In order to further improve the accurate detection signal, reduce interference between signals, this paper designs a new type of signal detection algorithm for satellite communication systems, using stochastic resonance technology improve the signal-to-noise ratio of the input signal, the signal by using energy detection, double threshold, accurate judgment. The first step in the conventional energy of double threshold detection, the second step into the energy detection method based on stochastic resonance detection process. The experimental results show that this algorithm under the condition of low SNR signals effectively detect, promoted the whole satellite communication system performance.
Traditional evolutionary fault-tolerant scheme can effectively repair circuit faults, but for large-scale integrated circuits, the evolution process consumes a lot of time and it is difficult to meet the real-time requirements. In this paper, a real-time system fault-tolerant scheme based on improved chaotic genetic algorithm is proposed. The scheme uses a built-in test detection mechanism with feedback to detect the running state of the circuit in real time. When a fault occurs, normal system operation is maintained by the fault compensation mechanism. At the same time, the system uses the evolution repair mechanism to repair the faulty circuit. Evolution process uses an improved chaotic genetic algorithm, which can quickly converge to obtain a repair circuit through adaptive chaotic crossover and mutation. This paper builds a fault-tolerant system on the FPGA. In the experiment, the fault is randomly injected into circuit so that to simulate the actual circuit fault. The proposed algorithm and fault-tolerant scheme are used to verify the self-repairing ability of the system. The experimental results show that under real-time constraints, the repair rate of the fault circuit reaches 94%.
In image retrieval, the shape feature is one of the key features of image content description. At present, most of the widely used Fourier descriptors in shape descriptors are invariant in translation, rotation and scale expansion. But Fourier descriptors are susceptible to the location of the starting point. In this paper, an improved image retrieval method based on Fourier descriptors is proposed. First, the image is preprocessed and the edge of the image is extracted. Secondly, the starting point of the contour is determined by the minimum inertia axis. Then Fourier transform is used to get eigenvectors. Finally, the correlation coefficient is used to calculate the similarity. Experiments show that the Fourier Descriptor Image Retrieval Method based on the minimum inertia axis is more efficient than other methods in Swedish Leaf database.
In this paper, resource allocation problem for ultra-dense D2D communications is studied. In ultra-dense scenarios, the number of D2D user equipments is far bigger than the number of cellular user equipments. The dense user equipments increase the complexity of resource allocation problem. Firstly, the system model of ultra-dense D2D communications is described. Then the resource allocation problem of ultra-dense D2D communications is formulated. Next a fast resource allocation algorithm based on predatory search algorithm is proposed and analyzed. Finally, the analysis and simulation results validate that the performance of proposed scheme is very efficient and has a low algorithmic complexity. This scheme can be applied into the ultra-dense D2D communication networks.
The low-density parity check codes (LDPC codes) are block codes whose performances are close to the Shannon limit. LDPC codes have the strong ability for error correction. The decoding algorithm of LDPC codes has a great influence on their performances. The belief propagation (BP) algorithm is a commonly used soft decision decoding algorithm. The algorithm decodes by information iterations, and its complexity does not increase rapidly with the increase of code length. This paper mainly analyze the probabilistic domain BP decoding algorithm, log-domain BP decoding algorithm and minimum sum decoding algorithm, the bit error performance of LDPC codes under BP algorithm is studied, and the influence of the number of iterations on the BP decoding algorithm is also shown by simulation results.
For the purpose of Spacecraft Networking, considering the factor of the deployment of SPP within SOIS protocol system and the deployment of IPv6 in ground network with its development in the future, we present a set of scheme about design and realization, with software, of onboard router based on IPv6 and SPP. The design and realization of onboard router provides a new idea on constructing the Integrated Space-Ground Network, establishing the foundation of Spacecraft Networking in the future.
Secondary users in cognitive radio system use spectrum sensing technology to detect the primary users in the frequency band and use spectrum holes to communicate. Spectrum prediction technology is based on the existing spectrum sensing results to predict the future channel occupancy, so as to reduce the blocking rate, avoid malicious dynamic interference and other purposes. In this paper, a spectrum prediction method based on convolution neural network is proposed and some applications of this method in practical communication systems are given. This method can be trained in real time and has a certain adaptability to the dynamic environment. Using this method, the predicted results can be used to allocate resources reasonably, and the spectrum resource utilization rate is high. In addition, the time-consuming of broadband spectrum sensing can be shortened by combining the spectrum prediction method based on convolution neural network. At the end of this paper, the simulation results of spectrum prediction method based on convolution neural network are given and the efficiency of the algorithm is discussed.
