ACE Lab

A new data-aided frequency estimator is introduced for phase-shift keying signals transmitted over frequency-selective fading channels. This estimator is developed based on a maximum likelihood criterion. It assumes the use of a special class of pilots, called near-i.i.d. (independent identically distributed) sequences, with impulsive fourth-order moments. With the help of such pilots, the proposed method can estimate frequency offsets without the need for channel information. The pilots of GSM and IS-136 mobile communication systems have been observed as being near-i.i.d., and statistical analysis indicates that the proposed estimate is almost unbiased if the pilot is near-i.i.d. The advantage of the proposed estimator over conventional methods is demonstrated via computer simulation.

A new data-aided frequency estimator for frequency-selective fading channels is introduced. The proposed estimator is developed based on a least squares (LS) error criterion and can estimate frequency offsets without the need for channel information. Statistical analysis indicates that the resulting estimate is unbiased and tends to approach the Cramer-Rao lower bound (CRLB). Simulation shows that the proposed LS method is preferable to existing techniques in mobile communications. The application of the LS estimator to systems with transmitter antenna diversity is also considered. In particular, it is demonstrated that the LS method can be successfully applied to third-generation wireless communication systems.

A digital signal processing technique for compensating both the I/Q mismatch and the DC offset in communication receivers is derived with an emphasis on direct-conversion architectures. The I/Q mismatch and DC offset are estimated in a least-squares sense using a training sequence. Also, a group of training sequences that minimizes the mean square error of the estimate is determined. The advantages of the proposed technique are demonstrated through computer simulation.

The application of double space-time transmit diversity (DSTTD) scheme to multicarrier systems, such as orthogonal frequency division multiplexing (OFDM) systems, requires calculating the determinations of the antenna permutation matrices for all subcarriers, resulting in a heavy computation load. In this paper, we show that the signal-to-noise ratio (SNR)-based antenna shuffling criterion for DSTTD systems can be reduced to a simple criterion that evaluates determinants of 2 times 2 submatrices of the 4 times 4 equivalent channel matrix. The new criterion can lighten the computational load by about 95%. Furthermore, it is shown that the minimum mean square error (MMSE)-based criterion for antenna permutation can also be reduced to the same criterion.

This correspondence proposes techniques that jointly compensate for amplifier nonlinearity and quadrature modulation/demodulation (QM/QDM) errors. The proposed methods are derived based on the polynomial predistortion (PD) employing the indirect learning technique and do not require any additional feedback loop for QM/QDM-error compensation. Compared with the existing joint compensation technique, the proposed methods need some additional parameters to be estimated but exhibit faster convergence and better performance. The advantage of the proposed technique is demonstrated through computer simulation.

We propose an efficient Wiener model for a power amplifier (PA) and develop a direct learning predistorter (PD) based on the model. The Wiener model is formed by a linear filter and a memoryless nonlinearity in which AM/AM and AM/PM characteristics are approximated as piecewise linear and piecewise constant functions, respectively. A two-step identification scheme, wherein the linear portion is estimated first and the nonlinear portion is then identified, is developed. The PD is modeled by a polynomial and its coefficients are directly updated using a recursive least squares (RLS) algorithm. To avoid implementing the inverse of the PA's linear portion, the cost function for the RLS algorithm is defined as the sum of differences between the output of the PA's linear portion and the inverse of the PA's nonlinear portion. The proposed direct learning scheme, which is referred to as the piecewise RLS (PWRLS) algorithm, is simpler to implement, yet exhibits comparable performance, as compared with existing direct learning schemes.

A two-way relay channel (TWRC) in which two terminals T 1 and T 2 exchange orthogonal frequency division multiplexing (OFDM) signals with the help of an amplify-and-forward (AF) relay T 3 is considered here, and an efficient technique for allocating powers to N parallel tones of OFDM is developed. A sum rate maximization problem is formulated by replacing the individual power constraints of the conventional sum rate maximization problem, which limit the power of each terminal, with the total power constraint limiting the sum of powers of all terminals. The maximization problem with the total power constraint yields a more efficient power allocation policy than the conventional problem with individual power constraints. It is shown that the closed-form solution of the maximization problem under the total power constraint can be obtained for a single-tone system ( N =1). Based on this result, a two-step suboptimal approach is proposed in which the power is optimally assigned to each tone first, and then at each tone the assigned power is distributed to the three terminals. The proposed method is shown to assign 50% of the total power to relay T 3 irrespective of the channels. It is demonstrated that the proposed method is considerably simpler to implement than the conventional dual-decomposition method (DDM), yet the performance of the former is almost identical to that of the latter.

This letter proposes an opportunistic power control for increasing the transmission rate of a secondary transmitter (ST) in frequency reused scenario. The environments considered herein have multiple primary transmitter/receiver pairs and one secondary transmitter/receiver pair using the same frequency band. In this scenario, we develop the opportunistic power control method for assisting successive interference cancellation (SIC) and optimal SIC order. Simulation results show that the proposed method improves the ST transmission rate significantly when compared to the method with no power control, and it exhibits a small throughput loss compared to the optimal technique while offering simpler implementation and using less transmission power.

This paper describes the design of a time-division duplexing frame with a variety of pilots for multiuser multiple-input-multiple-output orthogonal frequency-division multiple access (MU-MIMO OFDMA) systems, where the base station and users are equipped with four and two transmitting and receiving antennas, respectively. In addition, a simplified scheduling algorithm for the MU-MIMO OFDMA is proposed, and its computational complexity is analyzed. The proposed scheduling algorithm shows comparable sum achievable rates to the optimal MU-MIMO OFDMA scheduling that searches for user pairs in an exhaustive manner, whereas its complexity is fairly reduced. Furthermore, to verify the performance of MU-MIMO OFDMA systems that employ the proposed frame structure and scheduling algorithm, a system-level comparison of the average cell throughputs between the proposed MU-MIMO and the conventional MIMO OFDMA systems is numerically performed in a practical cellular environment. As a result, vital information on how we can apply MU-MIMO OFDMA schemes in cellular environments is provided.

An adaptive digital predistortion (PD) technique is proposed for linearization of power amplifiers (PAs) in multiple-input multiple-output (MIMO) transmitters. We consider a PD structure equipped with only one combined feedback path while conventional systems have multiple feedback paths. Hence, the proposed structure is much simpler than that of multiple feedback paths. Based on the structure, a new PD algorithm is derived. The simulation results show that linearization performance of the proposed method is almost the same as the conventional multiple feedback technique while the former is much simpler to implement than the latter.

We consider the design of a digital predistortion (DPD) scheme when there is a crosstalk between a local oscillator (LO) and a power amplifier (PA) called LO self-coupling. We first develop a baseband equivalent model for a direct conversion transmitter afflicted by LO self-coupling. Based on this model, a two-stage process, consisting of a feedback equalizer (FBEQ) and DPD, is proposed for the joint compensation of PA nonlinearity and LO self-coupling. The advantage of the proposed DPD over conventional DPD is demonstrated through computer simulation.

In this paper, it is proposed to use asymmetric complex signaling in full-duplex decode-and-forward single antenna relay channels in order to eliminate the self-interference signal and increase the throughput. Specifically, the relay system considered in this paper is as follows: the source has a transmit weight; the relay has an Rx processing whose operation is an inner product between the received signal and a weight; the relay has a transmit weight as well; the destination has a receive weight. The objective is to find the optimum weights to increase the smaller SNR between source-relay and relay-destination under the constraint of perfect self-interference nulling. We convert the complex variable into a 2-dimensional real vector. As a consequence, the problem becomes a joint vector optimization problem. The simulation result shows that the proposed signalling achieves higher rate than the conventional full-duplex relay.

An adaptive digital predistortion (PD) technique is proposed for linearization of power amplifiers (PAs) in multiple-input multiple-output (MIMO) transmitters. Instead of constructing a separate feedback path at each PA, we propose a PD technique based on a combined feedback, which adds all the PAs' outputs to form a single feedback. We develop an algorithm for finding the multiple PA characteristics jointly from the combined feedback. The PD parameters are calculated from the identified PA characteristics. Although the proposed method requires more computations for PA identification than the conventional method, its hardware implementation is much simpler and can overcome the crosstalk problem of a radio frequency (RF) selector, which is necessary for multiple feedback selection. The simulation results show that, despite its simple structure, the linearization performance of the proposed method is almost the same as the separate multiple feedback method. Published in: IEEE Communications Letters ( Volume: 16,

Cognitive beamforming exploiting spatial opportunity is an attractive technique for secondary users to coexist with primary users in cognitive radio environments. If perfect channel state information of the interfering link is available, interference from a secondary transmitter to a primary receiver can be perfectly pre-nulled by choosing the ideal transmit beam. In practice, however, there is channel estimation error due to noise and the time-varying channels. To minimize the residual interference due to those channel estimation errors, channel prediction based on auto regressive (AR) model is introduced in this paper. Further, to cope with extremely rapidly-varying channels, a cognitive transmit power control technique is proposed as well. By combining channel prediction and transmit power control in cognitive beamforming, the cognitive users can share the spectrum with the primary users with a limited interference level in time-varying channels.

This paper proposes an opportunistic feedback and user selection method for a multiuser two-way relay channel (MU-TWRC) in a time-varying environments where a base station (BS) and a selected mobile station (MS), one of K moving MSs, exchange messages during two time slots via an amplify-and-forward relay station. Specifically, under the assumption of perfect channel reciprocity, we analyze the outage probabilities of several channel feedback scenarios, including the proposed scheme. Based on the analysis, the transmission rates are optimized and the optimal user selection method is proposed to maximize the expected sum throughput. The simulation results indicate that, with opportunistic feedback, the performance can be significantly improved compared to that without feedback. Moreover, the performance is nearly identical to that with full feedback, and close to the case of perfect channel state information at BS for low mobility MSs. Copyright © 2013 The Institute of Electronics, Information and Communication Engineers.

4세대 이동통신시스템 '롱텀에볼루션(LTE)'의 핵심기술인 다중 송수신 (MIMO) 안테나 시스템은 추가적인 주파수나 송신전력의 할당 없이도 채널 용량을 안테나 수에 비례하여 증가시킬 수 있는 장점으로 인해 데이터 전송 속도를 획기적으로 늘리는 기술로 4세대 이동통신기술 진화의 핵심기술로 꼽히고 있다. 본 논문은 다중 송수신 안테나 시스템에서 사용되는 2개 이상의 DAC 대신에 고속 DAC와 아날로그 스위치를 각각 한 개씩 사용하여 기존 방식의 장점을 유지하며 설계비용 감소, 물리적 크기 축소화 등 경제적 효율을 증가시킬 수 있는 MIMO 시스템에서 데이터 컨버터를 줄일 수 있는 방안을 제안한다.

