Abstract:

Abstract:Harvesting energy from the transportation environment to monitor the health status of transportation infrastructure is not only convenient and sustainable, but also zero-carbon and environmentally friendly. It is beneficial to the realization of China′s “dual carbon” goal. At present, there are many researches on energy harvesting of transportation environment and health monitoring of transportation infrastructure. However, there are no reports that summarize and refine the self-powered transportation infrastructure health monitoring technologies. This article outlines the energy source and electromechanical conversion mechanisms of transportation environment, elaborates the basic content and research progress of the health state monitoring of traffic infrastructure, and emphasizes that energy harvesting of transportation environment is a potential solution to the problem of power supply for transportation infrastructure health monitoring. The detailed discussions include the energy harvesting technologies in transportation environments and the applications of self-powered transportation infrastructure health monitoring, and the challenges and prospects of self-powered transportation infrastructure health monitoring.

Abstract:In the field of structure health monitoring, compared to body waves, ultrasonic guided waves have advantages of long transmission distance, large coverage, and low inspection cost. This article proposes a synchronous implementation method of data transmission and defect detection based on Lamb waves, which realizes the multi-functional multiplexing of the ultrasound system. The Lamb wave mode tuning theory is evaluated by the theoretical simulation and frequency sweeping experiments. A piezoelectric wafer active sensor is used to excite S0 mode Lamb waves on the aluminum plate at a center frequency of 500 kHz for data transmission and defect detection. To solve the cross-talk problem caused by Lamb wave boundary reflection, a shift-invariant sparse coding method is used to recover the information. The information transmission rate of 100 kbps is successfully achieved on the aluminum plate with reflected boundary, and the bit error rate is 0. Meanwhile, atomic signals in shift-invariant sparse coding are used to detect structural defects. The precise location of defects is realized according to synthetic aperture focusing technology, and the positioning error is less than 0. 2% .

Abstract:To improve the quality and efficiency of coherent plane wave compounding ( CPWC) imaging for two-layered medium, a frequency-domain beamforming algorithm combined with sign coherence factor (SCF) is proposed. The phase symbol of plane wave data is extracted. And the wave field is extracted by the modified frequency-domain beamforming algorithm. The SCF weighting factor is established by using the field extrapolated phase symbol to weight the plane wave image after beamforming. Results show that, after SCF weighting, the average full-width at half maximum of defects in the time-domain and frequency-domain beamforming images are basically identical, which are 75% and 78% of the time-domain DAS respectively. The signal-to-noise ratio of the frequency-domain images is about 5 dB higher than that of the time-domain DAS. On the premise of the same resolution and signal-to-noise ratio, the imaging efficiency of the SCF weighting algorithm for frequency-domain beamforming are more than 4 times compared to the SCF weighting algorithm for time-domain beamforming. Meanwhile, the high image quality and the low computation complexity are both considered. The proposed ultrasound phased array imaging method is suitable for double medium in nondestructive detection.

Abstract:A fast transient component extraction method of the train bearing fault acoustic signal is proposed, which is based on Doppler modulated time-shifting Laplace wavelet. It includes two steps that are rough estimation first and precise identification. The first is rough estimation of transient parameters. The existing periodic Doppler modulated Laplace wavelet model is used to roughly estimate the transient parameters. The second is precise parameter estimation and transient component extraction. A Doppler modulated time-shifting Laplace wavelet model is formulated, which uses one-by-one matching strategy to accurately estimate the transient parameters and extract the transient components. The proposed method has two advantages, which are high accuracy and high efficiency. For high accuracy, the Doppler modulation time-shifted Laplace wavelet model has only one wavelet component for positioning the delay parameter in the time domain, which can solve the matching error problem caused by the pseudo-period of the transient component. For high efficiency, because the periodic transient model is used to roughly estimate the parameters of the transient components, the range of the wavelet parameters can be set very small in the process of extracting the transient components one by one. The experiment comparison and analysis results show that the efficiency is increased by 71. 46% , compared with the direct extraction method. This study provides a method to accurately and efficiently extract transient components from train bearing fault acoustic signals containing Doppler distortion.

