Yifan Zhang

ConFi: Convolutional Neural Networks Based Indoor Wi-Fi Localization Using Channel State Information

As the technique that determines the position of a target device based on wireless measurements, Wi-Fi localization is attracting increasing attention due to its numerous applications and the widespread deployment of Wi-Fi infrastructure. In this paper, we propose ConFi, the first convolutional neural network (CNN)-based Wi-Fi localization algorithm. Channel state information (CSI), which contains more position related information than traditional received signal strength, is organized into a time-frequency matrix that resembles image and utilized as the feature for localization. The ConFi models localization as a classification problem and addresses it with a five layer CNN that consists of three convolutional layers and two fully connected layers. The ConFi has a training stage and a localization stage. In the training stage, the CSI is collected at a number of reference points (RPs) and used to train the CNN via stochastic gradient descent algorithm. In the localization stage, the CSI of the target device is fed to the CNN and the localization result is calculated as the weighted centroid of the RPs with high output value. Extensive experiments are conducted to select appropriate parameters for the CNN and demonstrate the superior performance of the ConFi over existing methods.

Non-Cooperative Wi-Fi Localization via Monitoring Probe Request Frames

Most Wi-Fi based localization algorithms are cooperative as user device is required to associate with an AP. However, user may not associate with AP in scenarios such as supermarkets which calls for non-cooperative localization. In this paper, the probe request (PR) frame sent by device is analyzed and the weighted kernel density estimation assisted Bayes (w-KAB) algorithm is utilized for localization. The PR frame is sent in a sparse manner in time and the probability distribution of its receiving signal strength is complicated due to channel misalignment. Therefore kernel density estimation is adopted in the training stage to estimate the distribution of signal strength accurately with a limited amount of training data. In the localization stage, a weighted naive Bayes algorithm is used to estimate the location of user. Experiments are also conducted using off the shelf devices to validate the performance of the proposed algorithm.

A Phase Difference Based Cooperative Spectrum Sensing Scheme for Cognitive Radio Network

The increasing scarcity of spectrum resources is one of the most challenging issues to cognitive radio systems in 5G era. Traditional schemes fail to gain the balance between accuracy and complexity, which are the two of the most significant parameters to evaluate the performance of the spectrum sensing. In this paper, in order to improve the sensing accuracy and reduce the computation complexity, we propose a novel cooperative spectrum sensing scheme based on phase difference is proposed. By using the mean of Phase Difference (PD) as the test statistics, the proposed PD mean detection is formulated for efficient spectrum sensing and its performance is analyzed under Rayleigh fading channel and Gaussian noise, which has a low complexity of O(K) and is immune to the noise uncertainty in contrast to the energy detection scheme. Moreover, to improve performance of the sensing scheme based on phase difference by a single CR, we consider the cooperative scenario with multiple CR nodes. Simulation verifies that our scheme obtains 3–4 dB gains comparing with energy detection.

A Novel q-Weighed Sequential Cooperative Energy Detection Method for Spectrum Sensing

As traditional spectrum sensing approaches unable to deal with the contradiction between detection accuracy and complexity in cognitive radio network, a novel q-weighed sequential cooperative energy detection method for spectrum sensing in time varying channel is proposed in this paper to achieve better performance with lower complexity. By adding the q- weighted log likelihood ratio (LLR) of the past local observations from previous sensing slots to the current LLR sequentially, cognitive radio nodes can aggregate the current and previous received energy values to yield the improvement of sensing performance. Moreover, we pose a q-weighted K-out of-N voting rule at the fusion center to minimize the total error probability. For different probability of primary signal for turning its state from active to idle, we employ corresponding different weighted value q to make the sensing scheme more flexible and efficient.

TOA Based Localization Under NLOS in Cognitive Radio Network

In this paper, we consider cooperative localization of primary users (PU) in a cognitive radio network (CRN) using time-of-arrival (TOA). A two-step none-line-of-sight (NLOS) identification algorithm is proposed for the situation where both NLOS error distribution and channel model are not available. In the first step the TOA measurements are clustered into groups. The groups with a dispersion higher than a predefined threshold are identified as NLOS and discarded. In order to make the threshold more reasonable, Ostu’s method, a threshold selection method for image processing is utilized. The second step is introduced to correct the error of possible surviving NLOS. To increase the accuracy of estimated position when line-of-sight (LOS) paths are limited, we proposed a result reconstruction method. Simulation results show that our algorithm can effectively identify NLOS paths and improve positioning accuracy compared to existing works.