In wireless communication systems, the transmitter-receiver oscillator mismatch and Doppler effect result in a carrier frequency offset (CFO) of the received signals. Since the CFO results in a rotation of the phase of the received signal, the receiver performance is severely degraded without compensation for the CFO. Hence, estimation and compensation for the CFO is crucial at the receiver. This paper proposes a new frequency offset estimation method based on partial correlation. By using the partial correlation output for the CFO estimation, the computational complexity of the proposed technique is much smaller than the conventional method that uses the full training sequence. In addition, the estimation range is wider than that of the conventional method. To verify the estimation performance, computer simulation is conducted and comparison with the conventional method is presented.

In this paper, we propose a method of reducing the number of data converters in MIMO (multiple- input multiple-output) transmitters. While conventional MIMO transmitters require one data converter for each transmitting antenna, the proposed technique requires only one high-speed data converter and an analog switch, regardless of the number of antennas. Experimental results confirm the effectiveness of the proposed technique and also show that the proposed transmitted performs similarly to conventional systems with multiple data converters.

A digital predistortion (DPD) technique to linearize multiple power amplifiers (PAs) in analog beamforming systems is proposed. The analog beamforming system considered in this paper has one digital chain and multiple PAs/antennas controlled by phase shifters. Due to the system configuration, a single DPD should linearize the multiple PAs. To design the DPD, this paper introduces a cost function, squares of the sum of errors for all the PAs.The DPD solution minimizing the cost function is found by a recursive least squares (RLS) algorithm. Experimental results with commercial PAs show that the proposed DPD can effectively linearize multiple PAs.

This paper proposes a digital predistortion (DPD) technique to improve linearization performance when the power amplifier (PA) is driven near the saturation region. The PA is a non-linear device in general, and the nonlinear distortion becomes severer as the output power increases. However, the PA’s power efficiency increases as the PA output power increases. The nonlinearity results in spectral regrowth, which leads to adjacent channel interference, and degrades the transmit signal quality. According to our simulation, the linearization performance of DPD is degraded abruptly when the PA operates in its saturation region. To relieve this problem, we propose an improved DPD technique. The proposed technique performs on/off control of the adaptive algorithm based on the magnitude of the transmitted signal. Specifically, the adaptation normally works for small and medium signals while it stops for large signals. Therefore, harmful coefficient updates by saturated signals can be avoided. A computer simulation shows that the proposed method can improve the linearization performance compared with the conventional DPD method in highly driven PAs.

A digital predistortion (DPD) technique based on an iterative adaptation structure is proposed for linearizing power amplifiers (PAs). To obtain proper DPD parameters, a feedback path that converts the PA’s output to a baseband signal is required, and memory is also needed to store the baseband feedback signals. DPD parameters are usually found by an adaptive algorithm by using the transmitted signals and the corresponding feedback signals. However, for the adaptive algorithm to converge to a reliable solution, long feedback samples are required, which increases hardware complexity and cost. Considering that the convergence time of the adaptive algorithm highly depends on the initial condition, we propose a DPD technique that requires relatively shorter feedback samples. Specifically, the proposed DPD iteratively utilizes the short feedback samples in memory while keeping and using the DPD parameters found at the former iteration as the initial condition at the next iteration. Computer simulation shows that the proposed technique performs better than the conventional technique, as the former requires much shorter feedback memory than the latter.

A new digital pre-distortion (DPD) technique with low memory requirements is proposed for linearizing power amplifiers (PAs). Lookup table (LUT) and Polynomial methods are two popular techniques to determine DPD parameters. LUT DPD is simpler to implement but needs a longer convergence time than the polynomial DPD, leading to longer feedback samples. Thus, LUT DPD requires a large memory for storing long feedback and transmitted samples, which increases hardware complexity and cost. To solve this problem, we propose a new LUT DPD technique that iteratively utilizes a short feedback and transmit samples. The experimental results show that the proposed technique can linearize the commercial PA better than the conventional technique, as the former requires only 30 times shorter feedback memory than the latter.

This paper proposes a digital pre-distortion (DPD) technique for analog beamforming systems. The considering system has M PAs (M is the number of antennas) and one digital chain. In contrast to the conventional DPD problem where one DPD linearizes one PA, a single DPD should linearize M PAs simultaneously in the analog beamforming system. In this paper, we design a new DPD by minimizing the sum of linearization errors of M PAs. To verify the proposed DPD, computer simulation and experiment are conducted. The results confirm that the proposed DPD can successfully linearize the multiple PAs in analog beamforming systems.

A new digital pre-distortion technique to linearize power amplifiers (PAs) is being proposed for concurrent dual-band transmitters. In the considering system, dual-band signals are combined and amplified by a single wideband PA. The PA output signal is distorted by the cross-modulation between different band signals as well as their own inter-modulation. In conventional dual-band DPD techniques, two independent dualfeedback paths are required to convert the dual band PA outputs into baseband signals, respectively. However, it increases hardware complexity and expense. In this paper, we propose a new DPD method requiring only a single feedback path. In this new structure, the proposed technique first estimates the dual-band PA characteristics using the single feedback path. The DPD parameters are then extracted from the estimated PA characteristics. Computer simulation demonstrates that the proposed technique can achieve comparable performance with the conventional dual-feedback DPD, with significantly reduced hardware complexity.

Background/Objectives: A new non-uniform lookup table (LUT) based Digital Pre-distortion (DPD) technique to reduce hardware complexity without degradation of DPD performance is proposed. Methods/Statistical analysis: The performance degradation of the power amplifier (PA) occurs due to the PAs nonlinearity. This paper considers the use of non-uniform LUT digital pre-distortion (DPD) to linearize the PA. In the proposed method, we reduce the LUT quantization error by allocating more tables in severe non-linear region while less tables in linear region. To recognize the nonlinear characteristics of the PA, we first estimate the PA characteristics and differentiate the characteristics. Findings: In recent mobile communication systems, linearization technology of PA is getting more and more important. Various linearization techniques have been studied to find the cost-effective technique, and DPD is known as one of the most effective methods to linearize PA. In particular, the LUT based DPD technique is used in a limited hardware resource environment due to its low computational complexity. However, it causes a quantization error when the size of the table is not enough big. To solve these problems, we propose a non-uniform LUT based DPD scheme. The conventional LUT DPD divides the magnitude of the input signals equally. The proposed LUT DPD, on the other hand, allocates a small number of tables in the linear region and a large number of tables in nonlinear region. By doing this, the linearization performance can be improved. We will show through simulation that the proposed method provides better linearization performance than the conventional LUT DPD if the same number of tables is used. Improvements/Applications: By applying this technique, we can achieve the same performance with less hardware resources, or the proposed method can improve linearization performance if the same LUT size is used. Key Words: Non-uniform LUT (Lookup Table), DPD (Digital pre-distortion), Nonlinear PA (Power amplifier), polynomial, Low hardware complexity

Background/Objectives: A jamming technique was proposed for frequency hopping signals in commercial drones to raise awareness of secure communication, and it can be used for disabling flight of unauthorized drones. Methods/Statistical analysis: We analyze the characteristics of the communication signal between the drone and the controller of a commercial drone, and show that the communication system is vulnerable to jamming attacks. The wideband sampling, windowing and fast Fourier transform (FFT) are employed to observe the frequency characteristics. RF signal measurement equipment and MATLAB are used to analyze the frequency hopping signal of the commercial drone. Findings: Unmanned aerial vehicles, or drones, have attracted considerable attention due to their versatile availability as military and civilian uses. However, as interest on the drones grows, they may be exposed to various harmful environments such as ambient noise/signal interference or intentional communication disturbance attacks. Therefore, it is essential to design a secure communication system between the drones and the controllers. In this paper, we first analyze the communication signal between a commercial drone and the controller, which is turned out to be a frequency hopping signal. Improvements/Applications: Frequency hopping signals are considered as one of the most secure communication technologies. This paper shows that the hopping signals are not secure and can be damaged by an intentional reactive jammer. Key Words: Commercial Drone, Unmanned aerial vehicles, Frequency hopping spread spectrum, Reactive jammer

Background/Objectives: Increasing use of UAVs causes many social problems. The purpose of this paper is to develop drone-jamming technique that protects social safety against indiscreet and illegal use of UAVs Methods/Statistical analysis:First, we analyze the frequency hopping pattern of the commercial drones in order to design jamming system. Specifically, the instantaneous Tx frequency is detected within several micro seconds. Therefore, the frequency analyzer tracks the communication frequency of the drone that changes every hop. After that, it transmits a strong jamming signal to the corresponding frequency, thereby disrupting the communication of the drones. All these procedures are completed in one-hop period, which enables reactive jamming. Findings: As commercial drones become widespread, drones are often used for malicious purposes. One example is the problem of drones flying into restricted areas. In this paper, we propose reactive jamming to disable unauthorized drones. The proposed method is as follows. When a jamming signal is transmitted to drones that use FHSS communication, the communication between the drones and the controller is disabled. In this case, we confirmed that the drone was switched to failsafe mode, and the connection to the drones was disabled on the controller screen. When the transmission of the jamming signal is turned off, the communication between the controller and the drone is connected again. When using the reactive jamming technique proposed in this paper, it is possible to prevent bomb terrorist, unauthorized reconnaissance, narcotics smuggling etc. Improvements/Applications: The conventional jamming method transmits a strongwideband interference for frequency hopping signals.The proposed reactive jamming technique, however, transmits a relatively weak interference signal to the narrow band to disable the frequency hopping communication.

In this paper, a multiuser space-time line code (STLC) system is proposed. In the proposed multiuser STLC system, a zero-forcing (ZF)-based precoder decomposes multiuser multiple-input multiple-output (MIMO) channels into multiple single-user MIMO channels, and the STLC is performed with the effective single user MIMO channel independently for each user. Here, to maximize the sum rate, an optimal power allocation method is designed based on a water-filling strategy. Also, a simple suboptimal method called fairness-aware per-user (FAPU) power allocation is devised. It is analytically and numerically verified that each user of the proposed ZF-based STLC system asymptotically achieves the maximum of single-user achievable rate as the number of transmit antennas increases. The numerical results of rigorous simulation justify that the proposed ZF-based STLC with FAPU power allocation provides near optimal performance and outperforms the existing ZF-based maximum Eigen beamforming scheme and conventional multiuser STLC schemes. Furthermore, two simple search algorithms to find the optimal number of users that maximizes sum rate are devised. Numerical results verify that the proposed algorithms can effectively find the optimal number of users by reducing the search interval.

