Abstract:The internal force of rigid column base joints of steel frames is very large under earthquakes, making it prone to be damaged and destroyed, so the configurations of column bases of rocking structures should be optimized to improve the failure modes and structural performance. Lifting semi-rigid frame column base joints with buckling restraint steel plates (BRSPs) were proposed for seismic energy dissipation. The hysteretic behaviors of column base joints with BRSPs were analyzed based on three parameters: thickness ratio α, width ratio β, and axial stiffness ratio ω between BRSP and column flange, and the results were compared with those of traditional rigid column base joints. Results show that the energy dissipation capacity of BRSPs and lateral resistance of structures were improved with the increase in α and β, and the plastic damage of column bases was reduced. The structural performance was optimal under the conditions that α=0.85, β=0.50, and ω was in the range of [0.3,0.5]. The ratio of plastic dissipated energy of BRSPs to the total plastic dissipated energy was more than 80%, and the plastic damage was limited to the replaceable BRSPs. Contact analysis between BRSP, backing plate, cover plate, and base beam was conducted, and it was found that the cover plate could not only effectively restrain the out-plane buckling of the BRSPs, but also provide friction energy dissipation, 7% of the total dissipated energy. The proposed column base has great energy dissipation capacity, which can improve the seismic performance and resilience of structures.

Abstract:To further enhance the anti-interception performance of traditional weighted fractional Fourier transform (WFRFT) from the perspective of security of wireless communication system, the two-dimensional weighted fractional Fourier transform (2DWFRFT) was introduced. Based on the original one-dimensional transform, the kernel algorithm of 2DWFRFT further improved the anti-scanning property of the single-parameter WFRFT by taking advantage of multiple horizontal and vertical dimensional transforms with different parameters. First, the signal generation method was given, and the matrix expressions for specific cases were derived. Then, the implementation complexity of the proposed scheme and the advantages of constellation rotation characteristics were described. In order to fully illustrate the anti-interception performance of the scheme from the perspective of physical layer security information theory, the ergodic secrecy capacity was analyzed under the multiple-input single-output (MISO) wiretap channel model. Simulation results show that compared with the one-dimensional WFRFT, the order of 2DWFRFT consisted of multiple parameter vectors, which could effectively improve the interception resistance only at the cost of limited computational complexity, and provide richer constellation features by breaking the point limits of constellation scrambling. Finally, compared with the traditional artificial noise method, the proposed scheme had a higher security capacity without additional signal power consumption.

Abstract:Organic micropollutant is an emerging problem in drinking water treatment. Oxidation by peroxymonosulfate activation is an efficient measurement for the degradation of organic micropollutants. However, the cost of current catalysts is high, and their preparation and regeneration processes are complicated. This work attempts to modify the natural montmorillonite with Fe(III) salt, which is used as the peroxymonosulfate activator for the degradation of bisphenol A (BPA) in drinking water. Results show that BPA was quickly adsorbed onto the surface of montmorillonite, and the removal rate was up to 96.4% in 1 h. The nanoscale interlayer of Fe(III)-modified montmorillonite (Fe-MMT) accelerated the transformation of electrons between Fe(III) and PMS, which was beneficial to the decomposition of PMS to generate HO· and SO₄·^- radicals, accelerating the degradation of BPA in water. In addition, the mesoporous layer structure of Fe-MMT could effectively inhibit the effects of humic acids on the PMS activation. Fe-MMT with easy preparation and regeneration processes can be used as the PMS activator in drinking water treatment.

