Abstract:In repairing murals based on sparse representation， the dictionary construction is single， and the restoration of details is inadequate. Therefore， an improved sparse mural restoration algorithm using joint adaptive learning of multiple dictionaries is proposed. First， a non-subsampled shearlet transform （NSST） is used to perform multi-scale decomposition on the mural image to obtain low-frequency structural components and high-frequency texture components， overcoming the problem of neglecting the differences in texture and structure information in mural restoration using sparse representation. Then， a sparse method of multiple dictionary adaptive learning is proposed. The low-frequency texture image is clustered based on the similarity of features between pixels to construct multiple sparse sub-dictionaries， and the low-frequency component restoration is completed through singular value decomposition and split Bregman iteration optimization. Then， the pulse-coupled neural network mechanism is introduced to restore the high-frequency structural sub-band image of the mural image. Finally， the NSST inverse transform is used to merge and complete the restoration. Experimental results on actual murals show that the proposed algorithm effectively preserves significant information， such as the structure and texture layers of the mural image， and achieves better visual effects and objective evaluations than the compared algorithms.

Abstract:Multi-focus image fusion is an prominent branch of image fusion， which is widely used in microscopic imaging. Aiming at the problems of unclear texture details and misjudgment of focus areas in multi-focus fusion， this paper designs a global information encoding-decoding network from the perspective of global attention of spatial and channel information， combined with the shifted window self-attention mechanism in Swin Transformer and deep separable convolution. The comprehensive loss function is used to perform unsupervised learning of image reconstruction tasks. From the perspective of the importance of feature neighborhood information， an improved Laplacian energy sum function is introduced to discriminate the image focusing-properties in the feature domain， and the fine-grained effect of image focusing region discrimination is enhanced. Compared with seven classical image fusion algorithms， the proposed algorithm achieves advanced fusion performance in both qualitative and quantitative analysis and has a higher fidelity effect on the focus area information of the original image.

Abstract:Sonar equipment can produce severe speckle noise interference and degrade image quality when acquiring underwater images due to seafloor reverberation， scattered echoes and equipment. To overcome this challenge， a Non-subsampled contourlet transform （NSCT） sonar image denoising based on neutral ensemble with bilateral filtering is proposed. The method explores a clustering denoising method suitable for dealing with high-frequency subband noise of non-subsampled contourlet transform decomposition， where the neutrosophic set domain is applied as clustering to each layer of high-frequency sub-images， represented by true subsets， uncertainty sets and false sets， to achieve separation of noise and retention of valid information； then bilateral filtering is used to remove the noise mistaken as image information in true subsets； bilateral filtering for low-frequency coefficients smoothing is performed using bilateral filtering. Finally， the processed low-frequency subbands and high-frequency subbands are reconstructed by NSCT to achieve the purpose of removing speckle noise from the image. The experiments show that the proposed method can improve the visual effect of the image， improve the evaluation indexes such as mean square error and peak signal-to-noise ratio， and is also suitable for removing high-density noise.

Abstract:Aiming at the problem that the existing UAV detection algorithms cannot simultaneously take into account detection speed and accuracy， a lightweight UAV detection algorithm ，i.e.，Tiny Drone Real-time Detection-YOLO （TDRD-YOLO） based on YOLOv5s， is proposed in this paper. Firstly， the multi-scale fusion layer and output detection layer of YOLOv5s are used as the neck network and head network， respectively. MobileNetv3 lightweight network is introduced to reconstruct the original backbone network， and the channel behind the backbone network is compressed on the basis of the original YOLOv5s to reduce the size of the network model. Secondly， the attention mechanism of the Bneck module in the backbone network is modified from SE to CBAM（Convolutional Block Attention Module）， and the CBAM is introduced in the neck network to make the network model pay more attention to the target features. Finally， the activation function of the neck network is modified as h-swish to further improve the accuracy of the model. Experimental results show that the average detection accuracy of the TDRD-YOLO algorithm proposed reaches 96.8%. Compared with YOLOv5s， the number of parameters is reduced by 11 times， the detection speed increases by 1.5 times， and the model size is reduced by 8.5 times. Experiments show that the proposed algorithm can greatly reduce the model size and improve the detection speed while maintaining good detection performance.

