Abstract:This article proposes a temperature field reconstruction algorithm based on a deep learning framework,providing a new method for reconstructing the two-dimensional temperature field of laser absorption spectra.This paper numerically simulates an integrated absorbance array detected from different projection angles,with a specification of N×64(where N=4,8 and 16),and the algorithm constructed in this paper is used to perform temperature field tomography reconstruction on the test set.At the same time,peak signal-to-noise ratio(PSNR),structural similarity(SSIM),and average error are used to evaluate the quality of recon-struction,and the results of two-dimensional temperature field reconstruction for the three modes are studied.The results show that when N=4,the average error of the reconstructed temperature field is 1.32%,the peak signal-to-noise ratio is 37.08,and the struc-tural similarity is 0.961.The reconstruction error decreases with the increase of N value,and the quality of reconstruction improves with the increase of N value,that is,the reconstruction effect is better when N is greater than 4.As the complexity of the tempera-ture field increases,its reconstruction effect relatively decreases,but the reconstruction effect is still good.The research on this al-gorithm is of great significance for the construction of measurement systems and temperature field reconstruction using deep learning methods in actual combustion environments.

Abstract:This paper proposes a deep deterministic policy gradient algorithm based on Multi-Experience Replay,which aims to solve the problem of time-constrained trajectory optimization of unmanned aerial vehicles.Traditional methods often face the prob-lems of slow optimization speed and poor optimization results when dealing with time-constrained track planning problems.In order to overcome the above shortcomings,this paper innovatively introduces a multi-experience pool playback mechanism to improve the traditional DDPG algorithm.Experimental results show that compared with the traditional optimization methods,the DDPG algo-rithm based on multi-empirical pool playback proposed in this paper has significantly improved the optimization ability and can achieve convergence faster.

Abstract:Active electric field detection is not affected by light attenuation and the magnetic properties of the buried objects themselves,and is less affected by the marine environment.Therefore,it can be used as an effective supplement to address the tech-nical shortcomings of sonar equipment and passive magnetic detection under shallow sea sediment burial conditions.In this paper,the electric field model of seabed landfills is constructed by using the COMSOL finite element modeling software.Receiving measure-ment points are set up around the transmitting device driving a specific measurement line.Simulation calculations are conducted for the disturbed electric fields generated by landfills of different materials,positions and depths.The results of the law analysis are giv-en,and laboratory verification tests are carried out,it provides certain methodological support for the detection of offshore towing electric fields.

Abstract:The use of ski poles in cross-country skiing is primarily based on coaches'experience imparted to athletes,fol-lowed by repetitive practice.Such training methods often rely on subjective judgment and lack effective data support.In order to un-derstand the effectiveness of athletes'exertion on ski poles during skiing,a multi-parameter data monitoring system for ski poles is designed.The embedded design of this system is based on Free RTOS,with the STM32F1 series microcontroller as the core.Key in-formation about the athlete's exertion on ski poles is obtained through sensors such as pressure sensors and inertial measurement units,and is wirelessly transmitted to the mobile end in real-time for convenient acquisition of ski pole data.This design effectively collects key information such as pressure,angle,and acceleration exerted by athletes on ski poles,meeting the requirements of training.Experimental results demonstrate significant advantages of the system in monitoring ski pole states.

Abstract:In order to achieve early diagnosis of breast lesions and improve imaging resolution and accuracy,this paper pro-poses a feature extraction algorithm based on ultrasound transmission imaging.The research mainly includes placing the breast mod-el at the center of a ring array,controlling the ring array to collect ultrasound transmission signals,selecting signal attenuation and signal flight time difference as inversion parameters to construct an ultrasound feature matrix,extracting edges from the ultrasound feature matrix,removing singular values,and then imaging through a filtered backprojection algorithm.Quantitative analysis of the sound velocity,attenuation,and size of the imaging results shows that this method can improve imaging resolution and accuracy.