Abstract:Due to the large number of parameters and large amount of calculation of deep convolutional networks, it is difficult to quickly and accurately deploy multi-scale target detection networks on many platforms with limited resources and power consumption. To solve this problem, based on the Python productivity for ZYNQ (PYNQ) framework, this paper realizes the IP core design and heterogeneous system architecture deployment of CTiny model, which is an anchor-free object detection model. First, a method of segmental quantization of the overall scaling factors in the convolution kernel was proposed, so that the pre-trained high-precision algorithm could be deployed on the field programmable gate array (FPGA) with low loss. Then, the system of the CTiny model was constructed based on the PYNQ framework, including ResNet backbone network, deconvolution network, and branch detection network. Finally, the time-consuming calculation such as picture preprocessing and post-processing was moved from serial ARM to parallel FPGA, further reducing the total processing time. Experimental results show that after deploying the CTiny model on the PYNQ-Z2 development board, the proposed quantization method achieved a mean average precision of 81.60% in the public optical remote sensing dataset NWPU VHR-10, which increased by 14.27% than truncated quantization. It has realized the requirement of deploying a tiny anchor-free object detection network with low loss. In addition, the processing time of post-processing was reduced from 9.228 s on the ARM side to 0.008 s on the FPGA side, which improved the speed of the detection model.

Abstract:Traditional wireless communication transmission and network access rely on radio frequency (RF) communication technologies. With the influx of massive mobile devices into the network, the problems such as spectrum scarcity and energy constraint pose challenges to the sustainable development of the new generation Internet-of-Things (IoT). Through the integration of traditional RF communication technology and emerging visible light communication (VLC) technology, an efficient and robust hybrid VLC/RF cooperative communication system with simultaneous lightwave information and power transfer (SLIPT) was proposed. To bridge the light-emitting diode (LED) source and the destination far away, an off-the-grid relay that can move randomly within source coverage was introduced. The end-to-end link could be divided into two hops (i.e. VLC link and RF link) through relay forwarding. Specifically, in the 1st hop, the optical signal received at the relay was separated into alternating current and direct current components for information decoding and energy harvesting respectively. Then, the harvested energy was used to forward the decoded source information to the destination by using RF in the 2nd hop. Considering the constraints imposed by the average power and peak power on the LED source, the distinct channel conditions across the two hops, the relay location, signal processing, and the energy harvested by the relay were taken into account in this study. On this basis, the end-to-end outage probability, data throughput, and energy efficiency of the system were analytically derived in closed-form expressions. Simulation results demonstrate that with proper parameter configurations, a balance could be reached between the transmission performances of the two hops. The proposed method could effectively promote the matching between the information flow and the energy flow in the system, thus significantly improving the reliability, throughput, and energy efficiency of the system.

Abstract:As a key ceramic matrix composite material, SiC fiber has good high temperature resistance and oxidation resistance properties. High temperature heat treatment of SiC fiber is necessary in the preparation and actual service of high temperature composite materials. In order to understand the performance of SiC fiber during high temperature process and high temperature service process, through the high temperature argon and air treatments of SiC fiber, the influence of heat treatment on the tensile properties and microstructure properties of SiC fiber was analyzed, and the process performance and service performance of SiC fiber were investigated. In this study, the domestic second-generation SiC fibers were subjected to argon and air heat treatments at 1 000 ℃ for 30 h, and the tensile properties of the fibers were tested by fiber electronic strength tester. The microstructure and chemical composition evolution of the SiC fibers were tested and analyzed by means of scanning electron microscope (SEM), X-ray diffraction (XRD), and Raman spectrometer. The relationship between the oxidation process of SiC fibers at high temperature, the variation of fiber surface, and the deterioration of the mechanical properties was discussed. Results show that the properties of the domestic second-generation SiC fibers were basically unchanged under the conditions of argon environment and heat treatment (1 000 ℃, 30 h), and the increase in crystallinity of SiC micro-crystal in the fiber was useful for its performance improvement. While the air environment and heat treatment (1 000 ℃, 30 h) led to the surface oxidation of the SiC fibers with SiO₂ layer and porosity, and the surface of the initially smooth fiber became roughness, resulting in decrease in the tensile strength of the SiC fiber by about 26%.