Abstract:For the uplink resource allocation problem of cognitive heterogeneous cellular networks， a resource allocation algorithm based on bandwidth and power constraints is proposed and solved using an improved swarm intelligence algorithm. Based on the characteristics of cognitive radio technology， a range of bandwidth and power allocation values for cognitive home users are derived， and more resources are allocated to other users under the guarantee of satisfying user quality of services （Quality of Services，QoS） to enhance the transmission demand of users in the network and relieve the uplink access load of the network. To address the shortcomings of the bottleneck swarm algorithm such as low convergence accuracy and slow convergence， the crazy operator and dynamic elite learning factor are introduced into the leader and follower， respectively， to improve the algorithm's optimality-seeking efficiency and optimality-seeking accuracy. The improved Salp swarm algorithm is solved for the resource allocation algorithm based on bandwidth and power constraints. Simulation experiments show that the resource allocation algorithm with the introduction of bandwidth and power constraints can be effective in improving network performance， and it can effectively improve system efficiency and user access fairness under the condition of ensuring user QoS.

Abstract:In unmanned driving scenes， dynamic objects can significantly affect the global accuracy and robustness of simultaneous localization and mapping （SLAM） systems. Most existing laser SLAM systems are prone to odometry drift， positioning failure， and mapping ghosting in dynamic environments. To address these issues， this paper proposed a semantic laser SLAM system for dynamic scenes that integrates a lightweight PointPillars target detection network and a multi-object tracking method. The system first uses the PointPillars network to obtain the bounding boxes of potential dynamic targets and filters the feature points within these bounding boxes to obtain a preliminary estimation of odometry. Secondly， the system constructs a temporal sliding window based on the tracking results obtained by the multitarget tracking algorithm， which is based on the uniform velocity Kalman Filter. This enables the establishment of a robust and efficient spatiotemporal association of target-level data， which removes dynamic objects and recovers static targets， further optimizing the odometry. Finally， the proposed method is compared with the state-of-the-art methods on KITTI and NUSCENES datasets in challenging dynamic environments. The experimental results demonstrate that our system significantly improves the accuracy and robustness of odometry and global mapping， while the system also maintains real-time performance，which meets the requirements of autonomous robot systems and intelligent transportation applications in dynamic scenes.

Abstract:Deep learning （DL） based speech enhancement methods can be divided into time domain methods and frequency domain methods， each of which has its own pros. To make full use of the advantages of methods in both domains， a cross-domain speech enhancement model based on the neighborhood adaptive attention mechanism is proposed. The model enhances the input waveform and spectrum at the same time， and the final enhancement result is obtained by cross-domain fusion of the enhancement results in time domain and frequency domain. In order to take advantage of the information complementarity between the enhanced results in two domains， an information communication module is proposed to realize the information exchange between the enhanced results. In order to improve the feature extraction ability of the time-domain and the frequency-domain enhanced models， and to make full use of the signal characteristics of the two domains， the neighborhood adaptive attention module is proposed. The neighborhood adaptive attention module adaptively aggregates local self-attention with different neighborhood sizes according to the input information and then models the stationary features of different scales. The experimental results show that the complementary characteristics of waveform and spectrum can be effectively utilized to further improve the enhancement performance by adding the neighborhood adaptive attention module and cross-domain information exchange and fusion module.

Abstract:Aiming at the problems of high hardware resource consumption and low output accuracy in CORDIC algorithms， this design proposes an improved Coordinate Rotation Digital Computer（CORDIC） algorithm based on interval merging iteration. Based on the two-stage CORDIC algorithm， the algorithm uses interval merging iteration to complete the merging iteration operation in the second stage. For the truncation error caused by shift operation in merge iteration， interval merge iteration reduces the data error and resource consumption generated during the merge iteration process by reducing the size and number of data shifts. The simulation results show that the improved CORDIC algorithm not only retains the good characteristics of the two-stage algorithm in low latency but also reduces register consumption by 36.8% compared with the basic algorithm， 14.8% and 9.5% compared with the three-segment and two-stage algorithms， respectively. When a 16 bit output bit-width is given， the average error of the improved algorithm is reduced by 37.0% compared with the basic algorithm， 19.4% and 24.5% respectively compared with the three segment and two segment algorithms. Therefore， it is more suitable for modern digital communication with high speed， high accuracy， and low consumption.