Abstract:In order to solve the problems of poor accuracy,slow detection speed and single detection background of fall detec-tion algorithm,an improved YOLOv8 algorithm for multiple background environments is proposed.The neck structure of the bench-mark model is replaced by Bic FPN-PAN structure,and the detection accuracy is improved by the reuse of the same layer feature in-formation and the bottom detail information.The addition operation is used for feature fusion to ensure the lightweight of the model.Replacing the C2f module in the backbone of the benchmark model with GS2C module can reduce the false detection rate and the number of parameters in the model.After experiments,the accuracy of the improved model is increased by 5.27%,mAP is in-creased by 2.83%,model computation is reduced by 1%.Compared with the current mainstream YOLO series algorithm,the im-proved algorithm has a better effect.

Abstract:Although the ray tracing method based on the normal gradient sound velocity model can solve the problem of nonlin-ear bending of sound lines,it is easily affected by the outliers of sound velocity and has high computational complexity.In order to solve this challenge,an underwater node localization algorithm based on outlier detection and sound refraction is proposed.Firstly,the outlier detection method based on the Median Deviation of the Local Moving Window(MBW)is used to eliminate the anomalous sound velocity value,and the actual propagation path of the sound wave is fitted by ray tracing method based on the normal gradient sound velocity model.Then,the travel time is measured using a round-trip time(RRT)measurement.Then,based on the normal gradient sound velocity equation and depth information,the equivalent slant distance is calculated,that is,the straight-line dis-tance between the starting and ending points of the sound wave.Finally,the position of the underwater node is determined by solv-ing the equation between the position of the underwater node and the slope distance.The simulation results show that the proposed al-gorithm has certain advantages,the average error of the final positioning is about 1.5 m,the positioning efficiency is about 1.4 times higher than that of the dichotomous method,and about 5 times higher than that of the traversal search method.

Abstract:The whole-angle mode hemispherical resonant gyroscope exhibits impressive dynamic characteristics.However,the presence of nonlinear drift in the standing wave of the gyro is attributed to the circumferential non-uniformity in the quality fac-tor of the harmonic oscillator.To address this issue,a scheme for suppressing the drift of the standing wave based on the virtual Cori-olis force is proposed.Firstly,the mechanism behind the anchor drift in the gyro is deduced in the presence of uneven circumferen-tial quality factors.Secondly,a scheme for suppressing the drift of the standing wave in the hemispherical resonant gyro based on the virtual Coriolis force is proposed.Finally,this method is employed to suppress the standing wave drift in the full-angle mode hemispherical resonant gyro,and its effectiveness is verified through experiments.The experimental results unequivocally demon-strate that implementing the full-angle mode control scheme leads to a substantial reduction in the maximum drift rate of the hemi-spherical resonant gyroscope.Specifically,the drift rate decreases remarkably from 66.59°/h to an impressively low 7.46°/s.This re-markable achievement translates to a staggering decrease of 88.79%when compared to the corresponding period last year.

Abstract:Combat command data serves as the fundamental basis for ensuring the smooth execution of combat command.In the informationized warfare characterized by computer networking,it is imperative to construct a supporting combat command data resource platform,thereby continuously enhancing the stability,real-time performance,and accuracy of combat command.As a key component of engineering support forces,the combat command of the Engineers Corps features complex and variable command relationships,high repeatability of command decisions,difficulties in coordination and control,strong time sensitivity of command activities,and high technical requirements for command foundations.This paper thoroughly analyzes the relevant requirements of combat command data for the Engineers Corps,clarifies the specifications and standards of combat command data,and proposes the basic composition and model construction of the data resource platform.

