For signal processing related to localization technologies, non line of sight (NLOS) multipaths have a significant impact on the localization error level. This study proposes a localization correction method based on convolution neural network (CNN), which extracts obstacle features from maps to predict the localization errors caused by NLOS effects. A novel compensation scheme is developed and structured around the localization error in terms of distance and azimuth angle predicted by the CNN. Four prediction tasks are executed over different building distributions within the maps for typical urban scenario, resulting in CNN models with high prediction accuracy. Finally, a thorough comparison of the accuracy performance between the time difference of arrival (TDOA) localization algorithm and the results after the error compensation reveals that, generally, the CNN prediction approach demonstrates great localization error correction performance, improving TDOA accuracy by 75%. It can be observed that the powerful feature extraction capability of CNN can be exploited by processing surrounding maps to predict the localization error distribution, showing great potential for further enhancement of TDOA performance under challenging scenarios with rich multipath propagation.
A series of fast protection methods for multi-terminal DC (MTDC) grid have been proposed. However, most of these methods simply use voltage change rate or current change rate, but do not make full use of transient voltage and current information to identify fault regions. This paper proposed a protection method based on the cosine similarity of the transient voltage and transient current. Firstly, the characteristics of transient voltage and current in the case of short-circuit fault are analyzed. Then, the cosine similarity of the transient voltage and the current is introduced, and a protection method is constructed based on the cosine similarity difference of the internal faults and external faults. Finally, a simulation model is built in PSCAD, and the validity of proposed protection method is verified.
In a number of cutting-edge application fields, including image identification, speech and natural language processing, and autonomous driving, machine learning has had considerable success recently. In this study, we use the western black-crowned gibbon song data as the research object and the Kubernetes-based OpenPAI machine learning platform to categorize five environmental sounds in the western black-crowned gibbon habitat in the Mourning Mountains with an accuracy of 97.24%.
. The objective of this paper is to identify loads on suspenders in bridges. The load value at each time was selected as the parameter of identification, and the objective function was built based on measured and relevant simulated data. The differential equation of suspenders vibration was employed, and the relationship between load and acceleration response was given by superposition method. In addition, regularization technology was used in calculations to alleviate ill-posedness of the problem. Taking one tied arch bridge as an example, the load on one suspender of the bridge was obtained. The identification results indicated that the simulated responses were basically consistent with the measured data, which showed effectiveness of the method.
Visual-Inertial Odometry (VIO) is a new technology to determine position and attitude of the object which it is mounted on. Because of its small size and powerful performance, VIO system is widely used in virtual and augmented reality, autonomous driving and robot navigation. Usually, we assume the timestamps of the Inertial Measurement Unit (IMU) and the camera is synchronized, and there is no temporal misalignment (time offset) in VIO system. However, in practice, the different triggering time and propagation delays cause the sensors to be temporal misaligned. If we calibrate the temporal misalignment, the performance of VIO system can be greatly improved. To this end, we propose a novel method to calibrate the time delay between visual and inertial devices. This method accurately estimates the IMU state, time offset and camera pose in the VIO system by using Adaptive Kalman Filter. Simulation experiments show that our method can quickly and accurately estimate the time delay. The results compared against other methods proves that the proposed approach can significantly boost the performance of VIO system.
The new energy vehicle battery management system test platform built by hardware in the loop technology can verify the control strategy of the new energy vehicle battery management system, which is of great significance for reducing the test cost of the bench and the real vehicle and improving the development efficiency. In this paper, a hardware in the loop simulation target machine control model is designed to simulate the voltage, current, temperature and other parameters of the battery pack. Analyze and process the data collected by the battery management system based on the hardware in the loop, verify the control strategy of the battery management system, and verify that it can meet the simulation of most working conditions through testing, and further verify the correctness of the battery management system controller, and reduce the development cost.
Aiming at the problem of multi-model forecast of situation, a situation combination forecast model based on information entropy is proposed. Firstly, the forecast sub-model is trained to obtain the precision error sequence of each sub-model. Then the forecast stability of the sub-model is evaluated based on the information entropy. Finally, the combination forecast model is constructed to complete the situation combination forecast. The MATLAB simulation results show that this method can automatically adjust based on the change characteristics of the situation curve, and is superior to the existing typical methods in stability and forecast accuracy.
In order to improve the surface quality and production efficiency of NC milling parts, the feed speed, cutting depth and feed rate per revolution are selected as optimization parameters, and the surface roughness and material removal rate are used as evaluation indicators. The response surface method is used to design milling experiments. The influence of process parameters on surface roughness is analyzed through contour map, and a multi-objective optimization model of NC milling process parameters is established. By comparing the calculation accuracy of response surface regression equation with the fitting accuracy of BP neural network, it is concluded that the fitting accuracy of BP neural network is high, and then the FBGA optimization algorithm combining BP and GA is proposed to apply to the optimization of NC machining process parameters. The process parameter scheme optimized by the algorithm is verified by secondary experimental processing, and the workpiece surface processing quality and processing efficiency are effectively improved compared with the original workpiece.
