Search

The World Health organization (WHO) stated that the concept of health has been widened to subjectively experienced dimensions such as fatigue and chronic fatigue syndrome (CFS). With the increasing pressure of the current life, persistent fatigue caused by sustained high-pressure work will not only be hazardous to health, but also give rise to unexpected consequences. In particularly, fatigue driving induced by long time driving has become a leading cause of accidents and death in the transportation. In this study, we investigate electroencephalography(EEG)-based fatigue detection of drivers through the spatial-temporal changes in the relations between EEG channels. EEG signals are firstly partitioned into several segments and the covariance matrices obtained from each segment are fed into a recurrent neural network to extract high-level temporal features. Then, the covariance matrices of whole signals are leveraged to extract spatial characteristics, which will be fused with temporal features to obtain comprehensive spatial-temporal information. Experimental results on a benchmark dataset showed that our method obtained an optimal classification accuracy of 91.042% and outperformed some state-of-the-art methods. These results indicate that our method is reliable and feasible for fatigue detection, which also provides a novel solution for EEG modeling.

Ubiquitous Electric Internet of Things (UEIoT) is the next generation electrical energy networks. The distributed and open structure of UEIoT is weak and vulnerable to security threats. To solve the security problem of UEIoT terminal, in this paper, the interaction between smart terminals and the malicious attackers in UEIoT as a differential game is investigated. A complex decision-making process and interactions between the smart terminal and attackers are analyzed. Through derivation and analysis of the model, an algorithm for the optimal defense strategy of UEIoT is designed. The results lay a theoretical foundation, which can support UEIoT make a dynamic strategy to improve the defensive ability.

Simultaneous location and parameter estimation of vital sign signal captured with a MIMO-FMCW radar sensor system is presented. The suggested system is able to deal with safety inspection emergency counter-terrorism and fire rescue. Micro-Doppler information, estimated from received signal, is used to classify heart and breath frequency of human. The Capon BF measurement algorithm is used to estimate accurate angles. The received signal synthesize precise angle information to get position of target. Measured data collected from MIMO-FMCW radar received encouraging results after these algorithms. Results of human object with deferent range or angle show effectiveness to achieve great estimation performance.

Unmanned aerial vehicle (UAV) swarms is an emerging technology that will expand the application of UAVs and present the new requirements for its reliability and robustness. However, in the current research for the task decision-making of UAV swarms, few papers have considered the system robustness. That will be hard to adapt to some tasks with high uncertainty. Therefore, based on our previous research, this paper proposes a task decision-making method for UAV swarms considering system robustness under two kinds of dynamic evolution processes. Fuzzy comprehensive evaluation is used to solve this model. The validity of the method is verified by a case study. The method presented in this paper can provide some guidance for the mission planning and formation size control for UAV swarms.

The group tracking method always apply to situations that a target can generate one measurement at most and multiple targets move in the same style. However, extended object appears as high-resolution sensors become available. This paper adopts the concept of group targets and proposes a group tracking method with adaptive gate in multiple extended object. As constant ellipsoid gate causes changing volume, this paper proposes the idea of making the volume changeable to calculate the ellipsoid gate for accepting into the group. The proposed method is compared with the LJPDA algorithm for multiple extended objects tracking. Tracking results demonstrate that two methods have equivalent performance while group tracking algorithm cost One-third of the time in tracking.

In this paper, we propose a novel sub-pixel motion extraction method, called as Two Dimensional Spatial Keystone Transform (2DS-KST), for the motion detection and estimation from successive noisy Occupancy Grid Maps (OGMs). It extends the KST in radar imaging or motion compensation to 2D real spatial case, based on multiple hypotheses about possible directions of moving obstacles. Simulation results show that 2DS-KST has a good performance on the extraction of sub-pixel motions in very noisy environment, especially for those slowly moving obstacles.

In this paper, a waveform design method utilizing the low maximum side-lobe in the synthetic range profile as the criterion for the random hopping frequency radar is proposed. Firstly, the center range ambiguity function which represents matched filtering output of transmitted signal is derived. And the output signal to noise ratios (SNR) of it and the synthetic range profile based on correlation processing are investigated. Then, the waveform design method based on uniformly splitting window function is proposed to improve the peak to side-lobe ratio of the synthetic range profile. Finally, numerous simulated experiments show that the proposed design method is effective.

