The close relationship between the digital economy and the construction of an accurate spatio-temporal infrastructure,as well as the roles and challenges of communication and navigation integrated perception networks in realizing accurate spatio-temporal perception,is analyzed.China's breakthrough in communication and navigation integration research is introduced.The following topics are discussed:how embedded signal-to-noise positioning technology improves 5G network positioning performance,how the fusion of heterogeneous multi-network and multi-source data ensures positioning robustness,and the incremental effects of BeiDou's deep integration with 5G.Finally,further planning of communication and navigation integrated networks is proposed from the perspective of developing the digital economy.
Vehicle positioning methods using Visible Light Communication(VLC)are studied,and four methods based on VLC are studied,analyzed and compared.Firstly,based on the assumed system model and the received VLC signal mathematical model,the measurement process of TX position and system physical parameters adopted by each method are analyzed.An observation model is constructed using the geometric relations between these parameters and TX position to obtain vehicle position estimation.Then,the Cramer-Rao Lower Bound(CRLB)for each method in regard to position accuracy is obtained by the observation model based on VLC positioning method.Finally,under the VLC channel model with general finite propagation delay,Line of Sight(LoS)and Additive White Gaussian Noise(AWGN),the measurement of system physical parameters of each method is simulated for collision avoidance and platooning driving scenarios of real roads,and the CRLB of positioning accuracy for each method is evaluated based on the measurement results.
The characteristics of low foreground brightness and distorted background brightness in images collected under backlit traffic scenes result in low clarity,serious information loss,and poor identifiability of the collected images.To solve the above problems,a sub-regional enhancement method is proposed to study the different characteristics of the foreground/background of backlit images.Firstly,the maximum inter class variance(OTSU)method is used to segment the foreground and background of the backlight image;next,the LIME method is used globally for the backlight image to enhance foreground brightness while maintaining color distortion;then,the global histogram equalization results on the three RGB channels of background portion are individually mapped to the corresponding limited intervals,improving the contrast of the background.The Canny operator is used to detect the black edges at the stitching part between the foreground and background,and three adaptive filtering templates are generated based on the black edges to perform step-by-step mean filtering on the black edges,eliminating the black edges and improving the visual quality of the image.On a dataset CHD_B self-built in the laboratory,the proposed method is superior in terms of four commonly used objective evaluation indicators.Experimental results show that the proposed image enhancement algorithm can effectively eliminate the backlighting in images.
With the increasing complexity of communication environments,signal modulation recognition has become increasingly important.A modulation recognition method is proposed based on wavelet threshold denoising and time-frequency image detection to address the difficulty of digital signal modulation recognition at low signal-to-noise ratios.The method firstly converts the received real signal into an analytical signal and then denoises the analytical signal by the wavelet threshold method.Then the time-frequency reassignment technology is introduced to convert the denoised one-dimensional signal into a two-dimensional time-frequency image,and bilinear interpolation is used to scale the image to obtain a time-frequency image adapted to the size of the network input.Finally,the time-frequency map is input into the VGG network for training and recognition.The experimental results show that the proposed modulation recognition method performs well for modulation recognition under low signal to noise ratio.
Communication of deep space measurement and control usually adopts some measures to enhance the tracking performance of the equipment,such as increasing the carrier frequency and the diameter of the parabolic antenna,but it also brings some problems such as narrowing the half-power beam width of the antenna,reducing dynamic characteristics,and weakening the antenna's ability to resist gust disturbances,etc.The antenna deformation and pointing deviation caused by gale disturbance can lead to gain loss in deep space exploration,while decrease of gain can cause unstable measurement and control tracking,increased bit error rate of data,receiver lost lock,and inability to complete remote control commands,influencing the normal execution of measurement and control tasks.In view of the influence of gale disturbance to carry out deep space exploration mission of 35 m A-E dual-reflector antenna rotating platform in Chinese deep space station,the changing characteristics of annual wind-speed and wind-direction in the area are analyzed statistically by measured data,and the wind-speed and the data of antenna pointing angle error within the tracking segment are compared and analyzed,the antenna pointing angle error and its influence on the gain of the uplink and downlink measurement and control under different wind-speeds and wind-directions are simulated.Based on the existing technical facilities of the equipment,the coping strategies for executing real-time tasks are proposed,thus reducing the influence of gale disturbance on measurement and control tasks,and improved the ability of the task execution of the equipment.
To solve the problem that it is difficult to accurately and equivalently model the interference sources by the existing dipole modeling of phaseless near-field data,the double-sided iterative method used in the equivalent modeling of interference source of existing phaseless near-field data is improved.The improved method mainly uses more than two magnetic field amplitude data of different heights.Then,the selection of sides in the double-sided iterative method and the selection of the number of sides and starting side in the proposed method are discussed.Finally,the Root Mean Square Error(RMSE)of the reconstructed magnetic field of the interference source radiation is used as the basis for evaluating the accuracy of dipole modeling.The dipole models obtained using the improved method are compared and verified with the existing double-sided iterative method and other optimization methods,and the results show that the proposed method has higher accuracy.