This paper studied the surface defect detection of shaft parts. First, CCD industrial camera was used to collect the image information of the tested shaft parts, then the collected image was denoised by median filtering, and the defects of the image were segmented and extracted by the gray histogram bimodal method of threshold segmentation. Finally, for problems with noise which happened after binarization of the image, methods of removing the small connected area after the binary image adopted to suppress the noise and extract the defects completely. The image was processed through openmv. Compared with the results of manual inspection, the surface defects of the parts were detected based on machine vision technology, with fewer false and missed inspections, higher detection accuracy. It was fully improved productivity and ensured product quality.
Recently, increasing studies turned the spotlight on machine learning-based automated citation screening for systematic review. Moreover, outcomes of the model were suggested to be highly influenced by the selection of bibliographic information. In the current study, we implemented automated citation screening algorithm based on literature similarity, and further evaluated the effects of single or combined feature engineering constituted by the title, abstract and publication type. In the single feature-based strategy, we found that abstract-based model showed the highest Work saved over sampling at 95% recall (WSS95) score (73.55 % ). In terms of multiple features-based approaches, the model with features pooling title, abstract, and publication type showed the highest WSS95 score (77.34%). The current findings suggest that literature information might be mainly contained in abstract and title, thus highlighting their crucial roles in systematic review citation screening.
Aiming at the problem of abnormal data mining in marine Argo buoy monitoring data, due to the poor performance of traditional machine learning algorithms in anomaly detection, a two-stage anomaly detection method based on variational autoencoder and K-means clustering is proposed. The anomaly detection of Argo data is realized by combining deep learning and machine learning. The proposed method is divided into two parts, data preprocessing and anomaly detection. Specifically, standardization and normalization techniques are used to preprocess the input data. Then, the data are input into the training process of encoding and decoding in the variational autoencoder. Finally, the reconstruction error between the original data and the reconstructed data is calculated, and the upper and lower limits of the appropriate threshold are set. If the threshold exceeds the upper limit, it is directly determined as the abnormal data, and if the threshold is lower than the lower limit, it is determined as the normal data. It is called the boundary sample between the thresholds, and the extra clustering enhancement is used to determine the anomaly. The experimental verification is carried out on the simulation data set and the real data set, respectively. The experimental results show that compared with other methods, this method significantly improves the accuracy of anomaly detection.
With the rapid development of the Internet, the development of live e-commerce as an important innovation in the form of marketing has also led to the further development of agricultural products e-commerce, which has changed from the original simple display of images to the current real-time video display and multi-faceted interaction. This progress not only strengthens the consumer experience, but also has an impact on consumers' willingness to purchase. Based on the Technology Acceptance Model (TAM), structural equation modeling (SEM) is applied to calculate the impact of the characteristics of live agricultural products on consumers' purchase intention. A model based on the Meta-SEM approach to influence consumer purchase intention is proposed, and AMOS 26.0 is used to verify that perceived trust and perceived risk mediate between economy, informativeness and inter activity, and consumer purchase intention. By analyzing the impact of live streaming characteristics on consumer purchase intention, an optimization method is proposed for agricultural live streamers in live streaming operations, which is relevant for live stream merchants to build a good image, increase loyalty with consumers and gain more revenue.
With the rapid development of science and technology, the cycle of upgrading CNC machine tools is getting shorter and shorter, but the cost is still expensive. For enterprises that rely on CNC machine tools to process, production efficiency is on the one hand, and extending the working time of machine tools is also the key to create value. Therefore, fault detection and diagnosis are imminent. This paper summarizes several representative methods according to all the solutions of fault diagnosis. Since the birth of CNC machine tools, various novel fault detection methods, such as vibration diagnosis, nondestructive testing, expert system, fault tree, and deep learning, have gone from artificial to intelligent, from the original "look, hear, and touch" to advanced neural networks, whether mechanical detection methods, Or deep learning is a method to solve the problem, and the most appropriate method should be selected according to the actual situation.
In forest environment monitoring, data collection from wireless sensor networks may be separated from each other due to different geographical factors, which poses great difficulties for wireless sensor data collection. Therefore, this paper proposes a multi-target unmanned aerial vehicle (UAV) path planning method with time window to solve the path planning problem of data collection using UAV-assisted wireless sensor networks (WSN). In this paper, considering the energy consumption of the UAV and the time window of data collection, a multi-objective UAV path planning model is constructed. For this problem model, this paper, based on improved particle swarm optimization algorithm and genetic algorithm, proposes a hybrid algorithm (IPSO-GA) to plan the path of UAV data acquisition reasonably. Therefore, simulation experiments are conducted in order to evaluate the performance of this algorithm. The experimental results show that the hybrid algorithm (IPSO-GA) based on improved particle swarm optimization algorithm and genetic algorithm, which has been proposed in this paper can successfully generate an effective path, which covers all data nodes with the least number of UAVs. Besides, the IPSO-GA is superior to particle swarm optimization (PSO), improved particle swarm optimization (IPSO) and genetic algorithm (GA) in the global optimal value, the number of UAVs and the flying distance of UAVs.