Based on the negative temperature characteristics of threshold voltage and positive temperature characteristics of a multiple of thermal voltage, adding them with proper weight coefficient A voltage reference circuit was proposed with a zero temperature coefficient (TC). The device consists of pure MOSFET operated in subthreshold region and uses no resistors and bipolar transistors. The triple-branch current reference structure is adopted for independence of supply voltage instead of cascade structure and embedded operational amplifier structure with the merit of chip area and power consumption. Simulation results showed that based on standard CMOS 0.18 um process, the circuit can operate at 0.75 V supply voltage with the output voltage only 563 mV. The TC of the voltage was 17.5 ppm/℃ in a range from −40 ℃–125 ℃. The line sensitivity was 569.5 ppm/V in a supply voltage range of 1.2 V–1.8 V, and the power supply rejection ratio (PSRR) was 66.5 dB at 100 Hz. The power dissipation was only 187.4 nW.
Gesture recognition is one of the key technologies in the field of computer vision, and hand gesture recognition can be divided into static hand gesture recognition and the dynamic hand gesture recognition. This paper presents a new static gesture recognition algorithm based on hidden markov model. It uses two kinds of new shape features, the specific angle shape entropy feature and the upper side contour feature. They are firstly used for parameters training of hidden makov model, and then identify gesture categories hierarchically. In order to further improve the recognition effect for those small shape differences gesture, this paper adopts wavelet texture energy feature which can reflect the internal details of the gesture image, and makes the final correction estimation based on minimum total error probability. The experimental results show that the method has good recognition effects for gestures no matter the shape differences are big or not, and it has good real time performance as well.
The integrated space-ground network and its architecture are the focus and difficulty in research. Aiming at the problems of the incompatibility between satellite networks, satellite networks and ground networks, and the insufficiency of space network address resources, we propose an integrated space-ground network architecture based on the next generation Internet protocol IPv6. It combines the space communication protocols defined by CCSDS. The network layer protocol based on IPv6 is the foundation of the architecture. It ensures the interconnection and interoperability among inter-satellite networks and intra-satellite networks. The protocol of each layer in the architecture is designed. The IPv6 protocol, inter-satellite and intra-satellite transmission format, inter-satellite routing and intra-satellite communications are analyzed and designed. The experimental results show that the designed satellite router realizes the network communication and routing of IPv6 packets in intra-satellite and inter-satellite networks. It shields the differences between the satellite and ground network systems at the protocol level. It makes the satellite networks have good scalability and adaptability as with as the ground networks.
Complementary codes (CCs) have opened up a whole new frontier in Code Division Multiple Access (CDMA) techniques due to the ideal correlation properties. However, the equal gain combination must be satisfied in such CDMA system, which constrains the system performance over frequency selective fading channels. This paper aims to propose a set of weighted complementary codes (WCCs) to enable variable combination parameters while maintaining ideal correlation properties. Such new WCCs can provide power allocation with more freedom, and more importantly, an optimized power allocation can improve the bit error probability performance of CDMA systems as proved at the end of this paper.
The multi-source multi-sink maximum flow problem can be of great significance in guiding network optimization, service scheduling, and capacity analysis. With intermittent connectivity and time-dependence characteristics of satellite networks, the existing flow algorithms for multi-source multi-sink without temporal dimension involvement can no longer maintain high efficiency in satellite networks. To overcome the problem, we propose a novel dynamic multi-source multi-sink flow algorithm. Specially, the storage time-aggregated graph (STAG) is adopted to depict the time-varying properties of satellite networks. Then, a novel dynamic multi-source multi-sink flow (DMMF) algorithm is proposed to enhance the satellite networks' resource utilization. At last, the simulation is conducted, and the results with obvious network performance gain verifies our proposed DMMF algorithm.
The number of Beidou intelligent terminal is dramatically increasing recently. Only during 2017, there have been More than 200 mobile communication terminals supporting BDS applied for telecom equipment access to network license. It's important to evaluate the Beidou antenna performance of these intelligent terminals but currently there is neither specification nor common methods. Over-The-Air (OTA) test evaluates three-dimension (3D) radiated antenna performance in anechoic chamber to approach the real user experience. In this paper, an over the air test method for Beidou intelligent terminal is proposed, including standalone mode and communication assisted mode. The test scenarios and test procedures for radiated 3D Carrier-to-Noise (C/N0) pattern measurement and radiated sensitivity measurement are described. An over the air test system is developed in China Telecommunication Technology Lab (CTTL) and a number of intelligent terminals supporting Beidou are tested. The results indicate that although large numbers of terminals are claimed supporting Beidou, its Beidou antenna performance is relatively poor and when Global Position System (GPS) is not available, Beidou positioning cannot meet the user requirements currently. Thus, it is urgent to standardize the Beidou OTA technical requirements and test methods of intelligent terminals.