Recently, air-to-ground wireless communication systems have been used in various equipment such as unmanned aerial vehicle (UAV). In this paper, we analyze the air-to-ground channel in 400 MHz band with experiment using light aircraft. Methods/Statistical analysis: The characteristics of the air-to-ground channel are analyzed as follows. We generate and store a pilot signal in commercial vector signal generator and an aircraft equipped with the signal generator takes off. As a pilot signal, Zadoff-Chu sequence is used for its good autocorrelation characteristics. In the ground station, the received waveform is stored by using spectrum analyzer. When an airplane approaches near the ground station, airplane transmits the pilot signal and the ground station begins to store the received signals. After completing the signal storing, air-to-ground channel analysis is performed. Findings: From the experiments, we obtained several hundreds of received waveforms, and those waveforms are used for statistical air-to-ground channel analysis. According to the analysis results, most of the measurement channels are single path or line of sight channels. However, in a certain measurement, multiple paths or non-line of sight channels are observed. Those results confirm that multipath environments exist at certain air-to-ground condition. The maximum number of multipath components is 4 and the maximum delay spread is 0.4 µ sec. Therefore, when designing a wireless communication system considering air-to-ground environments, the design should consider 4-path fading channels and maximum delay spread over 0.4 µsec. When designing air-to-ground wireless communication systems such as aircraft and drone systems, the experimental and analyzed channels in this paper might be useful. Improvements/Applications: In this paper, we analyze the air-to-ground channel characteristics of a mountainous area in Korea with field experiment using light aircraft. The analyzed channel characteristics are helpful to those who want to design wireless communication systems between aircraft to ground or UAV to ground. In addition, the analyzed channel can be used for verification of

Background: Wearable devices have evolved as screening tools for atrial fibrillation (AF). A photoplethysmographic (PPG) AF detection algorithm was developed and applied to a convenient smartphone-based device with good accuracy. However, patients with paroxysmal AF frequently exhibit premature atrial complexes (PACs), which result in poor unmanned AF detection, mainly because of rule-based or handcrafted machine learning techniques that are limited in terms of diagnostic accuracy and reliability. Objective: This study aimed to develop deep learning (DL) classifiers using PPG data to detect AF from the sinus rhythm (SR) in the presence of PACs after successful cardioversion. Methods: We examined 75 patients with AF who underwent successful elective direct-current cardioversion (DCC). Electrocardiogram and pulse oximetry data over a 15-min period were obtained before and after DCC and labeled as AF or SR. A 1-dimensional convolutional neural network (1D-CNN) and recurrent neural network (RNN) were chosen as the 2 DL architectures. The PAC indicator estimated the burden of PACs on the PPG dataset. We defined a metric called the confidence level (CL) of AF or SR diagnosis and compared the CLs of true and false diagnoses. We also compared the diagnostic performance of 1D-CNN and RNN with previously developed AF detectors (support vector machine with root-mean-square of successive difference of RR intervals and Shannon entropy, autocorrelation, and ensemble by combining 2 previous methods) using 10 5-fold cross-validation processes. Results: Among the 14,298 training samples containing PPG data, 7157 samples were obtained during the post-DCC period. The PAC indicator estimated 29.79% (2132/7157) of post-DCC samples had PACs. The diagnostic accuracy of AF versus SR was 99.32% (70,925/71,410) versus 95.85% (68,602/71,570) in 1D-CNN and 98.27% (70,176/71,410) versus 96.04% (68,736/71,570) in RNN methods. The area under receiver operating characteristic curves of the 2 DL classifiers was 0.998 (95% CI 0.995-1.000) for 1D-CNN and 0.996 (95% CI 0.993-0.998) for RNN, which were significantly higher than other AF detectors (P<.001). If we assumed that the dataset could emulate a sufficient number of patients in training, both DL classifiers improved their diagnostic performances even further especially for the samples with a high burden of PACs. The average CLs for true versus false classification were 98.56% versus 78.75% for 1D-CNN and 98.37% versus 82.57% for RNN (P<.001 for all cases). Conclusions: New DL classifiers could detect AF using PPG monitoring signals with high diagnostic accuracy even with frequent PACs and could outperform previously developed AF detectors. Although diagnostic performance decreased as the burden of PACs increased, performance improved when samples from more patients were trained. Moreover, the reliability of the diagnosis could be indicated by the CL. Wearable devices sensing PPG signals with DL classifiers should be validated as tools to screen for AF.

In this Letter, the authors propose a novel convolutional neural network (CNN)-based estimation of the distance between an ultra-wideband (UWB) transmitter and receiver for a localisation. By exploiting the UWB signal characteristics, such as high-resolution in the time domain, the CNN is designed. The proposed CNN-based method estimates the distance from only the received signals, without signal-to-noise ratio information which is used for the conventional time of arrival (TOA)-based methods. Furthermore, as verified by simulation, the proposed CNN-based method significantly outperforms the conventional TOA-based method with respect to the distance estimation accuracy.

Background/Objectives: The objective of this paper is to develop a new distance estimation method for indoor localization by using ultra-wideband (UWB) signals.The new technique is based on recurrent neural network (RNN), one of the famous deep learning techniques. Methods/Statistical analysis: RNN is one of the most suitable methods for learning time series data. Hence, it is useful in processing data that change with time. The proposed method estimates the distance based on RNN from the received UWB signals. Specifically, from the received signals via IEEE802.15.4a indoor channels, the proposed RNN regressor estimates the distance. The proposed method is validated by computer simulation. Findings: In order to estimate the distance using UWB signals, it is necessary to accurately detect the first arrived signal along the shortest path.To find this signal or time instance by using RNN, we convert 1-dimensional received signal into 2-dimensional signal. The converted signal is input to RNN regressor and trained so that the RNN output is the distance between the transmitter and the receiver. The conventional method needs received signal-to-noise ratio (SNR)estimation, and the threshold is determined by the estimated SNR. When the received signal exceeds the threshold, the first arrived signal is detected and the arrived time is called the time of arrival (ToA). However, the proposed method estimates the distance directly from the received signal without SNR estimation.According to the simulation results, the proposed method shows RMSE of less than 2 [m] in low SNR, and less than 0.5 [m] in high SNR.Those performancesare much better than the conventional method. Improvements/Applications: The performance of the proposed RNN based distance estimator is examinedthrough computer simulation. To compare estimation accuracy, the root mean square error (RMSE) is measured. According to the results, the proposed estimator is superior to the conventional method. Keywords:distance estimation, indoor localization, ultra-wideband (UWB), recurrent neural network (RNN), regression, root mean square error (RMSE)

A new frame synchronization technique based on convolutional neural network (CNN) is proposed for synchronized networks. To estimate the exact packet arrival time, the receiver typically uses the correlator between the received signal and the preamble or pilot in front of the transmitted packet. The conventional frame synchronization technique searches the correlation peak within the time window. In contrast, the proposed method utilizes a CNN to find the packet arrival time. Specifically, in the proposed method, the 1D correlator output is converted into a 2D matrix by reshaping, and the resulting signal is inputted to the proposed 4-layer CNN classifier. Then, the CNN predicts the packet arrival time. To verify the frame synchronization performance, computer simulation is performed for two channel models: additive white Gaussian noise and fading channels. Simulation results show that the proposed CNN-based synchronization method outperforms the conventional correlation-based technique by 2 dB .

We propose a new transmission structure for single carrier – frequency domain equalization (SCFDE) to reduce pilot overhead. The proposed structure transmits multiple data frames between two pilot frames. At the receiver, two channel estimates for the pilot frames are obtained first, and the channels for the data frames are found by linear interpolation of the two channel estimates from the pilots. Computer simulation shows that the proposed structure reduces the pilot overhead significantly

본 논문은 기존의 SC-FDE구조에 기반하여 협대역 재머 대응을 위한 새로운 SC-FDE 구 조를 제안한다. 기존의 SC-FDE구조는 협대역 재머가 발생했을 시 높은 전력의 재밍 신호 간섭에 의해 시간영역에서의 채널 추정이 어려워지고, 그로인해 데이터 복원 성능이 저하 된다. 이러한 문제를 해결하기 위해 협대역 재머가 발생했을 시 채널추정이 가능한 SC-FDE 프레임 구조를 제안한다. 본 논문에서는 주파수영역에서 채널추정이 가능한 변형 된 SC-FDE의 구조를 제안하며 그에 따른 성능을 컴퓨터 모의실험을 통해 검증하였다.

The demand for the Internet of things (IoT) technologies is growing rapidly. Hence, it is of paramount importance to address the battery problem associated with IoT devices. Although the required battery life of most of these devices is two or three years, in practice, the battery can last only a few months. The research presented in this paper aims to reduce the power consumption of IoT devices by using the wake-up modem proposed herein. To design low-power IoT devices, we analyze which parts of the devices consume the most power. Based on findings in the literature, the conventional IoT devices account for 60% of the total power consumption in the communication subsystem. Hence, the design of a low-power communication subsystem is critical for the low-power design of IoT devices. IoT devices have a very small amount of data for transmission; besides, data communication occurs very occasionally. Thus, the IoT device is in a standby mode most of the time. To reduce the standby mode power consumption, we propose a wake-up modem and verify its effectiveness and efficiency via computer simulation. Moreover, this paper suggests separating the wake-up modem to detect the wake-up preamble and the main modem for data recovery. While the detection block of the wake-up preamble is working in standby mode, only wake-up receiver blocks are operated, and the remaining receiver blocks are turned off. This greatly reduces the standby mode power consumption. In addition, the wake-up receiver can decrease the portion of the data receiver to the total receiver power consumption, which enables the data transceiver design without considering power consumption. The main radio is turned on when a wake-up preamble is detected. Considering that IoT devices are in standby mode most of the time, the proposed scheme can significantly reduce the overall power consumption since the proposed wake-up modem can reduce power consumption in standby mode significantly. Therefore, by applying the proposed scheme, the total power consumption for communication can be minimized in communication systems whose data transaction does not frequently occur, for example, IoT devices.