In the work, we exploit the preconditioned conjugate orthogonal conjugate gradient (PCOCG) algorithm to solve the complex symmetric linear systems of equations arising from 2D spatial complex fractional Ginzburg-Landau equation (GLE). The matrix in the linear systems of equations is equal to a symmetric block Toeplitz matrix plus a complex diagonal matrix. Moreover, the effective $\tau$ preconditioner is proposed to accelerate the rate of convergence of the iterative algorithm. Two numerical experiments are carried out to indicate the computational merit of the proposed method.
This paper introduces a design of the control system of a piston prover micro liquid flow standard device. The core software of the system is designed based on Microsoft Visual Basic 6. The lower computer uses the Delta PLC to complete the control of the equipments and data acquisition of instruments. The control system adopts user-defined RS232 serial protocol as the main communication mode between each equipment and instrument. After the control system is designed, the standard device is used for the real flow test of flowmeter calibration, and the result shows that the system can successfully complete the equipments control, signal data acquisition and report output during the calibration process. So, with the characteristics of simple design structure, strong operability and maintainability of the system, it can be a good reference in the subsequent control system design of this kind of micro liquid flow standard device.
In order to explore the feasibility of in-situ detection of Chlorpyrifos in water by near-infrared spectroscopy, NIRQuest512 near-infrared spectrometer produced by American ocean company was used to build a near-infrared spectrum acquisition system to obtain spectrum data of Chlorpyrifos samples in different concentration ranges. Based on AdaBoost algorithm and PLS partial least square algorithm, the whole band quantitative model of Chlorpyrifos sample spectrum data was established respectively. The results showed that AdaBoost algorithm had better prediction ability than PLS. In addition, the characteristic band of Chlorpyrifos sample spectrum was selected by calculating the correlation coefficient through correlation analysis, and the characteristic band of the experimental spectrum with concentration range of 1–100 ug/mL was 898-1358nm, 1448-1626nm, 1661-1703nm. The quantitative model of the sample characteristic band spectrum data was established by combining the algorithm. Compared with the full band model, the calibration set determination coefficients of the experimental spectral characteristic band model with concentration range of 1–100 ug/mL were increased to 0.999 and 0.963, respectively. The decision coefficients of prediction set were increased to 0.944 and 0.748, respectively. The RMS error of the correction set was reduced to 0.657 and 6.521 respectively. The RMS error of the prediction set was reduced to 6.81 and 14.212 respectively, and the RPD value increased to 4.24 and 2.131. Compared with the prediction ability of the quantitative model of Chlorpyrifos based on the characteristic band and PLS method, the quantitative model of Chlorpyrifos based on the characteristic band and Adaboost method had a high precision prediction ability, which can meet the quantitative analysis conditions. Therefore, the rapid detection of chlorpyrifos pesticide by this method was feasible. The results of this study was applicable to the application of near infrared spectroscopy The detection technology provided theoretical basis and practical application value for the rapid in-situ detection of organophosphorus pesticide and other non-point source pollutants.
Aiming at the shortcomings of traditional MAF-PLL, an adaptive phase-locked loop based on moving average filtering frequency is proposed. The improved MAF-PLL adds Phase-Lead-Compensator, the control parameters are obtained based on the small-signal model. Compared with the traditional MAF-PLL, the MAF-PLL with Phase-Lead-Compensator has faster response speed and the ability of frequency adaptation. It can quickly and accurately extract the frequency and phase angle information of the fundamental positive sequence voltage under various disturbances. Finally, three operating conditions are simulated in Matlab/Simulink to verify the performance of the improved phase-locked-loop when the grid voltage signal changes. The simulation results show the effectives of the improved phase-locked-loop.
Range ambiguity has always been an inherent problem in pulse instrumentation radar system. At present, the traditional resolving range ambiguity algorithm is generally based on the Remainder Theorem. However, when radar detects medium or high orbit space target, the signal-to-noise ratio of the radar echo is low because the space target is far away and its radar cross section is small. In this case, the efficiency of the traditional resolving range ambiguity algorithm is reduced and even the algorithm cannot be applied. Therefore, in order to solve the problem of range ambiguity in the condition of low SNR radar echo, a coherent accumulation resolving range ambiguity algorithm based on chirp rate polarity modulation with M-sequence is proposed. In this algorithm, the M-sequence is used to modulate the chirp rate polarity of radar transmitting signal (chirp signal). Then the radar echo signal is compressed by reference matching filtering. Finally, the sliding window coherent accumulation process is applied to resolving the radial range ambiguity of space target. The simulation experiment shows that the proposed algorithm can improve the resolving range ambiguity efficiently, which is especially suit for engineering application.
This paper describes an intelligent medicine delivery trolley for epidemic prevention using the STM32 microcontroller. The system adopts power control unit with I / O interface. It realizes the forward and turning operation of the vehicle through PWM duty cycle modulation and the speed difference of reduction motor; The obstacle avoidance of the car is successfully completed by using the hc-sr04 ultrasonic ranging system in the front of the car. In addition, the hc-05 Bluetooth module is used to realize remote control. Finally, this design realizes the real human-machine interaction via Bluetooth and mobile APP.