A Novel Sequential Phase Difference Detection Method for Spectrum Sensing

As traditional spectrum sensing approaches in cognitive radio network unable to deal with the contradiction between accuracy and complexity, a novel sequential spectrum detection based on phase difference (SPDD) is proposed in this paper to achieve good performance with less complexity. The variance of phase difference of signal is utilized as the statistics to detect the signal under a realistic Rayleigh fading channel. Moreover, a variable sample size of proposed algorithm is conducted to minimize the complexity while maintained an acceptable performance. Simulation shows that our SPDD method yields about 2 dB gain over the conventional sequential energy detection. In addition, when the cutoff sample number is set to 1000, a substantial efficiency improvement is obtained compared to the fixed sample detection scheme.

Spatial Spectrum Holes in TV Band: A Measurement in Beijing

Spatial spectrum holes are areas where TV signal strength falls below a certain threshold and TV frequency can be utilized without license. In our measurement, we prove the existence of spatial spectrum holes considering shadowing and building penetration. To evaluate the influence of shadowing, two dimensional radio environment mapping (REM) is constructed for a (500,m,times ,530,m) area in the downtown. According to the REM, a maximum attenuation of 30 dB can be caused by building blockage and shadowing. To measure the loss of wall penetration, a three dimensional measurement is conducted in the outer and inner area of a 12-floor building. It is found that the wall attenuation approximately follows a normal distribution with a mean of 24.31 dB. The distribution of spatial spectrum holes is then plotted indicating spatial spectrum holes are abundant especially in the outskirts of Beijing.

Cooperative Interference Game in cognitive radio hidden terminal scenario

Hidden terminal problem in spectrum sensing is one of the most challenging problems in cognitive radio network (CRN). To tackle this problem, we propose a novel Cooperative Interference Game scheme in this paper. The scheme adopted full duplex (FD) mode to sense over multiple sub-bands in an iterative manner without extra sensing devices. The implementation algorithm of the proposed scheme is consisted of three modules: the formulation of the maximum transmit power limitation of second user (SU); the self-interference cancellation coefficient; and the optimal location of SU for an optimized low collision probability. Monte Carlo simulation proved that compared with cooperative spectrum sensing, the proposed scheme significantly improves the performance of spectrum detection and mitigates hidden terminal problem to a large extent with less energy consumption.

Low Complexity Joint Semiblind Detection for OFDM Systems over Time-Varying Channels

Orthogonal frequency division multiplexing (OFDM) modulation is widely used in communication systems to meet the demand for ever increasing data rates. In this paper, a low complexity joint semiblind detection algorithm for OFDM systems over time-varying channels is proposed based on the channel correlation and noise variance. The problem is relaxed to a continuous non-convex quadratic programming problem. Then an iterative method is utilized to deduce a sequence of reduced-size quadratic programming problems. These are solved by limiting the search in the 2-dimensional subspace. Furthermore, a low-bit descent search is employed to improve the system performance. Results are given which demonstrate that the proposed algorithm provides comparable performance with lower computational complexity than that of a sphere decoder.

Blind Polynomial Channel Estimation for OFDM Systems

Orthogonal frequency division multiplexing (OFDM) modulation is widely used in communication systems to meet the demand for ever increasing data rates. Characteristics of the transmitted signal can be employed for blind channel identification. In this paper, we propose a blind polynomial channel estimation algorithm using noncircular second-order statistics of the received OFDM signal. A set of polynomial equations are then formulated based on the correlation of the received signal. The solution of these equations provides an estimate of the channel coefficients. Results are presented which show that the proposed algorithm provides performance comparable to the least minimum mean square error (LMMSE) solution with low computational complexity. The performance is near-optimal for large OFDM systems.