Abstract:As a kind of electrostatic device,the core of CMUT is variable capacitance,that is,the capacitance value is differ-ent under different voltage,so its output characteristics are directly related to the voltage bias applied.The performance of ultrasonic transducers to emit ultrasonic waves has an important impact on the depth measurement and image quality.Firstly,the internal structure and working principle of CMUT are introduced.By making CMUT devices,the emission performance of CMUT devices in silicone oil under different voltage driving conditions is tested.By electrostatic driving the transducer,applying AC voltage and DC voltage,based on the complex environment where CMUT may be applied,silicone oil is selected to test the CMUT's emission perfor-mance and parameters,including the C-V characteristics of CMUT,the sending voltage response stage and the receiving voltage amplitude.The test results show that CMUT has good transmitting and receiving performance in silicone oil.The combinations of DC bias and AC excitation voltage of CMUT are tested,and the experimental data are recorded and analyzed.Finally,the variation of cmut performance under different driving voltages is obtained.

Abstract:With the widespread use of satellite navigation in aviation,military,traffic management,and other fields,satellite signal acquisition in high dynamic scenarios has become a research hotspot.This paper first introduces the principles of traditional acquisition algorithms and elaborates on the multiple challenges faced by satellite signal acquisition in high dynamic environments.It then reviews the current state of research both domestically and internationally,analyzing the advantages,disadvantages,and ap-plicable scenarios of various approaches.Finally,future development trends in signal acquisition technology for high dynamic satel-lite receivers are discussed.

Abstract:Compared to the real world,users observing virtual environments may experience inconsistent depth perception,which hampers the application of virtual simulation systems in training scenarios.To mitigate or eliminate these discrepancies,this paper proposes a depth perception optimization method based on scale variation.By establishing a depth perception mapping model between real and virtual observation spaces through scale variations induced by field-of-view(FOV)scaling in virtual scenes,the method maps corresponding depth perception relationships to the scaling adjustments of large virtual objects.This enhances depth perception effects for trainees.Additionally,to address the limited visible range of small objects in virtual scenes due to restricted display resolution,their physical dimensions are appropriately enlarged.This expands their visible range without significantly im-pacting depth estimation,thereby improving visibility distances consistent with real-world scenes.Tests conducted on a marine sim-ulator demonstrate that the proposed method effectively enhances users'depth perception experience.

Abstract:In response to the problems encountered during the testing process of MEMS inverted pressure sensors,this paper analyzes the correlation between pressure sensor faults and chip design,chip preparation,and integrated sensor packaging,provid-ing reference for preventive maintenance and failure analysis of pressure sensors.This article selects sensor chips successfully re-searched by the research group from 2021 to 2022,and packages them into integrated sensors.This paper analyzies the problems that occurred during testing and repairs their faults.The results show that the metal point electrodes in the glass through-hole of the inverted pressure sensor are warped after long-term operation,forming a fault with the surrounding sputtered metal and causing a circuit breaker.The through-hole is repaired by filling it with conductive slurry,and the repaired sensor's lifespan is greatly extend-ed under working conditions.This article proposes a complete failure analysis process for an inverted pressure sensor,and effective-ly analyzes the fault distribution and reliability of the sensor,providing a reference for developing targeted preventive maintenance plans.

Abstract:Aiming at the problem of mismatch between radar range assessment and actual implement,this study proposes a ra-dar range assessment method based on SNR.Within the theoretical framework of the radar equation,a systematic analysis model in-corporating key parameters such as traget RCS and range attenuation is established.The analytical relationship between SNR and ra-dar range is derived,enabling the construction of a radar range assessment applicable to multiple scenarios.The feasibility of the method is validated through comparion with simulation data.

Abstract:To meet the knowledge-based QA(Question Answering)needs of researchers and engineers in the field of auxilia-ry decision-making for modern management of military camps,this paper develops a FAQ(Frequently Asked Questions)system oriented to military camp management and decision-making.Aiming at the problems of low accuracy in matching similar questions in traditional FAQ systems and insufficient similar QA corpora in specific domains,this paper develops a similar question matching algorithm based on pre-trained models,and adopts a paraphrase generation model to perform data augmentation on similar question corpora.By constructing a FAQ dataset for military camp management and decision-making,this paper validates the model's perfor-mance on this dataset,and analyzes the results based on the Spearman correlation coefficient.The results show that the model can accurately match similar questions.