Abstract:Face image completion is an important image processing technique for reconstructing face images in the field of computer vision. The existing face image completion methods have the problem of unreasonable global semantics, which is mainly due to the lack of long-range transfer capability of the existing techniques that they are unable to reasonably transfer information from known regions in a broken image to occluded regions. To overcome the problem, a novel encoder-decoder face image completion network integrating style-aware and multi-scale attention was proposed under the framework of generative adversarial network (GAN). Specifically, the style-aware module was used to extract the global semantic information of an image, and the extracted information was employed to globally adjust the completion processing by rendering the encoding of the image level by level. The multi-scale attention module extracted patches of multi-scale features and performed a long-range transfer via matrix multiplication between a shared attention score and the extracted patches. Experimental results from the public dataset CelebA-HQ show that the style-aware module and the multi-scale attention module greatly enhanced the long-range transfer capability of the completion network. Compared with the existing state-of-the-art face image completion methods, the proposed model had significant improvement in various evaluation metrics. Meanwhile, the global semantics of the completion results were more reasonable and the completion effect was more natural under low lighting conditions.

Abstract:Shaking table tests on 36 telecommunication cabinets were conducted to investigate the seismic damage of cabinets under different earthquake intensities. Artificial ground motions were adopted as the inputs of the shaking table tests according to the standards. Results show that 13 cabinets met the standards, while other tested cabinets failed to meet the standards, and nine of them overturned during the tests. The quality of the telecommunication cabinets was uneven with poor seismic performance, which may pose a potential safety loophole for the normal operation of internet data centers. The acceleration amplification effects of the cabinets were compared and analyzed based on the recorded acceleration time-history curves during the tests. It was found that the acceleration amplification effect at the top of the unqualified cabinets was significantly larger than that of the qualified cabinets; the acceleration amplification effect at the middle of the cabinets was basically the same. The seismic fragility model of the telecommunication cabinets was developed based on the shaking table test data of 36 cabinets in this study. Peak floor acceleration was selected as the engineering demand parameter and the damage state of the cabinets was defined as hard to recover and loss of functionality. Test results show that when the peak floor acceleration was 1.118 1 g, the exceedance probability corresponding to the damage state of loss of functionality of the telecommunication cabinets was 50%. By comparing the results with the existing fragility curves in the literature, it was found that the dimensions of the telecommunication cabinets had little influence on their seismic fragility. The cabinets were more prone to be damaged and easier to lose functionality under bidirectional seismic excitations than under unidirectional seismic excitations.

Abstract:Sketch-based cross-domain image retrieval (SBIR) uses a sketch as query to retrieve the most similar image from the color image database. In this study, in order to better fuse the features from sketch and color image, a hybrid cross-domain joint network for sketch-based image retrieval was proposed, consisting of a sketch feature extraction branch and a color image heterogeneous feature fusion network branch. The network extracts the feature representations of sketch, positive and negative color image, and corresponding edge outline, and fuses the features of the color image and its sketch approximation (the edge outline of the color image) as the color image feature, which bridges the cross-domain gap between sketches and images. The network model parameters and network structure were further explored to optimize the purposed algorithm. Experiment on Flickr15K dataset shows that the proposed method performed better than other advanced image retrieval methods. The mean average retrieval accuracy of the proposed method was 0.584 8, which was 0.052 2 higher than the optimal value in other methods.