Abstract:A new BDDEE （Blind Detector based on Difference of Extreme Eigenvalues） multi-antenna spectrum sensing algorithm based on the difference between the extreme eigenvalues of the received signal sample covariance matrix （SCM） is proposed. It uses the ratio of the difference between the maximum and minimum eigenvalues of the SCM and the average energy of the received signal as the sensing decision. The proposed BDDEE algorithm breaks away from the dependence on the noise variance in the detection process and does not need to use the relevant parameters such as the primary user signal and the wireless transmission channel. On this basis， using the results of the ordered eigenvalue distribution of the finite-dimensional Wishart random matrix， an accurate analysis and calculation method for false-alarm probability and decision threshold is proposed theoretically. Furthermore， considering the limitation of computing and storage resources of secondary users， a decision threshold calculation method with low computational complexity is proposed by combining the decision threshold corresponding to the maximum and minimum eigenvalue limiting distribution by using the distribution theory of limiting eigenvalues in the high-dimensional Wishart random matrix. From the comprehensive consideration of detection performance and false-alarm performance， the proposed BDDEE algorithm has better sensing performance than the traditional CAV （Covariance Absolute Value）， MME （Maximum Minimum Eigenvalue） and DMME （Difference between the Maximum and the Minimum Eigenvalues） algorithms， and can obtain more robust detection results under the condition of limited sample number， which is verified by the various numerical simulation results.

Abstract:The effects of quenching water temperature （20~60°C） on the structure， strength and toughness of 7B50 aluminum alloy plate were investigated by using room temperature tensile test， fracture toughness test and the combination of scanning electron microscopy and transmission electron microscopy （TEM）. Besides， the sensitivity of different aging treatment systems to the quenching water temperature was also studied. The results show the room temperature tensile properties of 7B50 alloy are not sensitive to the quenching water temperature. However， the fracture toughness tends to increase as the quenching water temperature decreases. When the quenching water temperature decreases from 60°C to 20°C， the fracture toughness of the peak-aged and over-aged alloys increase by 12.9% and 11.4%， respectively. The fracture toughness of the peak-aged alloy is more sensitive to the quench water temperature when the quench water temperature is higher （40 °C and 60 °C） compared with the over-aged alloy. TEM obseration results reveal the relationship of the fracture toughness change due to the grain boundary microstructure. As the quench water temperature decreases， the size of the grain boundary precipitation phase of the quenched alloy gradually decreases. Furthermore， the size of the grain boundary precipitates decreases and the width of the precipitates free zone of the aged alloy becomes narrower. Especially， the grain boundary precipitates size and the width of the grain boundary precipitates free zone of the peak-aged alloy varied more significantly with the temperature change.

Abstract:To meet the requirements of the outsourced concrete for the relevant project， C60 fine stone Self Compacting Concrete （SCC） was prepared. The influencing mechanisms of water-binder ratio and anti-cracking agent on the workability， mechanical properties， volumetric deformation performance， and durability of fine stone SCC， such as resistance to sulfate erosion， chloride ion penetration， freeze-thaw， and carbonation， were investigated， and the performance change laws of fine stone concrete were also explored. The influence of different factors on the microstructure of concrete was investigated by Scanning Electron Microscope （SEM） and Nuclear Magnetic Resonance Spectroscopy （NMR） tests， and compared with the macroscopic properties. The results show that the water-binder ratio on the performance of fine stone concrete is similar to that of ordinary concrete. There is an optimum amount of anti-cracking agent， and the right amount of anti-cracking agent can generate Mg（OH）2 to fill matrix pores， which staggered grows with flocculent C-S-H and acicular AFt， improving the pore structure. It can also improve the mechanical properties and durability of fine stone concrete. However， when the dosage of the anti-cracking agent is too high， the compound erosion effect decreases the sulfate resistance of SCC， and the optimum content is 8%.

