Dynamic SLAM Visual Odometry Based on Instance Segmentation:A Comprehensive Review

Dynamic Simultaneous Localization and Mapping(SLAM)in visual scenes is currently a major research area in fields such as robot navigation and autonomous driving.However,in the face of complex real-world envi-ronments,current dynamic SLAM systems struggle to achieve precise localization and map construction.With the advancement of deep learning,there has been increasing interest in the development of deep learning-based dynamic SLAM visual odometry in recent years,and more researchers are turning to deep learning techniques to address the challenges of dynamic SLAM.Compared to dynamic SLAM systems based on deep learning methods such as object detection and semantic segmentation,dynamic SLAM systems based on instance segmentation can not only detect dynamic objects in the scene but also distinguish different instances of the same type of object,thereby reducing the impact of dynamic objects on the SLAM system’s positioning.This article not only introduces traditional dynamic SLAM systems based on mathematical models but also provides a comprehensive analysis of existing instance segmentation algorithms and dynamic SLAM systems based on instance segmentation,comparing and summarizing their advantages and disadvantages.Through comparisons on datasets,it is found that instance segmentation-based methods have significant advantages in accuracy and robustness in dynamic environments.However,the real-time performance of instance segmentation algorithms hinders the widespread application of dynamic SLAM systems.In recent years,the rapid development of single-stage instance segmentationmethods has brought hope for the widespread application of dynamic SLAM systems based on instance segmentation.Finally,possible future research directions and improvementmeasures are discussed for reference by relevant professionals.

A Big Data Based Dynamic Weight Approach for RFM Segmentation

Using the RFM(Recency,Frequency,Monetary value)model can provide valuable insights about customer clusterswhich is the core of customer relationship management.Due to accurate customer segment coming from dynamic weighted applications,in-depth targeted marketing may also use type of dynamic weight of R,F and M as factors.In this paper,we present our dynamic weighted RFM approach which is intended to improve the performance of customer segmentation by using the factors and variations to attain dynamic weights.Our dynamic weight approach is a kind of Custom method in essential which roots in the understanding of the data set.Firstly,Analytic Hierarchy Process is used to calculate the subjective weight,then the entropy method is applied to calculate the objective weight.Finally,we use comprehensive integration weighting method to combine the subjective and objective weight to obtain the final weight of the index to calculate the individual user value and quantify the user value difference.The experiment shows that the dynamic weight we used in RFM model is vital,affects the customer segmentation performance positively.Also,this study indicates that customer segments containing dynamic weighted RFM scores bring about stronger and more accurate association rules for the understanding of customer behavior.At last,we discuss the limitations of RFM analysis.

Nonlinear analysis on dynamic buckling of eccentrically stiffened functionally graded material toroidal shell segment surrounded by elastic foundations in thermal environment and under time-dependent torsional loads

The nonlinear analysis with an analytical approach on dynamic torsional buckling of stiffened functionally graded thin toroidal shell segments is investigated. The shell is reinforced by inside stiffeners and surrounded by elastic foundations in a thermal environment and under a time-dependent torsional load. The governing equations are derived based on the Donnell shell theory with the yon Karman geometrical nonlinearity, the Stein and McElman assumption, the smeared stiffeners technique, and the Galerkin method. A deflection function with three terms is chosen. The thermal parameters of the uniform temperature rise and nonlinear temperature conduction law are found in an explicit form. A closed-form expression for determining the static critical torsional load is obtained. A critical dynamic torsional load is found by the fourth-order Runge-Kutta method and the Budiansky-Roth criterion. The effects of stiffeners, foundations, material, and dimensional parameters on dynamic responses of shells are considered.

RO-SLAM: A Robust SLAM for Unmanned Aerial Vehicles in a Dynamic Environment

When applied to Unmanned Aerial Vehicles(UAVs),existing Simultaneous Localization and Mapping(SLAM)algorithms are constrained by several factors,notably the interference of dynamic outdoor objects,the limited computing performance of UAVs,and the holes caused by dynamic objects removal in the map.We proposed a new SLAM system for UAVs in dynamic environments to solve these problems based on ORB-SLAM2.We have improved the Pyramid Scene Parsing Network(PSPNet)using Depthwise Separable Convolution to reduce the model parameters.We also incorporated an auxiliary loss function to supervise the hidden layer to enhance accuracy.Then we used the improved PSPNet to detect whether there is a movable object in the scene.If there is a movable object,its feature points will be removed in the tracking thread,and the removed feature points will not participate in the pose estimation of the camera.In addition,we proposed a filling method based on Generative Adversarial Networks(GANs)for the holes caused by dynamic object removal in the map,which employs a new auxiliary descriptor to assist GANs in restoring static scenes based on semantic information.The proposed system is evaluated on the TUM dataset,and the results indicate that the proposed method performs better than DynaSLAM and DS-SLAM on the TUM dataset.We experimented on the Cityscapes dataset,the improved PSPNet achieving an Intersection Over Union(IOU)of 0.812.

Modeling and analysis of passive dynamic bipedal walking with segmented feet and compliant joints

Passive dynamic walking has been developed as a possible explanation for the efficiency of the human gait. This paper presents a passive dynamic walking model with segmented feet, which makes the bipedal walking gait more close to natural human-like gait. The proposed model extends the simplest walking model with the addition of fiat feet and torsional spring based compliance on ankle joints and toe joints, to achieve stable walking on a slope driven by gravity. The push-off phase includes foot rotations around the toe joint and around the toe tip, which shows a great resemblance to human normal walking. This paper investigates the effects of the segmented foot structure on bipedal walking in simulations. The model achieves satisfactory walking results on even or uneven slopes.