In the condition of high PRF waveforms, several pulse repetition intervals are needed to receive the echoes from a target, which makes the range of the target ambiguous. In this paper, we propose a novel method for range imaging. In our method, the additional phase term produced by range ambiguity is derived, and then considered in establishing the match filter. The high resolution range profiles can be obtained by applying the match filtering step to the echoes. Simulation results demonstrate the validity of this method.

A random distribution over the bandwidth makes the random hopping frequency radar widely use in the synthetic wideband application as stepped-frequency radar. However, the traditional inverse fast Fourier transform is not able to achieve range compression due to the random modulation. In this paper, firstly, the echo of the random hopping frequency radar is derived, and an efficient range imaging method is proposed. Then, the range-velocity ambiguity function of random hopping frequency radar using correlation algorithm is depicted to reveal the coupling interrelationship of the range and Doppler. At last, Monte Carlo simulations are carried on to estimate the relation between side-lobe level and the number of frequencies. At the same time, the theoretical prediction formula provides a solid proof for it.

For the targets whose maneuverability is not too severe, the azimuth echoes in a range bin are modeled as quadratic phase signal after the translational motion compensation. The FrFT algorithm can obtain a focused image by motion parameter estimation and compensation. However, when a range cell contains more than one scatterer, the effects of sidelobes and interference terms will deteriorate the performance of FrFT method. To solve this problem, we add a selecting step for range cells in the imaging scheme. By selecting the dominant-scatterer range cells the estimation performance can be improved obviously. The effectiveness of the proposed method is illustrated and analyzed with simulations data.

In this paper, we focus on the 3D SC model reconstruction from data with wide azimuthal aperture at a single elevation. Since the existing method is difficult to implement for high-frequency signal or large-size target, we propose a novel method to solve this problem. In our approach, the 3D positions of the SCs are estimated from the 1D SCs via a randomized Hough transform (RHT) method. Then the scattering coefficients are refined via a linear least squares algorithm. Comparison with the existing method, the approach has advantages of fast speed, small storage, and high parameter resolution. Experiments demonstrate the effectiveness of the proposed approach.

Composite laminates are widely used due to the advantages of low density, high specific modulus and specific strength, strong internal damping and excellent chemical stability. Invisible damage by impact and fatigue such as matrix crack, internal delamination and fiber break may lead to sharp decline of structural strength and stability and finally makes composite structural failure. This paper proposes a damage index weighted average localization algorithm (DIA), which can realize a quick damage identification and damage localization of high accuracy. The proposed algorithm does not require damage imaging, and thus has the potential of realizing Structural Health Monitoring on line and in situ. Pre-experiments of real impact damages and artificial bonded bolt are done on composite laminates of small sizes. At last, to reduce experiment cost, the bolts are bonded on laminates to replace real impact damage for validation of DIA. Damage identification is successfully realized with damage localization error lower than 5 mm. A regular PZT arrangement with fewer sensors verifies that 6 sensors are enough for localization error of 20 mm and a random PZT arrangement with one or two sensors missing from the initial sensor array will not influence the localization accuracy.

Target tracking using unthresholded raw data under low signal-to-noise ratio circumstance, also referred to as track-before-detect, is a challenging task, especially for the case with varying target number and uncertain target dynamics. This paper deals with the problem of tracking multiple maneuvering targets using raw image observation. The multi-target state is formulated as random finite set and its posterior distribution is approximated by multi-Bernoulli parameters. Multiple model approach is proposed to accommodate the uncertainty of the possible target dynamics, and sequential Monte Carlo method is presented to implement the multiple-model multi-Bernoulli (MM-MeMBer) filter. The state estimates are obtained by combining the result of mode-dependent filtering for the Bernoulli components with high existence probabilities. Simulation results for multi-target track-before-detect application show the improved performance of the proposed method over MeMBer filters in the single-model fashion under the condition of equivalent computational complexity.

This paper adopts EP1K100QC208–3N FPGA chip as the main controller of intelligent automobile. We also added OV6620 camera module, parking space module and control of motor module. We use these modules to achieve the automatic parking system.

In this article, the tail dependence coefficient of the multivariate elliptical distribution and its properties are obtained. Furthermore, some specific tail dependence coefficients (TDC) and regularly varying elliptical distribution are given. From these results, it is easy to see that the TDC is only determined by its tail index and the correlation coefficient of the multivariate elliptical distribution. Finally, for comparing with the results in [4], some simulations are presented.