Abstract:In recent years, Janus membrane, whose one side is hydrophilic and the other is hydrophobic, has attracted much attention due to its special properties. Janus membrane is generally composited by hydrophobic material and hydrophilic material, while it has the defects of low interfacial bonding strength and irreversible conversion. In this study, a graphene-based Janus membrane with different wettability properties on both sides was prepared, whose hydrophobic surface could be reversibly transformed into super-hydrophilic surface. First, the graphene oxide (GO)/diphenyl ether system was treated with high-speed shear to prepare stabilized Pickering emulsion, and the emulsion was lyophilized and reduced with hydrogen iodide (HI) to fabricate graphene-based hydrophobic material with bowl-shaped and thin-walled spherical shell structure. Then, oxygen functional groups, such as COOH, CO and C—O, were attached to one side of the membrane by means of plasma treatment, so the hydrophobic graphene-based membrane was transformed into the super-wettable Janus membrane with different wettability properties on both sides, and the super-hydrophilic surface could be restored to hydrophobic surface by electric Joule heat treatment. Finally, the stability of the hydrophilic-hydrophobic conversion was tested by static contact angle, and the dynamic spreading process of water droplet on the graphene-based Janus membrane was recorded via a high-speed video camera. Results show that it cost 10 s to change the hydrophobic graphene-based membrane into the super-hydrophilic under plasma treatment, and the hydrophilic-hydrophobic conversion was completed in 8 s under DC voltage (20 V). After 10 cycles of hydrophilic-hydrophobic conversion, the hydrophilicity contact angle remained 0° and the hydrophobicity contact angle was 152°. It only took 20 ms for the water droplet to complete the spreading process on the hydrophilic surface of the graphene-based membrane. The chemical functional groups of the material were analyzed by XPS, and the microstructure was characterized by SEM and TEM. The structure of bowl-shaped spherical micro shell connected by the holes in the wall was the key to fulfill the hydrophilic-hydrophobic conversion of the material under plasma treatment and electric Joule heat treatment, and the pure graphene component was the base of the two treatments.

Abstract:The interference of noise, the unreasonable distribution of receivers and target, and the number of receivers all affect the coefficient matrix in the multi-station passive positioning model based on time difference of arrival (TDOA) and frequency difference of arrival (FDOA). Therefore, the coefficient matrix may be ill-conditioned in the actual solution process, which will affect the positioning accuracy to a large extent. A closed-form analytical method based on regularized constrained total least squares (RCTLS) was proposed to further improve the positioning accuracy when the coefficient matrix was ill-conditioned. The method was divided into two steps. In the first step, the positioning model based on RCTLS method was established to solve the positioning problem using TDOA and FDOA, and the regularization parameter was solved based on the criterion of minimizing mean square error. Then, the closed-form analytical solution of the model could be obtained by mathematical operation. The second step was to establish the equation of the estimation error obtained from the first step utilizing the constraint conditions, which was then solved. Finally, the obtained solution was used to modify the estimation results of the first step. Simulation results show that the root mean square error (RMSE) of the proposed method was lower than those of two-stage weighted least squares (TSWLS) and constrained total least squares (CTLS) methods through sacrificing unbiasedness, and the positioning performance was more stable than those of TSWLS and CTLS methods in the case of ill-conditioned coefficient matrix.

Abstract:To alleviate the problem of large error in the parameter estimation of linear frequency modulation (LFM) signal with high chirp-rates by traditional algorithms, a parameter estimation algorithm of LFM signal was proposed based on a new discrete fractional Fourier transform (DFrFT). A new dimensional normalization method combined with scale transformation was introduced into the DFrFT on the basis of the traditional discrete algorithm, and its normalization factor could change adaptively with the change in the transformation order. Based on the proposed DFrFT, the mathematical model of the parameter estimation algorithm of LFM signal was established. First, the LFM signal to be estimated was preprocessed by the new dimensional normalization method. The signal was converted from time-frequency domain to two dimensionless domains, and then the signal was processed by DFrFT. Next, the resolution of chirp-rate parameter estimation in high chirp-rate band was improved using the scale transformation characteristics of the new dimensional normalization method, and the chirp-rate optimize-range of the parameter estimation algorithm was theoretically deduced in comparison with the traditional algorithm. Finally, the effectiveness of the proposed algorithm was simulated under different chirp-rates and signal-to-noise ratios, and the results were compared with those of the traditional algorithm. Simulation results show that when the chirp-rate of the signal to be estimated was high, the proposed dimensional normalization method could effectively improve the estimation accuracy of the chirp-rate, and the performance improvement was more obvious in the environment of high signal-to-noise ratio.