Abstract:Considering the effects of water to binder ratio （0.40， 0.45， 0.50）， NMK contents （0， 1%， 3%， 5%， 7%）， superplasticizer dosage （0， 0.04%， 0.08%， 0.12%） and temperature （10°C， 20°C， 30°C， 40°C）， the static yield stress test method and small amplitude oscillation shear test method were used to obtain the parameters of static yield stress and storage modulus of nano-metakaolin （NMK） cement paste. The internal structural build-up of NMK cement paste in early hydration （10~90 min） was discussed by the evolution of static yield stress and storage modulus with time. The test results show that NMK promote the internal structural build-up of cement paste in the early hydration period. Within 90 min of cement hydration， the average growth rate of static yield stress of 3% NMK cement paste was about twice that of ordinary cement paste； the average growth rate of storage modulus of 7% NMK cement paste was about 4 times that of ordinary cement paste. The higher the water to binder ratio or the higher the superplasticizer content， the lower the growth rate of internal structural build-up of NMK cement paste； NMK accelerates the build-up of internal structure of cement paste in different temperature environment； an increasing temperature increases the growth rate of static yield stress and storage modulus of NMK cement paste.

Abstract:The effect of the concentration of silane coupling agent KH560 on the dispersion and placement stability of the vitrified bond was investigated， together with the mechanism of modification. Furthermore， cBN grinding tools were prepared with vitrified bond of different dispersion， and the influence of modification methods of vitrified bond and storage time on the micro-mechanism， mechanical properties and grinding performance of cBN grinding tools were studied. The results show that when the KH560 addition was 3.0%， the vitrified bond powder displayed the best dispersion and placement stabilization. The microstructure of CBN abrasive tools prepared by using 3.0%KH560 modified vitrified bond powder for 360 h was uniform， and the bending strength and Rockwell hardness reached the maximum values of 189.3 MPa and 100.15 HRB， which were 12.72% and 5.82% higher than those of unmodified cBN grinding tools respectively. The surface roughness value of the workpiece was 0.054 μm when using this grinding tool to honed the inner holed of bearing steel， which was 48.6% lower than that of the workpiece processed by unmodified cBN grinding tools.

Abstract:Al was used for diffusion bonding with FeNiCoCrMn high-entropy alloy （HEA） and its six subsets （Ni、NiCo、FeNi、FeNiCo、NiCoCr、FeNiCoCr） by Spark Plasma Sintering （SPS）. The influences of chemical composition on the microstructure， element distribution and phase composition of the interfaces after bonding were investigated in detail， as well as the final microhardness. The results reveal that FeNiCoCrMn HEA had a better diffusion barrier effect on Al than its subsets， and the minimum thickness of the IMC diffusion layer after bonding with Al is only 14.5 μm. The intermetallic compounds （IMCs） generated at the interfaces of Ni， NiCo and FeNi after bonding were mainly Al3Ni-type IMC， while the interfaces of FeNiCo， NiCoCr， FeNiCoCr and FeNiCoCrMn after bonding were mainly Al13Fe4-type IMC， and the greater the proportion of Al13Fe4-type IMC in their respective IMC diffusion layers， the more significant the softening effect of the interface. The interface softening effect of the FeNiCo alloy and Al after bonding was the most obvious， and the lowest interface hardness value was only of 424 HV.

Abstract:Conjugated porous polymer， due to its diversified synthesis methods， high therm-and photo-stability， and tailorable band structure， has demonstrated promising application in visible-light-promoted photocatalysis. Herein， two polymers， i.e. conjugated microporous polymer CMP1 and linear polymer LP2， were prepared by Pd-catalyzed Suzuki-Miyaura cross coupling. The morphology and structure of polymers were characterized by scanning electron microscopy （SEM） and X-ray diffraction （XRD） measurements. Due to its high conjugation degree， CMP1 displays a higher visible light absorption capacity than that of LP2， as the optical property of those two polymers is investigated by the UV-vis diffuse reflectance spectroscopy （DRS）. The conduction band positions of the polymers were analyzed by cyclic voltammetry （CV） test. Both the valence band and conduction band positions of the two polymers met the thermodynamic requirement for photocatalytic NADH regeneration. The photocurrent response and electrochemical impedance were investigated by electrochemical tests. Compared with LP2， CMP1 has a faster photoexcitation response and charge transfer rate. Finally， the photocatalytic NADH regeneration experiments were conducted with CMP1 and LP2 as visible light photocatalysts. As expected， CMP1 demonstrates a higher reaction conversion of 58.3% than that of LP2 （50%）， indicating an excellent photocatalytic efficiency of conjugated polymers.