Background: Continuous photoplethysmography (PPG) monitoring with a wearable device may aid the early detection of atrial fibrillation (AF). Objective: We aimed to evaluate the diagnostic performance of a ring-type wearable device (CardioTracker, CART), which can detect AF using deep learning analysis of PPG signals. Methods: Patients with persistent AF who underwent cardioversion were recruited prospectively. We recorded PPG signals at the finger with CART and a conventional pulse oximeter before and after cardioversion over a period of 15 min (each instrument). Cardiologists validated the PPG rhythms with simultaneous single-lead electrocardiography. The PPG data were transmitted to a smartphone wirelessly and analyzed with a deep learning algorithm. We also validated the deep learning algorithm in 20 healthy subjects with sinus rhythm (SR). Results: In 100 study participants, CART generated a total of 13,038 30-s PPG samples (5850 for SR and 7188 for AF). Using the deep learning algorithm, the diagnostic accuracy, sensitivity, specificity, positive-predictive value, and negative-predictive value were 96.9%, 99.0%, 94.3%, 95.6%, and 98.7%, respectively. Although the diagnostic accuracy decreased with shorter sample lengths, the accuracy was maintained at 94.7% with 10-s measurements. For SR, the specificity decreased with higher variability of peak-to-peak intervals. However, for AF, CART maintained consistent sensitivity regardless of variability. Pulse rates had a lower impact on sensitivity than on specificity. The performance of CART was comparable to that of the conventional device when using a proper threshold. External validation showed that 94.99% (16,529/17,400) of the PPG samples from the control group were correctly identified with SR. Conclusions: A ring-type wearable device with deep learning analysis of PPG signals could accurately diagnose AF without relying on electrocardiography. With this device, continuous monitoring for AF may be promising in high-risk populations. Clinicaltrial: ClinicalTrials.gov NCT04023188; https://clinicaltrials.gov/ct2/show/NCT04023188.

In medical treatment, healthcare and home appliances, the demand for Internet of Things (IoT) services is increasing in urban areas. However, in fading environments of urban area, the existing IoT communication system can degrade severely. The purpose of this paper is to develop a new IoT communication technology that improves communication performance in fading channels. IoT communication systems are aiming at low power long distances. In addition, the amount of data transmission is small due to the service requirements. To achieve this goal, existing IoT communication modems are designed with narrowband signals. However, narrowband signals may cause performance degradation in a frequency selective fading channel environments in urban area. In this paper, we propose a split channel two tone OOK (SC-TT-OOK) that implements on-off keying (OOK) using two tones. Unlike the existing OOK by using one tone, SC-TT-OOK uses spaced two tones to improve communication performance in fading channels. The conventional OOK method has a similar characteristic to that of the existing IoT modems with a narrow bandwidth spectrum. The proposed SC-TT-OOK is OOK using two separated tones. Since the proposed modulation uses two tones, it requires twice the frequency resources of the conventional OOK. However, if the frequencies of the two tones are sufficiently spaced apart, frequency diversity gain can be achieved in the frequency selective fading channels, and communication reliability can be improved in urban environments. The proposed SC-TT-OOK is examined through computer simulation. The simulation results show that the proposed scheme outperforms the conventional scheme in the frequency selective fading channel. Therefore, the proposed method can be used as an alternative technique for IoT modem in urban environments.

Recently, high accuracy localization technique is required to provide indoor location services. The purpose of this paper is to propose a distance estimation technique based on deep convolutional neural network (DCNN) for indoor environments. Among distance estimation techniques based on wireless communication signals, the use of ultra-wideband (UWB) signals has the advantage of high accuracy in the time domain. The proposed distance estimation method uses UWB signals and proposes a new DCNN-based distance estimator. The superiority of the proposed method is confirmed through computer simulation. Widely used conventional distance estimators are based on the power threshold. The threshold is determined by signal to noise ratio (SNR) of the received signal. The arrival time of the received signal that exceeds the threshold is considered as the time-of-arrival (ToA) and the distance between transmitter and receiver is obtained from the ToA. On the other hand, the proposed distance estimator requires only the received signal without SNR estimation, which make the proposed technique simpler to implement. According to computer simulation, the conventional method is highly sensitive to SNR and distance. In contrast, the proposed method shows less than 2 m root mean square error (RMSE) performance in a wide range of SNR and the RMSE performance is not degraded in long distances. The proposed distance estimator shows excellent distance estimation performance at low SNR and long distance, so it can be applied to indoor localization system of large indoor space and can be used for precise location service.

A new frame synchronization technique based on convolutional neural network (CNN) is proposed for synchronized networks. To estimate the exact packet arrival time, the receiver typically uses the correlator between the received signal and the preamble or pilot in front of the transmitted packet. The conventional frame synchronization technique searches the correlation peak within the time window. In contrast, the proposed method utilizes a CNN to find the packet arrival time. Specifically, in the proposed method, the 1D correlator output is converted into a 2D matrix by reshaping, and the resulting signal is inputted to the proposed 4-layer CNN classifier. Then, the CNN predicts the packet arrival time. To verify the frame synchronization performance, computer simulation is performed for two channel models: additive white Gaussian noise and fading channels. Simulation results show that the proposed CNN-based synchronization method outperforms the conventional correlation-based technique by 2 dB .

Localization has been extensively studied owing to its huge potential in various areas, such as Internet of Things, 5G, and unmanned aerial vehicle services. Its wide applications include home automation, advanced production automation, and unmanned vehicle control. In this study, we propose a novel localization method that utilizes convolutional neural network (CNN) and ultra-wideband (UWB) signals. A localization problem is converted to a regression problem with the proposed CNN, in which the ranging and positioning phases are integrated. By integrating the ranging and positioning phases, the proposed CNN estimates the location of UWB transmitter directly without any additional step. To integrate both phases of localization, a simple-yet efficient input image generation method is proposed. In the proposed input image generation method, three oversampled two-dimensional input images are generated from the three received UWB signals and they are provided to the designed CNN through the three channels, which are represented by red-, green-, and blue-color channels, respectively. The proposed CNN-based localization system then estimates the location of the UWB transmitter directly using the three-channel image as an input of the CNN. Simulation results verify that the proposed CNN-based localization method outperforms the traditional threshold-based and existing CNN-based methods. Also, it is observed that the proposed method performs well under an asymmetric environment, unlike the existing method.

This paper presents the first full end-to-end deep learning framework for the swift prediction of lithium-ion battery remaining useful life. While lithium-ion batteries offer advantages of high efficiency and low cost, their instability and varying lifetimes remain challenges. To prevent the sudden failure of lithium-ion batteries, researchers have worked to develop ways of predicting the remaining useful life of lithium-ion batteries, especially using data-driven approaches. In this study, we sought a higher resolution of inter-cycle aging for faster and more accurate predictions, by considering temporal patterns and cross-data correlations in the raw data, specifically, terminal voltage, current, and cell temperature. We took an in-depth analysis of the deep learning models using the uncertainty metric, t-SNE of features, and various battery related tasks. The proposed framework significantly boosted the remaining useful life prediction (25X faster) and resulted in a 10.6% mean absolute error rate.

In this study, cognitive radio systems are examined. As it is difficult to include all the candidate frequency bands of the wide spectrum of the primary channel using the filtering approach, we propose a stair-type parallel predistortion (PD) and implement it using a commercial wideband Gallium Nitride power amplifier. Parallel predistorters centred at different integer multiples of are used for constructing the PD to effectively reduce spectral regrowth of the fundamental signal as well the harmonics at different frequencies. Experimental results demonstrate that the spectral regrowth at GHzand harmonic at 2GHz are reduced by 15dBand 10dB , respectively. The obtained results establish the proposed PD with parallel predistorters as a predistortion technique for dynamic spectrum allocation systems such as cognitive cellular systems and CR networks.

This paper proposes a new spectrum sensing technique for cognitive radio systems. To determine vacancy of the spectrum, the proposed method employs the recurrent neural network (RNN), one of the popular deep learning techniques. The proposed technique determines the spectrum occupancy of the primary user (PU) by observing the received signal’s energy and any information on the PU signal characteristic is not used. To this end, the received signal’s spectrum is obtained by fast Fourier transform (FFT). This process is performed on consecutive received signals and the resulting spectrums are stacked. Finally, a 2-dimensional spectrum (or spectrogram) is made. This 2-D spectrum is cut into sensing channel bandwidths and inputted to the deep learning model to decide the channel’s occupancy. While the recently published spectrum sensing technique based on convolutional neural network (CNN) relies on an empty channel, the proposed technique does not require any empty channel. Only the channel signal of interest to sense is needed. Since spectrum sensing results is two (busy or idle), binary classification deep learning model is developed. According to the computer simulation results, the proposed method has similar performance with the conventional CNN-based method while the spectral efficiency of the proposed method is much higher than that of the existing scheme. In addition, the overall learnable parameters of the proposed deep learning model is only 2/3 of the existing method

In wireless communication systems, the performance of the receiver is very sensitive to time and frequency offsets. In particular, orthogonal frequency division multiplexing (OFDM) systems are highly vulnerable to those offsets due to inter-carrier interference (ICI) and inter-symbol interference (ISI). To solve this problem, wireless local area network (WLAN) systems transmit a preamble for synchronization. In this paper, we propose a joint time and frequency offsets estimation technique based on convolutional neural network (CNN) for WLAN systems. In the proposed technique, the correlation between the received signal and the transmitted preamble is performed first. Then the frequency offset is coarsely compensated by several hypothesized offsets. The compensated signals are inputted to the proposed CNN and the CNN predicts the time and frequency offsets. The estimation performance is examined through computer simulation. According to the results, the proposed time offset estimator shows 3 dB to 6 dB performance gain, and the frequency offset estimator shows much lower root mean square error (RMSE) performance than the conventional technique at low SNRs

The objective of this paper is to propose a multiple target identification technique for orthogonal frequency division multiplexing (OFDM) radars. First, a 2-D (range & Doppler) periodogram is obtained from the reflected signal through 2-D fast Fourier transform (FFT) of the received OFDM symbols. Usually, the peaks of the periodogram indicates the targets. Conventionally, peak search algorithms are used to find the multiple targets. In this paper, however, a convolutional neural network (CNN) classifier is proposed to identify the targets. The proposed technique does not need any additional information but the 2-D periodogram while the conventional method requires the noise variance as well as the periodogram. The performance is examined through computer simulation. According to the results, if the number of maximum identifiable targets are small, the proposed technique performs well. However, as the number increases, the detection accuracy decreases. In the simulation environments, the proposed method outperforms the conventional one. The proposed OFDM radar technique can be applied to 6G mobile communications to identify the moving targets around the transmitter without additional frequency resource for radar systems.

