전극과 대등하거나 우수한 신호대 잡음비(signal to noise ratio, SNR)특성을 가짐을 확인하였다.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.1
• /
• pp.38-46
• /
• 2024

In general, EEG signal analysis has been the subject of several studies due to its ability to provide an objective mode of recording brain stimulation, which is widely used in brain-computer interface research with applications in medical diagnosis and rehabilitation engineering. In this study, we developed EEG reception hardware to measure electroencephalograms and implemented a processing system, classifying it into server and data processing. It was conducted as an intermediate-stage research on the implementation of a brain-computer interface using electroencephalograms, and was implemented in the form of predicting the user's arm movements according to measured electroencephalogram data. Electroencephalogram measurements were performed using input from four electrodes through an analog-to-digital converter. After sending this to the server through a communication process, we designed and implemented a system flow in which the server classifies the electroencephalogram input using a convolutional neural network model and displays the results on the user terminal.