Nowadays, the fast development of the digital circuits results in a more and more high digital level of radar system. Especially, the development of the solid-state active module, the high-speed multi-digital A/D convertor, the direct digital synthesizer (DDS), and universal use of the high-speed digital signal processor provide an outstanding basis of the radar communication integration. Minimum Shift Keying Linear Frequency Modulation (MSK-LFM) is a novel multifunctional radar waveform. For all above, this paper proposes an integrated working mode based on positive and negative frequency modulation in radar communication integration. In the mode, the radar main station transmits positive linear modulation frequency signals, however, the communication affiliated station transmits negative linear modulation frequency signals. Their orthogonality causes less interference. Through the derivation, the method for the orthogonality improvement is obtained. The effectiveness of this working mode is proved by the simulations.
High resolution image can be obtained with backprojection (BP) algorithm, but at the same time, significant grating lobes will be brought in radar image and reduce the quality of image. This paper presents a through-the-wall radar (TWR) imaging method based on deep learning to improve the quality of radar image. A convolutional neural network was designed for TWR imaging, the radar image can be obtained as the output of neural network. The simulation and real data experiments demonstrate the effectiveness of proposed method.
In this paper, we investigate spectrum sensing relying on multiple satellites, which can achieve global seamless spectrum sensing, due to the feature of wide coverage. We conceive a pair of satellite based spectrum sensing schemes, namely hard combination oriented energy detection based spectrum sensing (HC-EDSS) and semisoft-combination double-threshold oriented energy detection based spectrum sensing (SD-EDSS). In the HC-EDSS scheme, secondary users send their decision results to fusion center in order to get the final decision. By contrast, secondary users not only send their decision results, but also send some individual information to fusion center in the SD-EDSS scheme. We also derive the closed-form of the probability of detection over satellite fading channel. In our performance evaluations, the conceived HC-EDSS and SD-EDSS schemes outperform the conventional single user oriented energy detection based spectrum sensing (SU-EDSS) in terms of its probability of detection. Moreover, the SD-EDSS scheme achieves the best probability of detection among them, demonstrating the advantage of increasing the accuracy of spectrum sensing.
How to filter fluctuant RSSI signal has always been a difficult problem in an indoor localization system. This paper provides an efficient indoor localization algorithm in a BLE5.0 based scan-broadcast network by building RSSI path-loss model without a great deal of fingerprints. This method builds a RSSI-Distance fading model between one position node (PN) and one markup node (MN) by maximum likelihood estimation (MLE) based on Gauss distribution of RSSI data. Then the rough fading model about RSSI in data collecting intervals of 1 m will be get. In this paper we reduce the distance intervals in 0.1 m by fitting of path loss model and making discrete samples of confidence intervals to improve the accuracy of localization. Finally, the whole fading regularly will be fixed and the location errors of PN will be determined by centroid model (CM). The results show that sampling interval with high precision can benefit the accuracy performance in an indoor localization environment.
In this paper, we propose a Device-to-Device (D2D) seed node cluster generation strategy based on coalitional game in social trusted D2D communication system. First, in the premise that the D2D seed node not harm the interests of other nodes and the cooperative power cost considered, a simple distributed algorithm is adopted to form independent and disjoint coalitions to maximize the throughput of the seed node cluster. Then the social trusted framework is introduced to make the segmentation of coalitions effectively meet the requirements of social security. The simulation results show that compared with the node cluster of the traditional cellular network and non-coalitional game, the system throughput of which is maintained at a higher level with social security ensured as well.
Hyperspectral image (HSI) anomaly targets detection is always applied for timeliness and onboard mission. For high detection accuracy, deep learning based HSI anomaly detectors (ADs) are widely employed in recent researches. However, their huge network scale for high-level representation ability leads to great computation burden for the onboard computation system. To decrease the computation complexity of the detector, a lightweight network is expected for the HSI AD. In this paper, by creating a multiobjective optimization with nondominated sorting genetic algorithm II (NSGA-II), an automatic evolution based deep learning network HSI AD (Auto-EDL-AD) is proposed to explore a lightweight network. The experimental results on an HSI dataset show that the proposed Auto-EDL-AD can generate an optimal network for the HSI anomaly detection which reaches up to 170
In this paper, a DOA estimation method based on co-prime array is proposed to resolve the coherent and incoherent hybrid sources. Firstly, with respect to the difference co-array of co-prime array, the desired units with corresponding contiguous intervals in the correlation matrix are extracted and rearranged into an augmented correlation matrix. Then we decorrelate the augmented correlation matrix by reconstructing matrix algorithm, forward spatial smoothing and forward-backward spatial smoothing algorithm. Finally, through MUSIC spatial spectrum searching on the basic of the decorrelated correlation matrix, DOA estimation towards sources is obtained. The simulation results show that the proposed method can achieve DOA estimation of the coherent and incoherent hybrid sources with more number than physical array. Through comparison, it can be concluded that the reconstructing matrix algorithm obtains a larger number of distinguishable sources and the error performance of which is better under low SNR. However, the spatial smoothing algorithms have a better estimation error performance in the case of low snapshot.