A 900-MHz high-sensitivity split-channel binary frequency-shift keying (SC-BFSK) reconfigurable data/wake-up receiver (RX) for low-power wide-area networks (LPWANs) is presented. In the data mode, the proposed RX demodulates data through the RF/analog front end (RXFE), the analog-to-digital converter (ADC), and the modulator–demodulator (MODEM). The MODEM implemented in the external microcontroller unit (MCU) applies additional digital signal processing to improve about 17-dB sensitivity. In the wake-up mode, the proposed RX saves the power dissipation by turning off all blocks except for the RXFE and wake-up preamble detector (WuPD). The WuPD following the RXFE provides a 17-dB sensitivity improvement in place of the MODEM of the data RX. Adopting the proposed SC-BFSK, in which other channels are inter-allocated between the BFSK tones of one channel, increases the number of channels while taking advantage of wide tone spacing, which improves the bit error rate (BER) in a multipath fading channel. To implement the ultra-low-power (ULP) SC-BFSK RX, the RXFE employs the 2:1 LO sliding-IF architecture and a band-pass filter (BPF)-based frequency-to-energy detection baseband demodulator. Implemented in 55-nm CMOS, the proposed RX, including both data and wake-up RXs, exhibits −123-dBm sensitivity with 0.39-kbps data rate and 81.92-ms wake-up latency, while the RX chip dissipates 0.88 mW.

Background/Objectives: Atrial Fibrillation (AFib), one of the arrhythmias, causes the atrium to beat irregularly. AFib can be diagnosed by observing electrocardiogram (ECG) signals. However, the degree of irregular running depends on the patient, and it is difficult to detect AFib in early patients. As a result, it is important to accurately judge Sinus Rhythm (SR) and AFib, and only doctors with extensive experience in cardiology are known to judge accurately. In this paper, we propose a convolutional neural network (CNN) to perform accurate AFib diagnosis from ECG signals.

고속의 멀티미디어 통신서비스 수요 증가에 따라 주파수 스펙트럼 자원이 부족하고 이에 따라 동적인 환경에서의 고속 서비스 제공에 문제가 발생하고 있다. 따라서, 추가 주파수 자원 할당 없이 통신과 레이다 기능을 동시에 지원하는OFDM(Orthogonal Frequency Division Multiplexing) Radar를 사용함으로써 주파수 자원을 절약할 수 있고, 동적인 환경에서도 고속의 서비스 지원이 가능하다. 본 논문에서는 OFDM Radar 시스템과 CNN(Convolution Neural Network)을 사용하여 레이다가 감지한 표적의 거리와 속도를 동시에 추정하는 모델을 제안한다. 표적이 1개인 경우(단일 물체)와 표적이 2개인경우(다중 물체)의 성능을 비교하였다. 단일 이미지 입력일 때 SNR 2dB 기준, 단일 물체와 다중 물체의 거리 및 속도 추정MAE(Mean Absolute Error) 차이는 각각 1.03m, 4.53km/h로 다중 물체 대비 단일 물체의 MAE 성능이 우수하다. 다중 이미지 입력일 때 역시 모든 SNR 구간에서 단일 물체의 거리 및 속도 추정 MAE가 우수하다. 그 원인을 파악하기 위해 두 물체간의 최소 간격을 조정해 실험한 결과, SNR(Signal-to-Noise Ratio) 2dB 기준 두 물체 사이의 최소 간격이 커질수록 거리및 속도 추정 CNN의 성능이 개선됨을 확인하였다. 이는 두 물체의 최소 간격이 에러 발생 확률에 영향을 미침을 의미한다

Recently, as the importance of healthcare has increased, researches are being conducted to measure health status in real time. Heart Rate (HR) measurement is one of the important health conditions that measure heart beat rates. HR measurement can be performed using Photoplethysmogram (PPG) or Electrocardiogram (ECG) signals. Since the PPG or ECG signals are different from people to people, conventional HR estimator occasionally results in large errors. To develop a reliable HR estimator, an HR estimation technique using PPG is proposed in this paper, based on a deep learning technique. The proposed HR estimation technique has the following key features. We develop a new artificial neural network which is 1-Dimensional Convolutional Neural Network (1D-CNN) composed of ten convolutional layers and two fully connected layers. To assess the estimation performance, cross validation is used. The training and verification of the proposed 1D-CNN technique are performed on Python 3.7.5 with Keras 2.0 library. The proposed HR estimation technique performs training and verification using field PPG data. Overfitting is prevented by increasing the limited training data by data augmentation. In training, the loss function is the Mean Square Error (MSE), which is commonly used in regression problems. In the verification, the error between the predicted HR and the actual HR is compared using Mean Absolute Error (MAE). As a result of the final performance verification through cross validation, the proposed technique shows an MAE of 1.23 Beats Per Minute (BPM). This results indicate that the proposed technique enables quick and accurate HR estimation with only PPG signals. Therefore, if this technique is applied to medical and wearable devices, the proposed technique can replace the existing HR monitors.

The objective of this work is to propose a new method of estimating velocity and distance based on multi-output convolutional neural network (CNN) for orthogonal frequency division multiplexing (OFDM) radars. The two-dimensional (2D) periodogram is extracted from the received reflected waveforms through radar signal processing of received OFDM symbols. Conventionally, constant false alarm rate (CFAR) algorithm is used to estimate distance and velocity of targets. In contrast, this paper proposes a novel deep-learning based approach for the estimation of the targets in OFDM radar systems. The proposed multi-output CNN-based target detector estimates the distance and velocity of the target simultaneously. The proposed technique is verified through computer simulation. The results show that the proposed multi-output CNN-based method demonstrates more accurate distance and speed estimates than the conventional CFAR. Specifically, the distance and speed estimates of the proposed method are 9.8 and 12.3 times accurate, respectively, than those of the conventional CFAR.

The objective of this paper is to propose a new target distance and velocity estimation technique for OFDM radar systems. First, the 2D periodogram is collected from the reflected signal via FFT of received OFDM symbols. The largest value of a 2D periodogram often represents the target so that its position indicates the distance and velocity. The CFAR is one of the famous conventional techniques to find the peak in the 2D periodogram. In this paper, a CNN based estimator is proposed. The proposed CNN directly finds the distance and velocity from the 2D periodogram. The proposed method requires only 2D periodogram to estimate the target’s distance and velocity. On the other hand, the conventional methods need noise variance as well as the periodogram. The performance is examined through computer simulation. In the simulation, the MAEs are compared between the conventional and proposed methods. According to the results, the MAEs of the proposed method are lower approximately 8 m in distance and 7 km/h in speed to the conventional method. The proposed OFDM radar technique can be applied to 6G mobile communications to identify the moving targets without additional frequency resource allocation for the radar system. In other words, by using the proposed technique, the convergence of the communication and radar can be possible

In this study, utilizing the structure of the space–time line code (STLC) signals, a blind symbol and carrier synchronization method considering the carrier frequency offset (CFO) is developed. The proposed method can achieve almost perfect synchronization performance with more than three transmit antennas, irrespective of the CFO values. Based on the results, the proposed blind symbol and carrier synchronization method is expected to accelerate the practical deployment and operation of the low-complexity STLC systems that achieve full spatial diversity gain with partial or no channel state information at the receiver.

Continuous blood pressure (BP) measurement is vital in monitoring patients’ health with a high risk of cardiovascular disease. The complex and dynamic nature of the cardiovascular system can influence BP through many factors, such as cardiac output, blood vessel wall elasticity, circulated blood volume, peripheral resistance, respiration, and emotional behavior. Yet, traditional BP measurement methods in continuously estimating the BP are cumbersome and inefficient. This paper presents a novel hybrid model by integrating a convolutional neural network (CNN) as a trainable feature extractor and support vector regression (SVR) as a regression model. This model can automatically extract features from the electrocardiogram (ECG) and photoplethysmography (PPG) signals and continuously estimates the systolic blood pressure (SBP) and diastolic blood pressure (DBP). The CNN takes the correct topology of input data and establishes the relationship between ECG and PPG features and BP. A total of 120 patients with available ECG, PPG, SBP, and DBP data are selected from the MIMIC III database to evaluate the performance of the proposed model. This novel model achieves an overall Mean Absolute Error (MAE) of 1.23 ± 2.45 mmHg (MAE ± STD) for SBP and 3.08 ± 5.67 for DBP, all of which comply with the accuracy requirements of the AAMI SP10 standard.

Advancements in wireless sensor network technologies have enabled the proliferation of miniaturized body-worn sensors, capable of long-term pervasive biomedical signal monitoring. Remote cardiovascular monitoring has been one of the beneficiaries of this development, resulting in non-invasive, photoplethysmography (PPG) sensors being used in ambulatory settings. Wrist-worn PPG, although a popular alternative to electrocardiogram, suffers from motion artifacts inherent in daily life. Hence, in this paper, we present a novel deep learning framework (CorNET) to efficiently estimate heart rate (HR) information and perform biometric identification (BId) using only a wrist-worn, single-channel PPG signal collected in ambulant environment. We have formulated a completely personalized data-driven approach, using a four-layer deep neural network. Two convolution neural network layers are used in conjunction with two long short-term memory layers, followed by a dense output layer for modeling the temporal sequence inherent within the pulsatile signal representative of cardiac activity. The final dense layer is customized with respect to the application, functioning as: regression layer-having a single neuron to predict HR; classification layer-two neurons that identify a subject among a group. The proposed network was evaluated on the TROIKA dataset having 22 PPG records collected during various physical activities. We achieve a mean absolute error of 1.47 ± 3.37 beats per minute for HR estimation and an average accuracy of 96% for BId on 20 subjects. CorNET was further evaluated successfully in an ambulant use-case scenario with custom sensors for two subjects.

An efficient number estimation method is proposed for target detection by using the forward-backward matrix pencil method. Compared with the traditional parameter estimation methods, the proposed method can work well even for sparse signals. At first, the radar cross section of the group targets arranged in linear array is obtained at some certain angles or frequencies. Then, the Hankel-Toeplitz matrix is constructed from the known radar echoes, and the singular value decomposition is performed on this matrix. Finally, the number of the group targets arranged in linear array can be accurately estimated by setting an appropriate threshold to select the singular value with a large proportion. In addition, the influence of different signal-to-noise ratios (SNR) on the estimation results is also discussed. In order to enhance the accuracy, several groups of different echoes are used to estimate the number of group targets jointly. The simulation results demonstrate that the proposed method is effective and accurate in estimating the number of group targets arranged in linear array. Moreover, the estimation result of the proposed method in this paper is not affected by the noise, which shows that this method has better robustness.

In this study, a cooperative communication system that employs multiple decode-and-forward relay nodes (RNs), in which the associated/active RNs perform phase rotation of the regenerated signals before retransmitting them to a destination node (DN), is examined. In the first phase, i.e., communication from a source node to RNs, a received signal-to-noise ratio (SNR) threshold-based RN association method is proposed. The optimal SNR thresholds are designed to maximize the bit-error-rate (BER) performance at the DN under various communication environments, such as modulation types and channel code rates. Furthermore, the number of phase rotations (PRs) in a frame is examined. Intensive numerical results show that more PRs in a frame provide better BER performance at the DN, irrespective of the communication environments. This study provides a valuable guideline for designing practical cooperative networks with multiple decode-and-forward RNs with PRs.