Abstract:For biomedical magnesium alloys， Mg-Zn-Ca ternary alloys were prepared by Zn and Ca elements microalloying and high extrusion ratio deformation. The microstructure of the materials was characterized by metallographic analysis and scanning electron microscopy. According to the application scenario， the corrosion resistance of the alloy was evaluated by the combination of conventional electrochemical test and micro-electrochemical test in human simulated body fluid （SBF）. At the same time， the sliding wear property of the material was analyzed by friction and wear test. The results show that the alloy microstructure is the single-phase solid solution by microalloying and hot extrusion. The grain size of the alloy is relatively uniform， the corrosion current density is about 119.33 μA/cm2， and the corrosion mechanism is uniform corrosion. Compared with dry friction， the better corrosion resistance of the material itself co-improves the wear resistance of the material in the simulated body fluid environment. At the same time， the wear rate of the material is significantly reduced with the lubrication effect brought by the simulated body fluid. At this time， the wear mechanism is mainly corrosion wear accompanied by slight abrasive wear.

Abstract:Based on the structure and morphology of the armadillo outer skin armor， this paper designed a bionic double-layer composite scale composed of SiC ceramic and ultra-high molecular weight polyethylene （UHMWPE） as well as a new type of flexible protective gear. The flexible protective gear consisted of an upper scale layer and a lower backing layer， in which the scale layer consisted of periodically arranged composite scales， and the cushion layer was made of UHMWPE layers. The flexibility test was carried out using the MTS C43 electronic universal testing machine， and ballistic tests were conducted in accordance with the ballistic protection performance requirements of standard Ⅲ in the "Military bulletproof vest safety technical performance requirements" （GJB 4300A―2012）. Then， the influences of the thickness ratio of the composite scales， ceramic thickness， number of backing layers， and penetration position of the armor-piercing incendiary bullet on the protective performance of the torso’s flexible protective gear ware analyzed，and the change process of the rigidity of the protective gear under concentrated load was discussed. The results show that the new type of flexible protective gear has good protective performance， and different penetration areas have a significant impact on the damage range； the cushion layer limits the bending deformation of the scale layer under concentrated load and reduces the overall flexibility. The research results of this paper can provide valuable references for the design and performance improvement of torso flexible protective gear.

Abstract:Aiming at improving the prediction and judgment of rollover risk in rollover warning， a more efficient and accurate rollover warning algorithm was proposed to provide an important basis for drivers or other driving assistance systems to determine the intervention time of vehicle control. Firstly， a 3-DOF vehicle pre-warning reference model was established. The phase-plane method was selected to divide the roll stability region as the rollover index， an improved time-to-rollover （TTR） algorithm was designed and TTR was calculated according to the response of the 3-DOF vehicle model. The analysis results show that the phase plane rollover index is close to the actual lateral-load transfer rate （LTR）， which is more accurate than the common expression of LTR， and the improved TTR is closer to the actual TTR. Then， to improve the computational efficiency of pre-warning， a long short-term memory （LSTM） model was established to replace the improved TTR algorithm and the TTR value output by the model was used as the basis for vehicle pre-warning control. Finally， the LSTM model was trained by collecting data through driver-in-the-loop （DIL） tests. In two working conditions， the proposed rollover warning method was verified to have the accuracy of rollover risk prediction and higher real-time performance.