Abstract:In the process of deep drawing, the common failure modes of fibre metal laminates are wrinkling, fibre tensile fracture, and interface delamination. Inspired by the deep drawing test of metal, fibre strain can be introduced into the forming limit diagram when fibre tensile fracture and interface delamination failure are the main failure modes of fibre metal laminates. However, since the shape of the traditional metal strip specimen used for deep drawing test is an obstacle for the calculation of fibre strain in fibre metal laminates, a new notched specimen was designed in this study. The formability of glass fibre/polyamide resin (Gf/PA) composite-aluminum alloy laminate was investigated by comparing the forming tests of traditional strip specimens and the proposed notched specimens. Results show that the proposed notched specimen could ensure the calculated fibre strain level within 6%. Through the classification of different fibre strain intervals, it could be clearly divided that the failure form of deep drawing of laminates was fibre tensile fracture or interface delamination. It was found that the forming limit diagram of the notched specimen consisted of three regions: forming safety zone, fibre tensile fracture zone, and interface delamination zone. In addition, the forming limit curves obtained from the notched specimens could combine the deformation evolution with the failure mechanism, and more effectively characterize the deep drawing formability of fibre metal laminates.

Abstract:In view of the problem that the moving target may be occluded or tracked under uncertain conditions during the aerial target tracking of UAV in complex scenes, resulting in the gradual damage, drift, and irreversible failure of the visual model, a long-term tracking algorithm for UAV was proposed. Firstly, the complementary classifier was designed for multi-feature adaptive fusion. The color histogram feature was used in Bayesian classifier, and the directional gradient histogram, grayscale, and color name features were used in the correlation filter. In combination with the advantages of multiple features, the robust appearance of the target was built to adapt to the complex scenes. Secondly, an adaptive spatial-temporal regularization term was added to the correlation filter. Local changes were introduced into the spatial regularization parameters for the implementation of the filter with low pixel credibility during learning. In temporal regularization, the learning of the filter was adaptively adjusted according to the global response, and the initial filter was used to constrain the update range, which effectively prevented filter degradation while mitigating boundary effects. Finally, a re-detection module was added to make sure the accuracy of the tracking process. Experimental results show that the proposed algorithm could adapt to the complex scenes of UAV aerial photography, alleviate boundary effects, and prevent filter degradation. In comparison with similar mainstream algorithms, the proposed algorithm could still meet the real-time requirements and achieve better tracking effect even when the target experienced serious occlusion or moved out of view.

Abstract:Vehicle-to-vehicle (V2V) communication is an important component of the intelligent transportation system (ITS), and tunnel is an essential application scenario of ITS. Considering the reflection of microwaves by tunnel walls, the diffraction by vehicles, and the dense and complex layout of the tunnel ancillary facilities, it is of great significance to investigate the propagation characteristics of V2V channel in tunnel scenarios. In this study, the V2V channel was measured by means of a broadband channel sounder in the tunnel environment. The carrier frequency and bandwidth were 5.2 GHz and 120 MHz respectively. Regarding to different situations, the measurement was divided into three types: line-of-sight (LOS), obstructed line-of-sight (OLOS), and non-line-of-sight (NLOS). Based on the measured data, the path loss models and received signal amplitude fading distributions were established for different scenarios, so as to analyze the channel fading characteristics. Results show that the path loss index and shadow fading in NLOS were larger than those in LOS and OLOS. There were no significant changes in shadow fading between LOS and OLOS. The amplitude fading of the received signal in both LOS and OLOS scenarios followed the Rice distribution, and that in the NLOS scenario followed the Rayleigh distribution. The root-mean-squared delay spread conformed to the log-normal distribution in all the three scenarios. The shadow fading and time-domain dispersion were larger in NLOS. The K factor decreased with increasing distance, indicating that the K factor is dependent on the distance.