A 900-MHz high-sensitivity split-channel binary frequency-shift keying (SC-BFSK) reconfigurable data/wake-up receiver (RX) for low-power wide-area networks (LPWANs) is presented. In the data mode, the proposed RX demodulates data through the RF/analog front end (RXFE), the analog-to-digital converter (ADC), and the modulator-demodulator (MODEM). The MODEM implemented in the external microcontroller unit (MCU) applies additional digital signal processing to improve about 17-dB sensitivity. In the wake-up mode, the proposed RX saves the power dissipation by turning off all blocks except for the RXFE and wake-up preamble detector (WuPD). The WuPD following the RXFE provides a 17-dB sensitivity improvement in place of the MODEM of the data RX. Adopting the proposed SC-BFSK, in which other channels are inter-allocated between the BFSK tones of one channel, increases the number of channels while taking advantage of wide tone spacing, which improves the bit error rate (BER) in a multipath fading channel. To implement the ultra-low-power (ULP) SC-BFSK RX, the RXFE employs the 2:1 LO sliding-IF architecture and a band-pass filter (BPF)-based frequency-to-energy detection baseband demodulator. Implemented in 55-nm CMOS, the proposed RX, including both data and wake-up RXs, exhibits -123-dBm sensitivity with 0.39-kbps data rate and 81.92-ms wake-up latency, while the RX chip dissipates 0.88 mW.

인지통신(cognitive radio)과 같이 가변 스펙트럼 할당이 필요한 시스템을 위한 새로운 전치왜곡기를 제안한다. 본 논문에서 고려하는 시스템 모델에서 신호는 한 순간에는 작은 대역폭을 차지하지만 그 중심 주파수가 시간에 따라 변화할 수 있는 상황을 가정한다. 이러한 시나리오에서는 전력 증폭기 출력 단에 위치하는 종단 필터로는 전력증폭기에 의한 하모닉을 제거하지 못하는 상황이 발생할 수 있다. 제안된 전치왜곡기는 기본 주파수 (ω_ο) 신호의 비선형 왜곡을 선형화할 뿐만 아니라, 2ω_ο, 3ω_ο, ...에서 발생하는 하모닉도 동시에 제거한다. 제안된 전치왜곡기는 ω_ο 주파수의 정수배에 대응하는 여러 개의 전치왜곡기가 결합된 구조를 가지고 있다. 기본 주파수 ω_ο 에 해당하는 전치왜곡기는 기본 주파수 신호의 선형화를 담당하며, 나머지 주파수에 대응하는 전치왜곡기는 하모닉을 제거하는 역할을 담당한다. 제안된 전치왜곡기에서 필요한 변수는 최소 평균 자승 에러 알고리즘에 의해 동시에 계산되며, 모의실험 결과에 따르면 제안된 방법을 이용하면 기본 주파수의 스펙트럼에 발생하는 스펙트럼 왜곡이 20dB 감소하며, 2차 및 3차 하모닉도 기본 신호의 전력대비 약 -70dB로 작아지는 것을 확인할 수 있다.

본 논문에서는 다중 대역 전송 시스템에서 광대역 전력 증폭기의 선형화를 위한 새로운 전치왜곡 기법을 제안한다. 특히, 한 시스템에서 동시에 다중대역/다중모드 신호를 전송함에 있어 다중대역 신호가 하나의 전력 증폭기에 의해 증폭되어 전송되는 시스템을 고려한다. 상호 대역 간 비선형 왜곡을 포함한 비선형 특성을 보상하기 위하여, 본 논문에서는 다중 전치왜곡기 블록을 갖는 새로운 전치왜곡 구조를 제안하며, 다중 전치왜곡기의 계수를 동시에 갱신하는 적응 알고리즘을 제안한다. 제안하는 다중 대역 모델을 검증하기 위하여 상용 증폭기를 사용하여 증폭기 모델을 추출하였으며, 추출된 모델을 기반으로 제안한 알고리즘을 모의실험을 통해 검증하였다. 모의실험 결과는 제안 알고리즘이 효과적으로 다중 전치왜곡기의 계수를 구할 수 있으며, 다중 대역을 효과적으로 선형화 할 수 있음을 보여준다.

본고에서는 MIMO 송신기에서 비선형 전력증폭기를 선형화하기 위한 디지털 전치왜곡 기법을 제안한다. 기존의 시스템에서는 각 전력증폭기에 한 개씩 되먹임 회로가 필요한 반면 본고에서 제안하는 전치왜곡 시스템은 전력증폭기 출력 신호를 모두 결합하여 한 개의 되먹임 회로만 가지는 특징이 있다. 따라서 기존 시스템에 비해 훨씬 간단한 구조를 가진다. 이러한 구조를 바탕으로 결합된 피드백 신호로부터 각 전력증폭기를 선형화하는 전치왜곡 알고리즘을 제안한다. 모의실험 결과에 의하면 제안된 방식은 각 전력증폭기에 하나씩 되먹임 회로를 구성한 기존 방식과 거의 동일한 선형화 특성을 보임을 확인하였다.

본 논문에서는 W-CDMA 기지국용 envelope tracking 전력 증폭기의 선형성 특성을 개선하는 새로운 드레인 바이어스 기법을 제안한다. 기존의 envelope tracking 전력 증폭기에서 드레인 바이어스 전압은 트랜지스터의 문턱전압 근처까지 감소하여 선형성 특성이 크게 나빠진다. 이 문제를 해결하기 위해서 본 연구에서는 입력 신호가 작을 때는 드레인 바이어스 전압이 고정된 class AB로 동작하게 하고 입력 신호가 클 때는 envelope tracking 동작을 하도록 하는 방법을 제안한다. 또한, envelope tracking 동작에서 신호의 왜곡을 줄이도록 드레인 바이어스 전압과 입력 신호의 관계를 새로이 구한다. 제안된 기법의 효과를 검증하기 위하여 class AB Si-LDMOS 전력 증폭기를 사용하여 W-CDMA envelope tracking 전력 증폭기를 설계하였다. 제안된 드레인 바이어스 기법은 평균 효율을 저하시키지 않으면서 선형성 특성을 크게 개선하여 추가의 선형화 기법 없이도 W-CDMA 기지국용 전력 증폭기의 선형성 사양을 만족시키는 것을 시뮬레이션을 통해 확인하였다.

주파수 오차는 송·수신기내의 발진기의 오차와 송수신기의 이동속도에 의한 도플러 효과(doppler effect)로 인해 발생하는데 이러한 오차는 수신신호의 위상을 변화시켜 수신기의 성능을 떨어뜨리는 주요 요인 중 하나이다. 따라서 주파수 오차의 정밀한 추정 및 보상은 송·수신기의 필수적인 요소이다. 본 논문은 이러한 주파수 오차를 부분 상관을 이용하여 추정하는 새로운 방식을 제안하는데, 기존 방식에 비해 낮은 복잡도를 가진다. 또한 제안하는 방식은 주파수 추정 정확성의 손실 없이 주파수 오차 보상 범위의 조절이 가능하여 넓은 주파수 오차가 존재하는 시스템에 적합하다. 제안 방식의 검증을 위해 컴퓨터 모의실험 결과를 수행하고, 기존의 기법과 비교 분석하여 성능 및 복잡도에 이득이 있음을 보인다.

세대 이동통신시스템 ‘롱텀에볼루션(LTE)'의 핵심기술인 다중 송수신 (MIMO) 안테나 시스템은 추가적인 주파수나 송신전력의 할당 없이도 채널 용량을 안테나 수에 비례하여 증가시킬 수 있는 장점으로 인해 데이터 전송 속도를 획기적으로 늘리는 기술로 4세대 이동통신기술 진화의 핵심기술로 꼽히고 있다. 본 논문은 다중 송수신 안테나 시스템에서 사용되는 2개 이상의 DAC 대신에 고속 DAC와 아날로그 스위치를 각각 한 개씩 사용하여 기존 방식의 장점을 유지하며 설계비용 감소, 물리적 크기 축소화 등 경제적 효율을 증가시킬 수 있는 MIMO 시스템에서 데이터 컨버터를 줄일 수 있는 방안을 제안한다.

GPS (global positioning system)는 지구 궤도를 도는 여러 GPS 위성의 신호를 수신하여 이들 신호의 수신 시간 차이를 이용하여 자신의 절대 위치를 알아내는 시스템으로 민간 목적뿐만 아니라 군용 목적으로도 널리 사용되고 있다. 특히 군용 시스템에서는 무기 등에도 사용되기 때문에 안정적인 GPS 수신이 민간보다 더 중요하다. 하지만 GPS 는 사용 주파수가 알려져 있기 때문에 나쁜 의도로 인위적인 재밍신호를 송출하여 GPS 시스템을 무력화 시킬 수 있다는 문제가 있다. 이에 재밍신호를 제거하고 GPS 신호만을 안정적으로 수신하기 위해 GPS 수신기에 다중 안테나를 설치하고 빔 널링 기술을 이용하여 재머만을 제거하는 기술이 많이 연구되어 왔다. 이러한 기술 중에서 본 논문에서는 MVDR (minimum variance distortionless response) 기술에 기반하여 새로운 적응형 빔 제어 기술을 제안한다. 제안하는 방식은 위성 방향으로의 빔은 훼손하지 않으면서 재머 방향으로는 적응적으로 널을 형성하여 재머를 효과적으로 제거한다. 제안하는 방식의 성능은 컴퓨터 모의실험을 통해 검증한다.

본 논문에서는 이중 대역 송신시스템에서 비선형 전력증폭기를 선형화하기 위한 새로운 디지털 전치왜곡 기법을 제안한다. 기존의 시스템에서는 자기 대역의 비선형뿐만 아니라 타 대역간의 비선형을 보상하기 위해 두 개의 피드백 경로가 요구되는데 이는 하드웨어의 복잡도와 비용을 증가시킨다. 반면 제안하는 전치왜곡 시스템은 한 개의 피드백 경로만을 사용한다. 따라서 기존 시스템에 비해 훨씬 간단한 구조를 가진다. 제안하는 기법은 전치왜곡 계수를 얻기 위해서 한 개의 피드백 경로를 공유하여 두 대역의 전력증폭기 특성을 먼저 추정하고 추정된 전력증폭기 특성으로부터 전치왜곡 계수를 얻는다. 컴퓨터 모의실험 결과에 따르면 제안된 방식은 두 개의 피드백 회로를 구성한 기존 방식과 대등한 선형화 성능을 보인다.