Abstract:The design parameter selection of model predictive control significantly affects the performance of the controlled system. The current mainstream parameter tuning methods based on expert experience have the disadvantages of poor controller robustness and high calculation cost. To solve the above problems， this paper proposes a parameter tuning algorithm based on Fuzzy C-means-Extreme Learning Machine-Bare Bones Particle Swarm （FCM-ELM-BBPS）. Firstly， Fuzzy C-means （FCM） clustering is used to preprocess the data， and the complex data of the controlled system is clustered according to its own characteristics， so as to reduce the training error of the neural network and improve the prediction accuracy. Secondly， for each kind of characteristic data， the Extreme Learning Machine （ELM） was used to establish the mapping relationship model between predictive control parameters and performance indices， and the parameter tuning rules were further obtained. Then the Bare Bones Particle Swarm （BBPS） optimization algorithm is used to tune the predictive control parameters. The Gaussian distribution is adopted to update the particle position， which accelerats the convergence of the objective function and effectively reducs the parameter optimization time. Finally， simulation and experiment of the water tank system are carried out respectively to prove the effectiveness of the proposed algorithm. Experimental results show that， compared with existing methods， the proposed algorithm has more advantages， in which the tuning time is reduced by 34.84%， and the time domain performance indices such as the adjustment time are improved by 43.98%.

Abstract:Aiming at the modal aliasing problem of the Adaptive Local Iterative Filtering （ALIF） method， a Pseudo-extrema-based Adaptive Local Iterative Filtering （PEALIF） method is proposed， which uses the method of adding pseudo-extrema to make the distribution of signal extremely more uniform， effectively suppressing the problem of modal aliasing， and also ensuring the order of algorithm decomposition. The principle of the PEALIF method is introduced in detail. Simultaneously， simulation signals are constructed， and this method is analyzed and compared with the Empirical Mode Decomposition （EMD）， Ensemble Empirical Mode Decomposition （EEMD）， Complementary Ensemble Empirical Mode Decomposition （CEEMD） and ALIF methods. The results show that PEALIF has certain advantages in decomposition ability， suppression of modal aliasing and anti-noise interference. Finally， this method is applied to the fault diagnosis of double inner ring bearing. The experimental results show that the PEALIF method can obtain more prominent and easily identifiable fault feature information， which confirms the application of this method in the bearing fault diagnosis analysis.

Abstract:Grinding dynamics and discharge efficiency within semi-autogenous （SAG） or autogenous （AG） mills are predominantly determined by the mill liner design. Due to the mill size， it is difficult to quantitatively investigate the dynamic discharge flow performance with physical experiments. This research aims to utilize a coupled discrete element modelling and smoothed particle hydrodynamics method to model the two-phase mineral slurry within grinding mills. Two distinctive mill liner designs， radial and curved discharge end， were selected and the developed numerical framework was employed to quantitatively compare the discharge efficiency of a selected mineral slurry. A rotary viscometer was initially used to calibrate the numerical modelling parameters to ensure the flow dynamics of the slurry to reflect its actual behaviour. Numerical modelling was conducted and the results indicated that the curved discharge end showed 13.4% throughput increase and 93.3% reduction on back-flow in pulp lifters. Modelling results were validated by site measurements.

Abstract:To solve the problem that the uncertainty of design variables is usually ignored in the process of optimization design in milling， which may lead to the violation of optimization design constraints， the multi-objective robust optimization design of milling parameters is studied based on the approximate model method. In the proposed method， the maximum milling force and the surface roughness after milling are the objective functions， and the milling linear speed and feeds per tooth are the design variables. The polynomial response surface model and radial basis function model are used to replace the implicit relationship between the objective functions and design variables. NSGA-Ⅱ multi-objective genetic algorithm combined with the Monte Carlo simulation method is used to compare and analyze the conventional deterministic optimization for milling parameters and the robust optimization considering the fluctuation of design variables. The results show that the method can effectively reduce the milling force in the milling process and improve the surface quality of the workpiece after machining with high reliability.

Abstract:The operating temperature and battery aging are the key factors that affect the accuracy of estimation of the state of charge （SOC） of lithium-ion batteries. A hybrid model combining Dynamic Time Warping （DTW） and Kinetic Battery Model （KiBaM） is proposed based on the second-order RC equivalent circuit mode. The DTW algorithm is used to calculate the aging state of the battery based on data on the charging voltage， and the KiBaM model can get the unavailable capacity of the battery due to the current effect. The second-order RC equivalent circuit model is combined to derive the new SOC calculation matrix. After that， the unscented Kalman filter algorithm is used to estimate SOC. The accuracy of the hybrid model is verified based on the Urban Dynamometer Driving Schedule operating conditions. The experimental results show that the max error of the model is less than 2% under the low temperature of 10 ℃ environments and 200-cycle aging conditions.