To meet the decoupling requirements of the power transmission link and communication link in the wireless power and data simultaneous transmission system,based on the spatial decoupling of DD communication coil and power transmission coil,the integrated design of filtering feeding network and coil antenna is carried out,which improves the passband impedance matching characteristics and out of band suppression ability of the communication link,and further introduces the bandstop filter network to enhance the frequency domain decoupling with power transmission coil.When the coupling distance is 30 mm,the hardware of magnetic coupling communication filtering feeding antenna with the same aperture as 167 kHz wireless power transmission coil antenna is built,and the-3 dB passband bandwidth of 4 MHz is realized at the center frequency of 14.8 MHz.Compared with DD coil antenna,the decoupling antenna system improves the suppression ability of wireless power transmission signal by 85 dB.
Change detection of remote sensing images is an important research direction in the field of remote sensing,which plays an important role in many fields such as agriculture,disaster assessment,and urban construction.At present,most change detection tasks are completed using deep learning methods,but many existing deep learning networks have problems such as weak image feature extraction ability and inability to finely distinguish between change regions.A deep U-shaped network MCFFNet with multi-channel and multi-scale feature fusion is proposed.Firstly,the Unet network is extended to a three-channel structure,and the pre-classification feature information and fusion features of the corresponding scale feature images are obtained during the down-sampling process.Then,during the up-sampling process,the feature information of the corresponding scale is fused.Finally,the feature map is mapped into a single optimal change detection result map through convolutional activation and other operations.Experiments on the commonly used datasets CDD and WHU in the field of remote sensing image change detection have achieved higher change detection accuracy than the methods for comparison.
The uplink of an environmental backscattering communication system based on NOMA in the power domain often completes user pairing through distance.The dual channel fading of the passive backscattering system makes the remote edge devices unable to meet the communication power demands or the minimum Signal-to-Interference-Noise Ratio(SINR)constraint for decoding.In order to solve the above problems,a hybrid uplink NOMA scheme of OFDM and OFDM with Index Modulation(IM)is proposed,which is composed of a tag modulation scheme and a reflection coefficient adjustment scheme.OFDM-IM provides better energy efficiency and enables devices with low power level to meet communication conditions.The Backscatter Devices(BD)flexibly selects OFDM or OFDM-IM according to the received power level,and superimposes them in the power domain.By adjusting the reflection coefficient to ensure the power difference of superimposed signals,the receiver uses the power level to perform multi-user detection.The experimental results show that the scheme can effectively improve the success rate of remote user decoding,and the increase in the number of decoding users improves the system capacity.
For the problem of premature node death caused by unbalanced energy consumption of Low Energy Adaptive Clustering Hierarchy(LEACH)routing protocol in wireless sensor networks,an improved LEACH routing protocol based on genetic algorithm and ant colony algorithm is proposed.In the stage of clustering,the reasonable cluster head nodes are selected by genetic algorithm and the cluster is divided according to the distribution of nodes;In the data transmission stage,the cluster head node is made to choose the path with sufficient energy and short distance for data transmission through ant colony algorithm.The simulation results show that compared with the traditional clustering routing protocols LEACH and LEACH-C,the improved algorithm can make the energy consumption of the network more balanced and prolong the network life cycle.
To address the problems of gradient vanishing and limited feature extraction capability of traditional CNN spectrum sensing methods in deep network structures and to effectively avoid network degradation issues under deep network structures, this paper proposes a collaborative spectrum sensing method based on Residual Dense Network and attention mechanisms. This method involves stacking and normalizing the time-domain information of the signal, constructing a two-dimensional matrix, and mapping it to a grayscale image. The grayscale images are divided into training and testing sets, and the training set is used to train the neural network to extract deep features. Finally, the test set is fed into the well-trained neural network for spectrum sensing. Experimental results show that, under low signal-to-noise ratios, the proposed method demonstrates superior spectral sensing performance compared to traditional collaborative spectrum sensing methods.
To overcome the defect that Two-dimensional Parabolic Equation(2DPE)can only model two-dimensional communication links without considering lateral diffraction and backward reflection of electromagnetic waves near buildings in urban environment,a modified 2DPE model based on Deep Neural Network(DNN)is proposed.The three-dimensional model of urban buildings is built through digital elevation map,from which seven characteristics such as propagation distance,propagation angle,and building coverage are extracted to characterize the distribution of buildings on the propagation path and the deployment of transceiver antennas.Then combined with the measured data,the dataset for correcting 2DPE is constructed;through the training of DNN,the modified model of 2DPE is constructed to make it suitable for the prediction of radio wave propagation in a complicated three-dimensional environment.The simulation results show that compared to linear regression,support vector regression,and decision tree model,the calculation accuracy of the 2DPE correction model based on DNN is high in the three-dimensional propagation environment,and the prediction error on the test set is reduced by a maximum of 46.8%.