Abstract:The existing engineering signal processing methods are based on complete data acquisition, which do not consider the missing signal processing. However, in engineering practice, due to human factors and natural irresistible factors, the sensor may fail and result the lack of signal acquisition. To eliminate the negative influence of signal loss on engineering signal processing, a signal recovery method based on the variational Bayesian parallel factorization is proposed. Firstly, the collected vibration signal is constructed into a three-dimensional tensor by the parallel factor analysis theory. Meanwhile, combined with the Bayesian method, potential variables and super parameters are introduced to formulate Bayesian parallel factor probability graph model. Then, the posterior distribution of the factor matrix and the super parameters are derived by the variational Bayes algorithm. Therefore, the distribution prediction of the missing element can be further deduced. Finally, the proposed algorithm can better solve the problem of signal loss by analyzing the lower bound of the model and the selection of initialization parameters. Two evaluation indexes ( i. e. root mean square error and root relative squared error) are used to evaluate the performance of the algorithm. The simulation and experiment results show that with the increase of missing ratio, the variational Bayesian parallel factorization algorithm has smaller error than the traditional low rank tensor completion algorithm, which can more effectively restore the missing signal. The proposed method provides an effective way to solve the problem of signal missing caused by sensor failure in engineering signal processing.

Abstract:This article proposes a method based on the deep feature fusion of multi-sensor data for accurate motor fault diagnosis under varying speed condition. First, vibration, acoustic, and leakage magnetic signals are sampled from the data acquisition node. The accumulative rotating angle of the motor rotor is calculated from the leakage magnetic signal. Then, the order analysis is conducted on the vibration and acoustic signals based on the angle curve. Finally, the features of the pre-processed signals are extracted and fused by using the double-layer bidirectional long short-term memory (DBiLSTM) networks for fault pattern recognition. Experimental results show that the proposed method can identify 10 types of working conditions including high-resistance connection, eccentric, broken wire of the Hall sensor, interphase short circuit, and bearing faults with the accuracy of 99. 86% , by extracting and fusing of 8 channels of motor vibration and acoustic signals. The method is promising to be deployed into the internet of things edge computing node for remote online condition monitoring and fault diagnosis.

Abstract:The performance degradation process of the airborne fuel pump has of multi-stage and nonlinear characteristics, which requires real-time life prediction. To address these issues, an online degradation model and a life prediction method based on failure physics and data driven are proposed. The fuel pump degradation stage is identified online by the switching Kalman filter, the degradation model of rapid degradation stage is formulated based on failure physics and data-driven method, the model parameters are continuously updated based on the unscented Kalman filter, and the failure life is predicted by using the updated model. The proposed method is compared with the data-driven method, the fusion method without degradation stage identification or parameters update. The root mean square value is less than 0. 3 during the whole parameter update process, and the percentage error of lifetime prediction is less than 2% , which are smaller than the values of the compared method. The effectiveness and superiority of the proposed method are verified.

Abstract:Damage degree assessment is critical for the prognostics and maintenance of rotating equipment. Lempel-Ziv complexity has been widely used for rotating equipment quantitative fault diagnosis. However, traditional Lempel-Ziv complexity indicator extracts fault information only at the single scale, and it is difficult to fully explore fault features. Thus, scholars proposed the multiscale Lempel-Ziv complexity. However, multiscale analysis would shorten the length of time series and lead to inaccurate assessment results easily. Therefore, this paper proposes a damage degree assessment method for rotating equipment based on variable-step multiscale fusion Lempel-Ziv complexity (VSMFLZC). Firstly, the variable step length strategy is adopted to optimize the coarse-grained procedure and explore the fault information more comprehensively. Then, a fusion method based on Laplace score weighting is applied to evaluate the importance of each scale and the method can convert variable-step multiscale complexity sequence into a single but comprehensive evaluation indicator, i. e. the proposed VSMFLZC, which is used to explore the characteristics of the vibration signal comprehensively and achieve the damage assessment of the rotating equipment. The effectiveness of the proposed method is verified with bearing singlepoint defect dataset, bearing life cycle dataset and gearbox fatigue test dataset. Meanwhile, the indicator is compared with other complexity indicators. Results show that the proposed indicator can assess the fault severity of bearings and the wearing degree of gears with 100% accuracy, detect early failures and realize the quantitative diagnosis of rotating equipment.