Abstract:Aiming at the anomaly detection of ship track data,combining the advantages of neural network weight calculation and Boosting the idea of voting model,a multi-model weighted anomaly detection algorithm is proposed,which selects different types of unsupervised anomaly detection algorithms as the base model and trains feedforward neural network parameters with labeled data sets in different fields.The weight of each model is optimized by AUC score.By comparing the accuracy rate and processing speed of each model,the feasibility and correctness of multi-model weighted anomaly detection algorithm in ship track data analysis are verified.

Abstract:The explosive growth of computeintensive scientific workloads—such as AI training and bigdata analytics—poses triple challenges to cloud schedulers:highly dynamic task arrivals,pronounced resource heterogeneity and conflicting optimization objectives.To overcome the limited responsiveness and objective entanglement of existing approaches,this paper presents PRHS-MDP,a predictive rolling horizon scheduling algorithm formulated as a Markov Decision Process(MDP).An LSTM network is employed to forecast multistep node loads and augment the system state with foresight.A rollingwindow optimizer then maximizes a composite reward that jointly considers completion rate,mean response time,and energy consumption.Extensive simulations on a heterogeneous CloudSim testbed demonstrate that PRHS-MDP outperforms FIFO,MinMin,HEFT and PSO:task completion rate increases by 18%,mean response time drops by 23%,and total energy consumption is reduced by 11%,while preserving fast con-vergence and high robustness.The proposed method offers a practical and theoretically sound path towards intelligent,selfadaptive scheduling for scientific cloud platforms.

Abstract:To study the optimization problem of aircraft type allocation in airlines under the dual carbon target,firstly this arti-cle combines two carbon emission calculation methods,ICAO and EEA,and proposes a carbon emission intensity calculation meth-od for aircraft types,which can evaluate the carbon intensity of different aircraft types in different route distance segments.On this basis,a configuration optimization model for airline aircraft models considering carbon emission costs is established,taking into ac-count various practical constraints.The genetic particle swarm optimization algorithm is chosen as the solution algorithm.Finally,multiple experimental plans are set up and validated by using the flight schedule of a large domestic airline.The results show that all optimized solutions can effectively reduce carbon emissions while slightly increasing profits.In the carbon emission prices taken in the experiment,the carbon dioxide emissions reduction continuously increases with the increase of carbon emission prices,with a maximum reduction of about 12.40%.Further experiments have shown that replacing the new generation of models can further re-duce carbon emissions by approximately 11.29%.

Abstract:In recent years,supported by big data,cloud computing,deep learning,brain control,imitation and other technol-ogies,a new wave of evolution has been ushered and an intelligent military revolution is being opened.Starting with the fields of land,sea,air,virtual space and cognition,this paper analyzes and studies the application mode of intelligent combat force in differ-ent battlefields in the future,which is of great practical significance for strengthening the joint command capability of man-machine integration.

Abstract:Currently,terahertz has attracted much attention in the field of biomolecular detection.This study proposes a novel terahertz sensing structure using a combination of coplanar waveguide(CPW)and electric LC(ELC)resonator design.The sensing structure consists of an ELC resonator and a CPW structure deposited on the back of the substrate.By optimizing the structural pa-rameters,sensing can be achieved in the 200 GHz～360 GHz frequency band.The principle of the sensing structure is based on the frequency shift of the resonant peak in the transmission coefficient(S21).On the basis of this structure,a microporous structure is added to the capacitance gap of the ELC resonator.When particles pass through the pores,it causes a change in the resonant fre-quency of the resonator.This sensing structure can achieve miniaturized trace sample detection function and can be used for screen-ing particles and biological cells.