Abstract:When the existing spectrogram separation methods are used for acoustic scene classification research, the classification accuracy of these methods is not high. To solve the problem, an acoustic scene classification method based on Mel-spectrogram separation and long-distance self-calibration convolutional neural network (LSCNet) was proposed. Firstly, the working principles of spectrogram harmonic/percussive-source separation were presented. A Mel-spectrogram separation algorithm was proposed, which can separate the Mel-spectrogram into harmonic components, percussive source components, and residual components. Then, LSCNet was designed combining self-calibration convolutional network and residual enhancement mechanism. The model adopts frequency domain self-correction algorithm and long-distance enhancement mechanism to retain the original information of the feature map, strengthens the correlation between deep and shallow features through residual enhancement mechanism and channel attention enhancement mechanism, and combines multi-scale feature fusion module to further extract the effective information of the output layer in model training. Finally, acoustic scene classification experiments were conducted on Urbansound8K and ESC-50 datasets. Experimental results show that the Mel-spectrogram residual components (MSRC) could specifically reduce the influence of background noise, thereby indicating a better classification performance. The LSCNet could realize the attention to the frequency domain information in the feature map, and its best classification accuracy reached 90.1% and 88% respectively, which verified the effectiveness of the proposed method.

Abstract:To understand the in-plane stability of a new type of double-box open-spandrel circular steel arch, the in-plane elastic buckling and elastic-plastic stable bearing capacity of the steel arch were investigated by theoretical deduction combined with finite element numerical simulation, the influence of shear force on the failure modes of arch section was analyzed, and the design method for the in-plane stable bearing capacity of the steel arch was proposed. First, according to the distribution of shear force on the arch section, the influence of shear deformation of double-box open-spandrel circular steel arch and local shear deformation of chord web on the in-plane elastic buckling was studied. The formula of pure compressive elastic buckling load of double-box open-spandrel circular steel arch was derived considering double shear. Then, based on the design principle of axially compressed column of steel structure, stability coefficient and regularized slenderness ratio were introduced to draw the stability curves under pure pressure. Finally, the stable bearing capacity design formulas of the integral failure mode of double-box open-spandrel circular steel arch under several common load conditions were analyzed. The arch structure analyzed in this paper is novel, and the calculated results of the proposed elastic buckling load formula were in good agreement with the finite element analysis results. At the same time, the binomial formula of axial force and bending moment was used to check the overall stable bearing capacity. The results can be used as a reference for future scientific research and practical engineering design, which has important practical significance.

Abstract:In order to improve the practicability of segmentation algorithm in weld defect detection, a lightweight weld defect evaluation network MYNet was proposed. In the network, the lightweight residual structure could reduce the amount of calculation of the model, the feature pyramid network (FPN) combined with multi-layer visual fusion mechanism could improve the segmentation ability of the network, and the parallel mask mechanism could obtain a fast and high-quality defect segmentation mask. The open source computer vision library platform OpenCV was introduced to calculate different defect areas by pixel threshold, and Tencent’s ultra-high-performance mobile platform reasoning framework was introduced to accelerate the forward reasoning speed of the model in the central processing unit. In this study, a digital artificial intelligence evaluation device was built with the ARM Cortex-A72 architecture as the control core, and a suitable lightweight 64-bit Linux system was deployed for defect detection to verify the feasibility of the proposed weld defect evaluation algorithm. Experimental results show that the model could effectively locate and learn different types of defect features. The network evaluated the defect area and location information with the accuracy of 94.64%. Compared with the MS R-CNN network which has high accuracy but requires a large amount of calculation, the accuracy of the proposed method was only reduced by 1.93%, while the parameter weight was only 1/14 of the MS R-CNN network, and the computing power required for network execution was lower. The lightweight residual structure increased the forward reasoning speed of the network by 309%, and it only took 1.7 s to complete the task of segmentation of weld instances on the low-cost hardware based on the advanced RISC machine (ARM) architecture. The method proposed in this paper can effectively learn and evaluate X-ray weld defect images, and the algorithm applied to the evaluation device reduced the cost of welding quality inspection.