Abstract:To enhance the low precision of transformer fault diagnosis, a model based on multi-strategy improved sparrow algorithm (MISSA) and bidirectional long short-term memory network (BiLSTM) is proposed. Based on dissolved gas analysis (DGA) technology in oil, the uncoded ratio method is used to extract 9-dimensional fault features of the transformer as the input of the model for network training. The Softmax function is used to obtain fault diagnosis types in the output layer. The sparrow search algorithm ( SSA) is improved by logistic chaos mapping, uniformly distributed dynamic adaptive weights and dynamic Laplacian operator. In the initial solution set, the multi-strategy improved Sparrow algorithm (MISSA) is used to optimize the target hyperparameters. In this way, the transformer fault diagnosis accuracy is optimized, and the kernel principal component analysis (KPCA) is used to reduce the dimension of fault feature indexes, and the convergence speed of the model is accelerated. Compared with PSO-BiLSTM, GWA-BiLSTM and SSABILSTM fault diagnosis models, the diagnostic accuracy of the proposed model is 94% , which is 11. 33% , 8. 67% and 6% higher than those of PSO-BiLSTM, GWA-BiLSTM and SSA-BiLSTM fault diagnosis models, respectively. It is verified that the proposed method can effectively improve the performance of transformer fault diagnosis.

Abstract:The identification method of pipeline leakage aperture based on the deep neural network has a high identification rate. However, its application in industrial environment and real-time processing is greatly limited due to the large number of parameters and large memory consumption due to its complex structure. To address this issue, an optimized convolution improved dense block lightweight neural network is proposed for the pipeline leak aperture identification. Firstly, a new multi-convolutional dense block is constructed by combining the deeply separable convolution with the heterogeneous convolution to extract the features of leakage signals. Then, the convolutional attention mechanism is used to classify the weight of features to realize the importance distinction of features. Finally, the results are obtained by classifier. Experimental results show that the recognition accuracy of the proposed method is 96. 59% , and the number of parameters is only 781 KB. While ensuring high recognition accuracy, the number of parameters and floating point numbers are greatly reduced, the training time is also reduced, and the real-time response ability is improved, which has guiding significance for practical industrial monitoring applications.

Abstract:The fault diagnosis of the planetary gearbox often relies on strong professional knowledge, and the universality of the diagnosis model is poor. Based on deep reinforcement learning, an intelligent fault diagnosis method of the planetary gearbox using the deep wide convolution Q network is proposed. Firstly, fault diagnosis of the planetary gearbox is resolved into a sequential decision problem, which is described by the classification Markov decision process. The fault diagnosis simulation environment is established. Secondly, a deep wide convolutional neural network is designed as an action-value network in the deep Q network model to enhance the perception ability of the environmental state. Finally, the model learns the best diagnostic policy autonomously by interacting with the environment and according to the reward of the environment. In this way, the state identification of the planetary gearbox can be achieved. Experiment and case results show that this method can effectively and accurately realize the intelligent diagnosis of the planetary gearbox under multiple working conditions. The diagnostic accuracy is more than 99% , which enhances the generalization and universality of the diagnosis model.

Abstract:Unsupervised intelligent fault diagnosis under different operating conditions is still a challenge task. To obtain high-quality samples and strong model generalization performance, an unsupervised intelligent diagnosis method based on the symmetrical contrast learning framework is proposed for gearbox fault diagnosis. Firstly, a positive sample set and a negative sample set are constructed and enhanced from original signals by adding noise and sequence inversion, which can be fed into two convolutional neural networks (CNN) with the same structure to extract high-dimensional features. Then, a novel symmetrical contrast learning method is proposed to obtain the underline encoding information by measuring the degree of similarity between positive and negative samples. Further, the noisecontrastive estimation loss function is optimized through symmetrical self-supervised learning strategy. In this way, the label information of the sample itself could be effectively used, and the discriminative performance of extracted features from unlabeled samples is improved. Finally, the proposed method is tested and verified on the gearbox data set. Three indicators including clustering accuracy, classification coefficient and partition entropy are constructed for comprehensive evaluation. Experimental results show that the proposed method achieves at least 98% clustering accuracy, which has stronger cluster and generalization ability than other diagnosis approaches.