GPS (global positioning system)는 지구 궤도를 도는 여러 GPS 위성의 신호를 수신하여 위성 별 수신 시간 차이를 이용하여 자신의 절대 위치를 알아내는 시스템으로 민간 목적뿐만 아니라 군용 목적으로도 널리 사용되고 있다. 특히 유도무기 등에 사용되기 때문에 신뢰성 있는 GPS 수신은 군에서 훨씬 더 중요하다. 하지만 GPS 는 사용 주파수 대역이 알려져 있고 수신신호 전력이 미약하기 때문에 인위적인 재밍신호를 송출하여 쉽게 GPS 시스템을 무력화 시킬 수 있는 문제가 있다. 이러한 간섭신호에 대응하기 위한 한 가지 방법은 GPS 수신기에 다중 안테나를 설치하고 디지털 빔포밍 기술을 이용하여 GPS 신호는 훼손하지 않으면서 재머만을 제거하는 기술이다. 본 논문에서는 MVDR (minimum variance distortionless response) 기술에 기반하여 새로운 적응형 빔 제어 기술을 제안한다. 제안하는 방식은 알려져 있는 위성 방향으로의 빔을 형성하고 재머 방향으로는 적응형 알고리즘에 의해 널을 형성하여 재머를 제거한다. 제안하는 방식의 성능은 컴퓨터 모의실험을 통해 검증한다.

본 논문에서는 배열 안테나 기반의 MVDR (minimum variance distortionless response) 항 재밍 GPS (global positioning system) 수신기에서 안테나 경로 사이에 존재하는 이득, 위상, 지연 편차를 교정하기 위한 안테나 교정 기법을 제안한다. 제안하는 기법은 다중 안테나 시스템에서 안테나 경로들 사이의 이득, 위상, 지연 편차를 추정하고 이를 보상한다. 경로 간 이득, 위상, 지연 편차를 정확히 추정하기 위해 좋은 상관 특성을 갖는 파일럿 신호가 사용된다. 교차 상관에 기반하여 지연 편차가 먼저 추정되며 이후에 이득과 위상 편차가 추정된다. 정밀한 지연 편차 추정 및 보상을 위해 보간기법을 사용하는데 이산 푸리에 변환 (DFT) 기반의 보간기법으로 계산 복잡도를 감소시킨다 . 제안 된 기법은 MATLAB을 이용한 컴퓨터 모의실험을 통해 검증한다. 모의실험 결과에 따르면 제안하는 기법을 적용하면 이득, 위상 시간 지연 편차를 각각 0.01 dB, 0.05 도, 0.5 ns 이내로 줄일 수 있다.

본 논문은 협대역 간섭신호에 대응하기 위한 새로운 SC-FDE 구조를 제안한다. 기존의 SC-FDE구조는 협대역 간섭신호가 발생했을 시 채널 추정이 어려워지고, 그로 인해 데이터 복원이 어려운 상황이 발생한다. 이러한 문제를 해결하기 위해 본 논문에서는 큰 전력의 협대역 간섭신호가 발생했을 때에도 주파수영역 채널추정이 가능한 새로운 SC-FDE 프레임 구조를 제안한다. 구체적으로 기존방식은 시간영역 채널추정을 먼저 수행한 후 푸리에변환을 통해 주파수 영역 채널을 추정하지만 본 논문은 곧바로 수신신호에서 주파수영역에서 채널추정이 가능하도록 새로운 SC-FDE의 구조를 제안하며 제안하는 구조의 성능 향상은 컴퓨터 모의실험을 통해 검증하였다. 모의실험 결과 재머의 크기가 수신신호의 크기와 동일한 환경에서 제안하는 방식은 재머가 없는 경우 대비 약 2 dB 이내의 손실로 수신이 가능하지만 기존의 방식은 통신이 불가능 하다.

근접한 이중대역 신호에 대한 새로운 디지털 전치왜곡 기법을 제안한다. 본 논문에서 고려하는 시스템은 이중대역 신호를 한 개의 전력증폭기로 증폭하는 송신기이다. 이 경우 전력증폭기 출력신호는 두 대역 신호의 교차변조 및상호변조에 의해 왜곡이 발생한다. 특히 두 대역이 가까울 때에는 각 대역에서 상호변조에 의해 발행한 스펙트럼이서로 겹칠 수 있고 이는 전치왜곡 성능을 저하시킨다. 이러한 문제를 해결하기 위해서 제안하는 기법은 먼저 전력증폭기 특성을 추정하고 이렇게 추정한 전력증폭기 특성을 바탕으로 전치왜곡 계수를 추출한다. 이러한 두 단계를 통해 전치왜곡 계수를 구하면 근접한 이중대역 신호에 대해서도 서로 간섭 없이 전치왜곡이 동작할 수 있다. 제안하는기법은 컴퓨터 모의실험을 통해 검증하는데, 모의실험 결과에 따르면 제안하는 기법이 기존의 이중대역 전치왜곡 방법보다 우수한 성능을 보인다.

본 논문에서는 파일럿으로 발생하는 파일럿 오버헤드를 감소시키면서 협대역 간섭신호 혹은 재머에 대응할 수 있는 새로운 SC-FDE (single carrier frequency domain equalization) 구조를 제안한다. 기존의 SC-FDE 구조는 협대역 재머가 존재할 때 시간영역 채널 추정이 어려워 수신기의 성능이 저하되는 단점이 있다. 또한 매 SC-FDE마다 채널추정을 위한 파일럿을 전송하여 스펙트럼 효율이 저하되는 문제도 있다. 이와 같은 문제를 해결하기 위해 제안된 구조는 시간영역 채널 추정 없이 곧바로 주파수 영역 채널 추정을 수행할 수 있도록 설계된다. 또한 파일럿 오버헤드를 감소시키기 위해서 두 개의 파일럿 사이에 여러 개의 SC-FDE 데이터 블록을 전송하고, 두 개의 파일럿을 이용한 선형 보간 채널 추정을 통해 각 데이터 블록에서의 채널 특성을 찾는다. 이렇게 구한 채널 추정값을 사용하여 주파수 영역 등화를 수행하면 협대역 재머가 있는 상황에서도 재머 제거와 함께 안정적인 등화가 가능하다. 제안하는 구조의 성능은 컴퓨터 모의실험을 통해 확인한다.

본 논문에서는 ultra-wideband(UWB) 시스템에서 합성곱 신경망(CNN)을 이용한 거리 추정 기법을 제안한다. 제안하는 기법은 UWB 신호를 이용하여 송신기와 수신기 사이의 거리를 추정하기 위하여 수신신호의 크기 샘플로 이루어진 1차원 벡터를 2차원 행렬로 재구성하며, 이 2차원 행렬로부터 합성곱 신경망 회귀를 이용하여 거리를 추정한다. IEEE 802.15.4a 표준의 UWB 실내 가시선 채널모델을 이용하여 수신신호를 생성하여 학습데이터를 만들며 합성곱 신경망 모델을 학습시킨다. 또한 실제 필드 시험을 통해 실내환경에서의 실험 데이터를 이용하여 거리추정 성능을 확인한다. 제안하는 기법은 기존의 문턱값 기반의 거리 추정 기법과의 성능비교도 수행하는데, 결과에 따르면 10m 거리에서 제안기법은 0.6m의 제곱근 평균 자승 에러를 보이는데 기존기법은 1.6m로 훨씬 큰 에러를 보인다.

본 논문에서는 인지 무선 통신을 위한 새로운 합성곱 신경망 기반 스펙트럼 센싱 기법을 제안한다. 제안하는 기법은 주 사용자 신호에 대한 어떠한 사전 정보도 알지 못하는 상황에서 에너지 검출을 통해 주 사용자 신호 유무를 판단한다. 제안하는 기법은 센싱하고자 하는 전체 대역을 고려하여 수신신호를 고속으로 샘플링한다. 이후 신호의 FFT (fast Fourier transform)을 통해 주파수 스펙트럼으로 변환하고 연속적으로 이와 같은 스펙트럼을 쌓아서 2차원 신호를 만든다. 이렇게 만든 2차원 신호를 탐지하고자 하는 채널 대역폭 단위로 자르고 합성곱 신경망에 입력하여 채널이 사용 중인지 비어있는지 판단한다. 판단하고자 하는 분류의 종류가 두 가지이므로 이진 분류 합성곱 신경망을 사용한다. 제안하는 기법의 성능은 컴퓨터 모의실험과 실제 실내환경에서의 실험을 통해 검증하는데 이 결과에 따르면 제안하는 기법은 기존 문턱값 기반 기법보다 2 dB 이상 우수한 성능을 보인다.

본 논문에서는 초광대역 (Ultra-wideband, UWB) 시스템에서 실내 위치 측위를 위한 새로운 거리 추정 기법을 제안한다. 제안하는 기법은 딥러닝 기법 중 하나인 순환 신경망 (RNN)을 기반으로 한다. 순환신경망은 시계열 신호를 처리하는데 유용한데 UWB 신호 역시 시계열 데이터로 볼 수 있기 때문에 순환신경망을 사용한다. 구체적으로, UWB 신호가 IEEE 802.15.4a 실내 채널모델을 통과하고 수신된 신호에서 순환신경망 회귀를 통해 송신기와 수신기 사이의 거리를 추정하도록 학습한다. 이렇게 학습된 순환신경망 모델의 성능은 새로운 수신신호를 이용하여 검증하며 기존의 임계값 기반의 거리 추정 기법과도 비교한다. 성능지표로는 제곱근 평균추정에러 (root mean square error, RMSE)를 사용한다. 컴퓨터 모의실험 결과에 따르면 제안하는 거리 추정 기법은 수신신호의 신호 대 잡음비 (signal to noise ratio, SNR) 및 송수신기 사이의 거리와 상관없이 기존 기법보다 항상 월등히 우수한 성능을 보인다.

본 논문에서는 합성곱 신경망(CNN)에 기반한 프레임 동기 기법을 제안한다. 기존의 프레임 동기 기법은 프리앰 블과 수신 신호 사이의 상관을 통해 수신 신호와 프리앰블이 일치하는 지점을 찾는다. 제안하는 기법은 1차원 벡터로 이루어진 상관기 출력 신호를 2차원 행렬로 재구성하며, 이 2차원 행렬을 합성곱 신경망에 입력하고 합성곱 신경망 은 프레임 도착 지점을 추정한다. 구체적으로 가산 백색 가우스 잡음(AWGN) 환경에서 무작위로 도착하는 수신 신 호를 생성하여 학습 데이터를 만들고, 이 학습 데이터로 합성곱 신경망을 학습시킨다. 컴퓨터 모의실험을 통해 기존 의 동기 기법과 제안하는 기법의 프레임 동기 오류 확률을 다양한 신호 대 잡음 비(SNR)에서 비교한다. 모의실험 결 과는 제안하는 합성곱 신경망을 이용한 프레임 동기 기법이 기존 기법 대비 약 2dB 우수함을 보인다.