Abstract:In contrast with the Shannon sampling which requires large-scale samples, the compressive sampling has unique advantages in energy-efficient representation of video signals. When the sampling subrates of a keyframe and a non-keyframe are inconsistent, the existing interframe patch matching algorithms show unstable recovery quality under different subrate combinations. In order to fully utilize the temporal correlation between frames, a subrate-adaptive interframe patch matching algorithm was proposed for video compressive sensing reconstruction, where the differential interframe matching mechanism was performed according to the relative change between keyframe sampling subrate and non-keyframe sampling subrate for better adaptation of video bitstream generated from different subrate combinations. Firstly, the measurements of each frame were reconstructed to obtain the corresponding intraframe results respectively. Then, the adaptive interframe reconstruction was implemented according to the growth ratio of the keyframe sampling subrate to the non-keyframe sampling subrate. In the case of low growth ratio, the current frame selected the nearest keyframe and the non-keyframe in the same direction as its co-directional double reference frames for patch matching reconstruction. In the case of high growth ratio, the current frame gradually selected the reconstruction results of multiple reference frames for multi-stage patch matching reconstruction. Compared with the existing reconstruction algorithms, the proposed algorithm could achieve more stable reconstruction performance during the typical framework of video compressive sensing, and thus the recovery quality of video sequence was consistently improved.

Abstract:Translucent photovoltaic skylights have unique synergistic characteristics of shading and power generation, which have been widely used in public skylights of tall buildings. Its variable shading and light transmission characteristics aggravate the complexity of natural lighting mechanism when applied to skylights in tall and large public buildings. This paper aims to improve the natural lighting problem in the tall and large space of railway passenger station waiting halls, and reveal the influence mechanism of the key design parameters of translucent photovoltaic skylights on the natural lighting. In this paper, Beijing is selected as the representative area in northern China, where solar radiation level is moderate, and natural lighting shading and heat gain need to be taken into account. Based on extensive investigation, the typical translucent film photovoltaic skylight of high space waiting hall of typical railway station was selected as the research object, and the DAYSIM tool was used to complete the dynamic simulation and analysis of indoor light environment throughout the year. The study verified the decisive effect of roof area ratio of photovoltaic skylights on the quality of the indoor light environment. Results show that the critical values of roof area ratios of centralized skylights and “one column and many rows” distributed skylights were 40% and 15% respectively. A suitable minimum area ratio of skylights was recommended for the centralized application of photovoltaic modules with different visible light transmittances. Comprehensive design suggestions for the “one column and many rows” distributed skylights were provided, including the number of longitudinal skylights and the vertical and horizontal dimensions. Based on this, the natural lighting optimization design strategies for translucent photovoltaic skylights were proposed for the two typical distribution modes respectively.

Abstract:To further improve the convergence speed of the intelligent jamming decision-making algorithm based on value function in reinforcement learning and enhance its effectiveness, an improved Q-learning intelligent communication jamming decision algorithm was designed integrating the efficient upper confidence bound variance. Based on the framework of Q-learning algorithm, the proposed algorithm utilizes the value variance of effective jamming action to set the confidence interval. It can eliminate the jamming action with low confidence from the jamming action space, reduce the unnecessary exploration cost in the unknown environment, speed up its searching speed in the interference action space, and synchronously update the value of all actions, thus accelerating the optimal strategy learning process. The jamming decision-making scenario was modeled as the Markov decision process for simulation. Results show that when the correspondent used interference avoidance strategy against the jammer to change the communication channel, the proposed algorithm could achieve faster convergence speed, higher jamming success rate, and greater total jamming rewards, under the condition of no prior information, compared with the existing decision-making algorithms based on reinforcement learning. Besides, the algorithm could be applied to the “many-to-many” cooperative countermeasure environment. The action elimination method was used to reduce the dimension of joint jamming action, and the jamming success rate of the proposed algorithm was 50% higher than those of the traditional Q-learning decision algorithms under the same conditions.