Abstract:This article takes the high-speed train axle box bearing as the research object. A bearing fault diagnosis method is proposed to deal with finite variable working conditions, which is based on the supervised auto encoder feature representation transfer. The feature sequences of different working conditions are mapped to the reference condition feature sequences. In this way, the influence of condition change on bearing fault feature is decreased. The migrated features are inputted into the fault diagnosis model based on the convolution neural network, which is pre-trained by the reference condition training feature sets. Then, the axle box bearing fault diagnosis is achieved under variable working conditions. The open bearing data of Case Western Reserve University and the high-speed axle box bearing data are utilized. Experimental results show that the accuracy of fault identification has been greatly improved after feature migration. The method can achieve the feature migration under different working conditions and reduce the distortion of fault features caused by the change of working conditions.

Abstract:Aiming at the issue of high delay and reliability in the transmission of a large number of mechanical vibration wireless sensor networks, an efficient and reliable transmission method of mechanical vibration wireless sensor networks based on redundancy strategy is proposed. Firstly, the time consumption under ACK and Non-ACK transmission mechanism supported by IEEE 802. 15. 4 protocol is analyzed and compared, and the time delay caused by canceling transmitting ACK frame ensures the efficient transmission of a large number of mechanical vibration data; then, a reliable data transmission method based on redundancy strategy is proposed, which uses partial redundancy matrix for data coding to expand the original mechanical vibration data under packet loss rate; finally, the monitoring center decodes the received data to obtain the original data. The experimental results show that after the data size of the proposed method is increased by 25% after redundant coding, the transmission delay of each round can be reduced by about 3 s, the transmission energy consumption can be reduced by about 555 mJ, and the data integrity can be guaranteed.

Abstract:To study the influence of temperature on the measurement of wind tunnel strain-gage balance, the calibration technology under the action of thermo-mechanical coupling is taken as the focus of the research. The thermo-mechanical coupling loading matrix is established through the response surface experimental design method. The thermo-mechanical coupling loading is implemented on a six components wind tunnel strain-gage balance based on the six-degree of freedom calibration system and the constructed temperature environment. The balance formula including temperature, force and moment parameters is established by the multivariate regression method. Finally, the verification load is used to test the accuracy of the balance formula. Results show that the comprehensive loading error of each component of the balance is better than 0. 3% under the action of temperature, force and moment load. The balance formula can accurately characterize the comprehensive performance of the balance under the action of thermo-mechanical coupling. Results show that the calibration technology of wind tunnel strain-gage balance under the action of thermo-mechanical coupling can be used to evaluate and correct the influence of temperature on balance measurement.

Abstract:The eddy current pulsed thermography is a new technique for defect detection with high detection speed, high sensitivity, and large detection range. To adapt the dynamic detection of wheel tread, this article proposes a rectangular electromagnetic induction excitation sensing structure. By deducing the mathematical model of the magnetic circuit of the sensor structure, the feasibility of the sensor structure is proved theoretically. Through numerical simulation, this study analyzes the distribution of electromagnetic and eddy current field in the cracked area of tread surface. And the detection results of the straight coil are compared. Based on this, an automatic wheel flaw detection system is established, which is capable of dynamic measurement of defects at 65 mm/ s speed. The results show that, by optimizing the induction heating uniformity of the wheel tread, the rectangular yoke sensing structure improves detection results for fatigue cracks in the shallow surface of the wheel tread (axial surface opening).

Abstract:A graphene sound generator based on laser scribing is studied in this article. To study the influence of graphene sound generators with different structures on the output performance, four structures of graphene sound generators with solid, coarse mesh, medium mesh and fine mesh are designed. First, the thermoacoustic conversion models of solid structure and mesh structure are formulated, respectively. Then, the fabrication process and SEM characterization of laser scribed graphene are given. Finally, the output performance of graphene sound generators with four structures is measured. Experimental results show that the solid structure has the best output performance. The maximum SPL reaches 40. 68 dB in the conditions of 20 kHz with 5 cm distance when the input power is 0. 78 W. It is found that in the range of 0~ 0. 78 W, for every 0. 1 W increase in input power. The theoretical SPL is increased by an average of 2. 58 dB, and the experimental SPL is increased by 2. 48 dB on average. These results are basically consistent with the theoretical model. The graphene sound generator based on laser scribing has advantages of low cost, simple preparation, good acoustic performance, which can be widely applied in the fields of biology, medicine, and wearable electronic devices.