본 논문에서는 인지 무선 통신을 위한 새로운 순환 신경망 기반 스펙트럼 센싱 기법을 제안한다. 제안하는 기법은 주사용자에 대한 정보가 전혀 없는 상황에서 에너지 검출을 통해 신호 존재 유무를 판단한다. 제안 기법은 센싱하고자 하는 전체 대역을 고려하여 수신신호를 고속으로 샘플링 후 이 신호의 FFT (fast Fourier transform)를 통해 주파수 스펙트럼으로 변환한다. 이 스펙트럼 신호는 채널 대역폭 단위로 자른 후 순환 신경망에 입력하여 해당 채널이 사용중인지 비어있는지 판정한다. 제안하는 기법의 성능은 컴퓨터 모의실험을 통해 확인하는데 그 결과에 따르면 기존 문턱값 기반 기법보다 2 [dB] 이상 우수하며 합성곱 신경망 기법과 유사한 성능을 보인다. 또한, 실제 실내환경에서 실험도 수행하는데 이 결과에 따르면 제안하는 기법이 기존 문턱값 기반 방식 및 합성곱 신경망 방식보다 4 [dB] 이상 우수한 성능을 보인다.

고속의 멀티미디어 통신서비스 수요 증가에 따라 주파수 스펙트럼 자원이 부족하고 이에 따라 동적인 환경에서의 고속 서비 스 제공에 문제가 발생하고 있다. 따라서, 추가 주파수 자원 할당 없이 통신과 레이다 기능을 동시에 지원하는 OFDM(Orthogonal Frequency Division Multiplexing) Radar를 사용함으로써 주파수 자원을 절약할 수 있고, 동적인 환경에 서도 고속의 서비스 지원이 가능하다. 본 논문에서는 OFDM Radar 시스템과 CNN(Convolution Neural Network)을 사용하 여 레이다가 감지한 표적의 거리와 속도를 동시에 추정하는 모델을 제안한다. 표적이 1개인 경우(단일 물체)와 표적이 2개인 경우(다중 물체)의 성능을 비교하였다. 단일 이미지 입력일 때 SNR 2dB 기준, 단일 물체와 다중 물체의 거리 및 속도 추정 MAE(Mean Absolute Error) 차이는 각각 1.03m, 4.53km/h로 다중 물체 대비 단일 물체의 MAE 성능이 우수하다. 다중 이미 지 입력일 때 역시 모든 SNR 구간에서 단일 물체의 거리 및 속도 추정 MAE가 우수하다. 그 원인을 파악하기 위해 두 물체 간의 최소 간격을 조정해 실험한 결과, SNR(Signal-to-Noise Ratio) 2dB 기준 두 물체 사이의 최소 간격이 커질수록 거리 및 속도 추정 CNN의 성능이 개선됨을 확인하였다. 이는 두 물체의 최소 간격이 에러 발생 확률에 영향을

Distance and depth detection plays a crucial role in intelligent robotics. It enables drones to understand their working environment to avoid collisions and accidents immediately and is very important in various AI applications. Image-based distance detection usually relies on the correctness of geometric information. However, the geometric features will be lost when the object is rotated or the camera lens image is distorted. This study proposes a training model based on a convolutional neural network, which uses a single-lens camera to estimate humans’ distance in continuous images. We can partially restore depth information loss using built-in camera parameters that do not require additional correction. The normalized skeleton feature unit vector has the same characteristics as time series data and can be classified very well using a 1D convolutional neural network. According to our results, the accuracy for the occluded leg image is over 90% at 2 to 3 m, 80% to 90% at 4 m, and 70% at 5 to 6 m.

본 논문에서는 비선형 앰프를 선형화하는 LUT (look-up table)에 기반한 디지털 전치왜곡 알고리즘을 제안한다. 제안하는 방식은 기존 다항식 기반의 간접학습 방법을 이용하여 전력증폭기를 선형화하는 사후왜곡기를 먼저 구한 다음 이를 바탕으로 LUT 전치왜곡기를 구한다. 제안하는 기법에서는 기존 다항식 기법의 빠른 학습속도와 우수한 선형성능의 장점을 취하면서 많은 곱셈기로 인한 실시간 전치왜곡의 구현 복잡도를 LUT를 이용하여 낮추는 것이 장점이다. 제안하는 기법의 성능은 위성통신에 사용되는 실제 전력증폭기를 이용한 실험을 통해 확인한다. 구체적으로 약 8 GHz 대역에서 동작하는 광대역 TWTA (Travelling Wave Tube Amplifier)를 이용하여 제안하는 전치왜곡기법과 기존 다항식 전치왜곡 기를 비교한다. 실험결과 LUT의 크기가 64 이상이면 다항식 기법과 성능차이가 거의 없는 것을 확인할 수 있다. 제안하는 기법은 사용하는 곱셈의 개수가 1/10 수준에 불과하기 때문에 구현복잡도 측면에서 장점을 가진다.

본 논문에서는 이동 채널 환경에서 도플러 효과로 시간에 따라 채널이 빠르게 변할 때 수신성능을 개선할 수 있는 새로운 SC-FDE 전송구조를 제안한다. 기존의 SC-FDE 구조는 매 SC-FDE 심볼의 앞쪽에 파일럿을 삽입하여 전송하는 반면, 제안하는 구조에서는 SC-FDE 심볼의 중간에 파일럿을 삽입하여 전송한다. 수신기에서는 파일럿을 이용하여 채널 추정을 수행하고 추정된 채널을 이용하여 등화를 수행한다. 기존구조는 심볼의 시작 시점의 채널이 추정되는 반면 제안구조는 심볼의 중간지점의 채널이 추정된다. 시변 채널 환경에서는 중간지점의 채널이 SC-FDE 심볼의 대표 채널로 더 적합하므로 이동속도가 높아 채널이 빠르게 변하는 경우 제안구조가 더 좋은 수신성능을 보인다. 제안하는 전송구조의 성능 향상은 컴퓨터 모의실험을 통해 확인한다. 모의실험 결과, 기존구조에서는 이동속도가 높으면 신호대잡음비 를 높여도 비트오류율이 낮아지지 않는 문제가 관찰되지만 제안하는 구조는 이러한 문제를 해결할 수 있다.

다양한 QoS를 수준을 갖는 위성통신망 서비스를 원활하게 지원하기 위한 효율적인 스케줄러의 설계는위성통신 네트워크의 가장 중요한 핵심 설계 요소 중의 하나이다. 많은 스케줄러가 제안되었지만, 기존 방법들은실시간 및 비실시간 트래픽을 다양하게 갖는 실용적인 액션 시나리오를 반영하기보다는 특정 액션 시나리오에만집중하여 지연 조건을 만족시키고 처리량을 감소시키려는 일차원적인 접근 방식만을 고려한 문제점이 있었다. 본논문에서는 가상시간( V T ) 및 가상종료시간( V F T ) 기반 우선순위 결정의 개념과 단말기의 QoS 클래스별로 별도의 버퍼를 도입하여 위성통신 시스템에 효율적인 스케줄링 알고리즘을 제안한다. 시뮬레이션 결과, 제안된 계획은기존 방법에 비해 처리량 성능 이 향상되었음에도 실시간 서비스의 지연 요구사항을 충족한 것으로 나타났다.

This paper proposes a deep-learning based signal to noise ratio (SNR) prediction technique for wireless communication environments. The communication system considered in this paper uses time division duplex (TDD), and receives signal using multiple antennas while transmits with only one antenna. Based on the SNR measurements when receiving in the past, we proposed a convolutional neural network (CNN) model to predict the SNRs for all antennas at the time of future transmission. If there is no received signal nor SNR measurement, the SNR measurements are filled by linear interpolation of neighboring two received SNRs. According to the simulation results, the wideband signals show better prediction performance than the narrowband signals. In the case of wideband, the proposed technique is about 0.37～0.98 dB superior to the conventional method for 20 km/h. For narrowband, the proposed one is better by 0.29～0.88 dB.

본 논문에서는 두 개의 톤 간격을 넓게 할당하여 전송하는 SC-BFSK(Split Channel Binary Frequency Shift Keying) 시스템에서 주파수 자원을 효율적으로 사용할 수 있는 확산 코드와 최적의 채널 간격을 찾는 방법을 제안한다. 제안하는 방법의 송신 신호는 PN(Pseudo Noise) 코드를 사용하여 전송한다. 송신 스펙트럼 관찰 결과 확산 이득이 큰 경우에는 규칙적인 스펙트럼을 보이지만, 확신 이득이 작은 경우에는 불규칙적인 스펙트럼을보인다. 확산 코드를 설정한 후 인접 채널 간격을 가변 하면서 간섭에 의한 BER(Bit Error Ratio)을 관찰하여 가장 작은 BER을 보이는 인접 채널 간격을 찾는다. 모의실험 결과, 확산 이득 별로 최적의 확산 코드가 다르며 인접 채널 간섭의 영향 또한 다르게 나타난다. 제안하는 확산 이득 별 확산 코드와 인접 채널 간격을 찾는 방법론은SC-BFSK에서의 통신 성능과 주파수 자원의 문제를 해결할 수 있다.

본 논문에서는 이동하는 통신 환경에서 합성곱 신경망에 기반하여 최적의 송신 안테나를 선택하는 방법을 제안한다. 고려하는 통신시스템은 다중 안테나를 가지고 있으며 양방향 통신은 시분할 이중화로 수행한다. 수신 시에는 모든 안테나를 이용하지만 송신할 때는 최적의 안테나를 선택하여 전송한다. 합성곱 신경망의 입력은과거 수신 시 측정한 신호대잡음비이다. 기존의 방법으로는 과거 신호대잡음비들의 평균값에 기반한 방법과 가장최근에 수신할 때의 신호대잡음비에 기반하는 두 가지 방법을 고려한다. 제안하는 방법과 기존의 두 가지 방법은컴퓨터 모의실험을 통해 비교한다. 이동속도와 수신할 확률을 바꿔가며 모의실험 한 결과 제안하는 방식이 광대역신호에서 가장 높은 안테나 선택 정확도를 보이고 협대역 신호에서는 가장 최근 신호대잡음비를 사용하는 기존의방법이 미세하지만 가장 우수하다.