Land use transition is an important factor affecting ecological environment change.High-resolution land use data is selected from 2001,2011 and 2021 in Changchun.Firstly,land use transfer matrix and dynamic attitude of land use are used to analyze temporal and spatial changes of land use pattern.Based on ecological environment quality index and ecological contribution rate index,the response of ecological environment to land use transformation in Changchun is explored.The results show that:① Cultivated land and construction land are the core land use types required for the development of Changchun from 2001 to 2021.The cultivated land area increases first and then decreases,accounting for more than 75%of the total area,and the construction land area continues to grow.The grassland,water area and unused land area showed a decreasing trend to different degrees.The forest land maintained its ecological stability and changed little.The conversion among land types was most obvious in cultivated land and grassland.② In the past 20 years,the range of low-quality areas in central urban areas has gradually expanded,while that in sub-central urban areas has slightly decreased.From 2011 to 2021,the decline rate of ecological environment quality in Changchun slowed down compared with that in 2001 to 2011.The area of ecological environment improvement increased,but the overall ecological environment quality changed little.③ From 2001 to 2021,part of grassland in Changchun was deserted and water area was destroyed,resulting in severe ecological problems.Although the implementation of national ecological policies such as returning farmland to forest or grassland has improved the quality of urban ecological environment,the deterioration trend is still higher than the improvement trend.The research results can provide scientific basis for the optimization of land spatial pattern and the improvement of ecological environment in Changchun.
Considering the low detection accuracy caused by defect areas of PCB due to excessive background interference and the small scale of defective objects,a defect detection method of PCB based on attention mechanism and multi-scale fusion is proposed.Firstly,to enhance the saliency of defective object features and make the model focus more on object features,a 3D attention module is introduced in the feature extraction network based on YOLOv5.Secondly,to make full use of the multi-scale features of tiny defective object,a weighted Bi-directional Feature Pyramid Network(BiFPN)is introduced in the feature fusion network to reduce the loss of feature information of the defective object and improve the detection accuracy of the model for small defective object.Finally,the experimental results show that the method can accurately detect the defective objects in PCB images,and the average detection accuracy is improved by 3.9%compared with the original method while the real-time performance is ensured,which shows the effectiveness of the method.
Alzheimer's Disease(AD)is a neurodegenerative disease with high prevalence,which seriously affects the life of the elderly.Magnetic Resonance Imaging(MRI)can non-invasively obtain the morphological structure of the brain and reveal the pathological changes of the brain,which is currently the main means of AD diagnosis.Deep learning has powerful feature extraction and modeling capabilities in image processing,and the use of deep learning methods to process MRI for automatic diagnosis of AD has great application value.For three-dimensional brain images,the size and location of lesions are random and correlated,and local detailed features and global long-range dependency information are important.An attention based end-to-end network combining 3D CNN and Transformer is proposed to classify AD patients and normal individuals in response to such issues.Firstly,3D CNN is used to extract deep semantic feature-maps,which are then subjected to multi-scale feature weighted attention encoding and globally modeled by Transformer to obtain classification results.The method is validated on the AD dataset and publicly available 3D medical classification datasets.It is shown that the accuracy,sensitivity,and specificity are improved.The accuracy on the AD classification task reaches 95%,and the attention maps of the model highlight the disease-related areas such as the frontal lobe and the posterior cingulate cortex.The results show that the method has good classification performance and can be used as an automatic,effective,and convenient method for auxiliary diagnosis of AD and other medical tasks.
Based on the graph optimization framework,a multi-sensor fusion approach and an effective optimization method are designed,and a multi-sensor fusion Simultaneous Localization and Mapping(SLAM)scheme with robust localization effect is proposed,which can effectively deal with complex indoor and outdoor environments.The laser-vision back-end mapping fusion method is further developed to construct a point cloud grid map with a new form of map expression.At the same time,low-cost sensors are used to design and implement a high-performance low-cost backpack scanning system based on multi-sensor fusion,which can complete the self-localization and dense mapping in unknown environment as a whole,and reduce the track error per 100 meters caused by long-time movement to centimeter level on the low-performance CPU device.The multi-sensor fusion scheme proposed can match the existing mainstream schemes in terms of accuracy and computing power consumption,and is of great significance for obtaining accurate positioning results and rich spatial information of the system under various environmental conditions.
Mobile Edge Computing(MEC)is widely used in the new Internet of Things due to its low delay,low power consumption and high system capacity.To further make full use of the computing resource and reduce the power consumption,the scheme of computation offloading,computation processing sequence and power allocation are studied in green MEC system with D2D assisted offloading.To achieve green offloading,total power minimization of the terminal devices is put forward which considers many different application requests.To solve the original NP-hard problem,the Green Offloading and Application Processing Sorting Algorithm(GOAPSA)is proposed based on greedy method and bubble sort method.Simulation results show that the proposed algorithm can reduce the total system power consumption and increase system capacity by adjusting flexibly the processing sequence of applications and transmission power allocation and using D2D-assisted offloading.