Abstract:The measurement of seawater temperature in marine environment detection requires large-range and high-resolution. To achieve this objective, a reflective-type optical fiber sensor based on the cascaded no-core fiber ( NCF) and fiber Bragg grating ( FBG) is proposed, which is also evaluated. The manufacturing process of the sensor includes three steps. First, the coated-NCF is fused to the single-mode fiber (SMF) inscribed with FBG. Then, the gold film is plated on the other side of the NCF to form a reflector. Finally, the fiber structure is encapsulated in the capillary. It is worth noting that the NCF without coating is a multi-mode waveguide, while the NCF with coating can be regarded as an anti-resonant reflecting optical waveguide. Theoretical analysis shows that the output spectrum is achieved by the superposition of multi-mode interference and anti-resonance effect. Due to the high thermal optical coefficient of NCF polymer coating, the position of the interference wavelength in the spectrum shifts obviously with the temperature change. According to the position of the FBG center wavelength and the interference wavelength, the accurate temperature value can be calculated through the corresponding fitting curves. Experimental results show that the minimum detectable temperature resolution is 0. 000 1℃ in the range of -6℃ ~ 54℃ . This reflective sensor has advantages of convenient processing, compact structure, and high sensitivity, which has great potential for large-range and high-resolution seawater temperature measurement applications.

Abstract:The transmitting transducer is the core component of the acoustic logging while drilling tool (ALWD). Its working bandwidth and the strength of the radiated acoustic energy are two key technical indicators of the instrument. The transducer is a non-linear capacitive load. The impedance matching can be used to enhance the active power and broaden the working frequency band. This article combines the multi-mode equivalent circuit of the transducer with the impedance matching network. And the circuit grid equation is utilized to obtain the component parameters. Finally, the sound field test is implemented in an anechoic pool, while observing the electrical and acoustic characteristics of the transducer terminal active power, sound field radiation sound pressure, and emission voltage response. The effect of the impedance matching network is evaluated comprehensively. It provides help and guidance for the follow-up ALWD instrument development and experimental testing. Experimental results show that the designed impedance matching network can significantly improve the acoustic radiation performance of the transducer. The frequency bandwidth is expanded by 3. 3 times, the active power is increased by 1. 7 times, the sound pressure is increased by 1. 8 times, and the emission voltage response is increased by 5 dB. The transmit voltage response in the 12~ 16 kHz frequency band is up to 125 dB.

Abstract:In the magnetic interference environment, the yaw angle of the aircraft calculated by the magnetic sensor will deviate. Aiming at the estimation error caused by the magnetic interference, a magnetic sensor yaw angle diagnostic filtering method based on nonlinear extended state observer is proposed for the aircraft system with model uncertainty. Firstly, a two-parameter nonlinear extended state observer is designed to reduce the complexity of parameters. Then, based on the observer, an improved observer residual method is designed to diagnose and filter the magnetic interference data. Finally, an improved complementary filter based on error prediction is designed to further suppress the magnetic interference through fusion filtering. The static simulation results show that the data matching rate of the diagnosis method designed in this paper is more than 94% , and the fusion filtering result is phase ahead and smoother. The dynamic flight experiment results show that the diagnostic filtering method effectively suppresses the influence of the magnetic interference on the yaw angle estimation of the four-rotor aircraft, and enhances the stability and anti-magnetic interference ability of the aircraft.

Abstract:Under dynamic conditions, how to achieve accurate transmission shaft torque measurement data is an important way to evaluate the health status and running state of the transmission shaft. The drive shaft torque measurement is often implemented in harsh environment, such as vibration, high temperature and humidity, and strong electromagnetic. The electrical sensors have problems, such as leakage, electric sparks and difficulty in working in the strong electromagnetic interference environment. To address these issues, this article proposes a torque measurement method based on the extrinsic optical fiber Fabry-Perot interferometry. The working principle of the optical fiber sensor is described, and the installation mode of the sensor is given. The torque test platform of the sensor is established and the performance of the sensor is evaluated. Experimental results show that the sensor sensitivity is (0. 224±0. 06) μm/ Nm, which is in good agreement with the theory. The proposed measurement method has advantages of large dynamic range, high measurement accuracy and real-time speed in the torque measurement of drive shafts.

Abstract:To realize the quantitative characterization of combustion stability based on boiler flame images and overcome the training problem of insufficient unstable combustion samples, a real-time and quantitative characterization method of combustion stability based on the convolutional variational autoencoding model is proposed. First, the model is trained by using the flame images under stable combustion conditions, and the high-dimensional latent probability distribution of the stable combustion image is obtained by using the convolutional variational autoencoder. The distribution characteristics of the latent variables corresponding to the model are recorded, the KL divergence value between the distribution and the standard normal distribution is calculated. The KL divergence is used to realize the quantitative characterization of combustion stability. In the simulation verification, the comparison experiments show that the introduction of variational inference theory can improve the reconstruction quality of the model for the combustion image, and the root mean square error before and after image reconstruction is 0. 005 48. The accuracy and effectiveness of the evaluation method are verified through the experiment of adjusting the coal feeding amount of the coal mill to artificially create the combustion conditions of the burner with different degrees of stability, and the evaluation accuracy rate is as high as 92. 1%. The comparison results with the coal fire inspection and evaluation show that the method has the quantitative judgment function of the coal fire inspection system for flame, and the sensing ability is more sensitive. It can give the warning of combustion instability in 167 s before the burner fires, which has certain engineering application value.

Abstract:The detection of dim small ground targets in remote sensing images has problems of less target information and mixed information. To address these issues, a detection algorithm based on the multi-level feature fusion is proposed in this article, which is named as CC-YOLO. Firstly, the deep convolution neural network is used to extract features of the target image step by step, and the high-level and low-level feature spatial pyramid is obtained. Then, the cross-level channel feature fusion is implemented on the spatial pyramid, and features are aggregated along two spatial directions. The newly added CA is combined to retain the accurate location information of dim small targets. Finally, the end-to-end target detection method is implemented on the dual feature map generated after aggregation. And the detection results are output by combining multi-channel detection information. To solve the problem of lacking image data in algorithm experiment, this article establishes the ground-based dim small target dataset ( GDSTD) of remote sensing image. Experimental results show that the proposed algorithm achieves 42. 3% at AP0. 5 ∶0. 95 and 94. 6% at AP0. 5 , and the detection rate FPS reaches 58. 8 frames / s, which has certain robustness and real-time performance.

Abstract:The end pit is one of the important indexes for defect detection of the cylindrical lithium battery. It is very difficult to detect shallow pits automatically because the shallow pits with small contrast are easily interfered by strong noise such as bright spots and dark spots appearing randomly on metal surface. Therefore, a solution is proposed in this article. Firstly, to obtain a clear shallow pit image under a single light source angle, the six images of pit under different light source angles are collected. Secondly, the temporal averaging and outlier elimination method are used to fuse six images to obtain the datum image, and the spatial filtering method based on sliding window and Nyquist sampling theorem is utilized to weaken the interference noise with strong information intensity. Then, the average deviation is calculated according to the error analysis theory. According to the shape of pits in the gray distribution curve, the peak-tovalley difference and width ratio of concave-convex curve segment are extracted. Finally, the BP neural network is used to formulate a detection model to realize pit detection. The samples collected on site are tested, and the correct detection rate of the algorithm is 100% .

Abstract:Recognition of steel surface defect is essential to promote the improvement of steel production quality. However, the traditional image processing and recognition methods have low accuracy and are easily affected by factors such as light. However, the emerging algorithms based on deep learning have problems such as large amount of model parameters and difficulty in deployment, which cannot be widely used in practical production. In this article, a lightweight partial depth mixture separable network ( PDMSNet) is proposed, which is small model size and floating-point operations (FLOPs) and is easy to deploy on resource-constrained platforms. The test results of the standard strip steel surface defect data set NEU-CLS show that the performance of the proposed model is better than other defect classifiers in strip steel surface defect detection. The recognition accuracy reaches 99. 78% , and the number of parameters is only 0. 17 M and 272 M FLOPs. The average time of an image recognition on a single low-end GeForce MX 250 GPU is 0. 47 ms, which can meet the requirements of real-time detection in industrial field.

Abstract:The moving objects and structural deformation in dynamic environments bring the degradation of autonomous positioning accuracy of lidar. To address this issue, a Dynamic Lego-loam method is proposed in this article. To reduce the error caused by the mismatch of dynamic points to the lidar odometry, a point cloud coarse registration method is firstly proposed, which is based on dynamic point culling before the odometer′s precise calculation. The accuracy of laser odometry is improved. Then, to reduce the error accumulation and mapping ghosting caused by the dynamic environment, the traditional radius-based closed-loop detection method is optimized by the scene matching method. The radius-based rough search is used to quickly locate the local scene in a large range. The regional height difference descriptor is established in a small range to accurately match the most similar historical frames, which realizes an accurate closed-loop detection and improves the mapping accuracy in the dynamic environment. Compared with the Lego-loam algorithm, experimental results show that the Dynamic Lego-loam algorithm improves the autonomous positioning accuracy by 63% in a dynamic environment.

Abstract:To address the time-varying observation noise and particle position distribution on simultaneous localization and mapping (SLAM) accuracy in particle filter SLAM (PF-SLAM) for simultaneous localization and mapping of mobile robots, this article proposes an adaptive PF-SLAM algorithm based on variational Bayes, which adopts a Gaussian mixture model to formulate the time-varying observation noise and iteratively estimates the unknown parameters in the mixture model by using a variational Bayesian method. Meanwhile, the particles are divided into fixed particles and optimized particles according to the particle weights, and the particle positions are adjusted by the topological position distribution relationship between two particles, which handle the time-varying observation noise and optimize the particle position distribution. In this way, the optimized particle set could represent the robot position probability distribution and realize the adaptive observation noise and particle position distribution. Compared with the traditional PFSLAM algorithm, simulation results show that the positioning and map building error of this algorithm is reduced by 76. 45% . Compared with the traditional PF-SLAM algorithm, the actual experiments show that the environmental contour error of this algorithm is reduced by 61. 87% . It effectively improves the state estimation accuracy of mobile robot and provides a new reference for mobile robot real-time positioning and map construction.

Abstract:The smart-phone-based personnel indoor positioning is fragile to the phone attitude. To address this issue, an indoor positioning method integrating WiFi and the wearable inertial navigation module is proposed. The pedestrian dead reckoning (PDR) positioning is achieved by leveraging the wearable inertial navigation module fixed to the chest. And the influence from the smartphone attitude is avoided. WiFi fingerprint positioning is also adopted by using the proposed weighted Bayesian algorithm, which provides the initial position for PDR positioning. Meanwhile, the WiFi positioning are continuously fused with PDR positioning under the framework of the unscented Kalman filter to reduce the cumulative positioning error of pure PDR positioning. Finally, a large number of experiments are implemented in the real indoor environment. Compared with the traditional Bayesian algorithm, experimental results show that the positioning error achieved by the proposed weighted Bayesian WiFi positioning algorithm is reduced by 51. 9% . The proposed positioning method integrating WiFi and the wearable inertial navigation module has better accuracy and stability. Compared with the pure PDR positioning algorithm, the average positioning error is reduced by 65. 2% . Furthermore, compared with implementing the same algorithm on the smart phone, the average positioning errors under three different phone attitudes are reduced by 12. 3% , 39. 3% and 48. 4% , respectively.

Abstract:The traditional ant colony algorithm in large-scale and complex environment has problems of slow global search convergence, too many turns in the path and not smooth enough. To address these issues, an improved ant colony algorithm is proposed in this article. This method speeds up the convergence of the algorithm by dynamically updating the pheromones on different levels of ant paths. By introducing the distance function and the direction function as heuristic factors, the quality of path search is improved. An improved adaptive pseudo-random transition strategy is utilized to avoid the probability of falling into the local optimal solution. Based on the optimal path, the cubic uniform B-spline curve is introduced to improve the smoothness of the path. Compared with the traditional algorithm, the path planning experiments in two different scale environments show that the proposed algorithm reduces the number of turns by 55. 6% and 87. 5% , respectively. The convergence speed is improved by 87. 5% and 100% , which verifies the superiority of the proposed algorithm. Finally, taking QBot2e as the platform, the algorithm is applied to indoor automated guided vehicle (AGV) path planning to further evaluate the practicability of the algorithm.