基于改进DDS的应答器动态检测系统2FSK调制器设计

基于高速综合检测列车平台的应答器动态检测系统可实现对应答器相关设备性能的实时检测。为测试车载应答器动态检测系统的性能,需产生高精度可控波特率和载频的调制信号(2FSK)作为上行链路信号。基于直接数字频率合成器(DDS)技术,通过对相位累加模块和相幅转换模块进行设计,提出分段相位累加算法,计算同周期内相位增量和码元跳变处的相位增量;通过改进的坐标旋转数字计算方法(CORDIC)实时计算对应的信号幅值,用数字模拟转换器(DAC)将离散信号转换为模拟信号,从而实现频偏和数据速率可变且相位连续的信号调制。在相幅转换模块设计中,采用改进CORDIC算法替代传统的查表法,将相位增量转换成信号幅度,有效减少了现场可编程逻辑门阵列(FPGA)存储资源的开销。基于改进DDS的2FSK调制器设计方法,可为应答器动态检测系统提供高精度可控波特率和载频的2FSK调制信号,保证了应答器动态检测系统校准的准确性。

ANALYSIS OF DYNAMICAL BUCKLING AND POST BUCKLING FOR BEAMS BY FINITE SEGMENT METHOD

Based on the multi-rigid body discretization model, namely, finite segment model,a chain multi-rigid-body-hinge-spring system model of a beam was presented, then a nonlinear parametrically exacted vibration equation of multi-degrees of freedom system was established using the coordination transformation method, and its resonance fields were derived by the restriction parameter method, that is, the dynamical buckling analysis of the beam. Because the deformation of a beam is not restricted by the discrete model and dynamic equation, the post buckling analysis can be done in above math model. The numerical solutions of a few examples were obtained by direct integrated method, which shows that the mechanical and math model gotten is correct.

RELAXATION IN DOMAINS:STRUCTURE AND SEGMENTAL DYNAMICS OF LC/I DIBLOCK COPOLYMERS

A series of block copolymers consisting of an isotropic (polystyrene) block and a side-chain liquid crystallineblock (LC) have been studied using small-angle X-ray scattering and dielectric spectroscopy. The triblock copolymer (PS-LC-PS) displays an order-to-order transition (OOT) together with the isotropic/nematic transition of the LC phase. The seriesof diblock copolymers show no clear OOT but the phase diagram differs from that of non-LC block copolymers. Thesegmental dynamics as measured with dielectric spectroscopy is dominated by the α and δ relaxation of the LC block. Bothdisplay a WLF like temperature dependence. The relaxation times are influenced by the constraints of the nanoscale domains.They are decreased for the LC confined in the domain as compared to the LC in the continuous matrix.

Characteristic Analysis of DS18B20 Temperature Sensor in the High-voltage Transmission Lines’ Dynamic Capacity Increase

Dynamic capacity increase in high voltage electric power transmission line is currently the most economical method for solving electric power transmission bottleneck nowadays. DS18B20 temperature sensor is applied to the dynamic capacity increase of high voltage transmission lines to measure the conductor temperature and ambient temperature. The paper is focused on the experiment of DS18B20 both in the laboratory and outside. From the result of the lab temperature measurement data analysis, using 4 DS18B20’s is the most suitable plan, considering both accuracy and economical efficiency. In the experiment outside, we get four groups of conductor (uncharged) temperature and four groups of ambient temperature. The data proved that DS18B20 has good stability, and small measurement error. It is suitable for measuring the temperature of conductor and ambient in dynamic capacity increase, and helpful to improve the accuracy of the calculation of capacity increasing.

Dynamic Analysis of Wind Power Turbine's Tower under the Combined Action of Winds and Waves

To deal with the dynamic response problem of offshore wind power tower under the combined action of winds and waves, finite element method is used to analyze the structure and flow field around the outside flange of the segmentation part. The changes of pressure distribution and vorticity about the outside flange are obtained Focused on the analysis on the change of hydrostatic pressure and temperature of the tower cut surface, contour lines under the combined action of winds and waves are depicted. Results show that the surface of the offshore wind turbine tower presents instable temperature field when it suffers the action of winds and waves loads, the static pressure increases nonlinear!y with the increase of altitude, the fluid vorticity around the outside flange follows an parabolic curve approximately. Results provide a reference for the actual monitoring data of the offshore wind turbine tower under the combined action of winds and waves, so as to ensure the normal operation of tower.

Discrete-Time Dynamic Image Segmentation Using Oscillators with Adaptive Coupling

In this study, we propose a novel discrete-time coupled model to generate oscillatory responses via periodic points with a high periodic order. Our coupled system comprises one-dimensional oscillators based on the Rulkov map and a single globally coupled oscillator. Because the waveform of a one-dimensional oscillator has sharply defined peaks, the coupled system can be applied to dynamic image segmentation. Our proposed system iteratively transforms the coupling of each oscillator based on an input value that corresponds to the pixel value of an input image. This approach enables our system to segment image regions in which pixel values gradually change with respect to a connected region. We conducted a bifurcation analysis of a single oscillator and a three-coupled model. Through simulations, we demonstrated that our system works well for gray-level images with three isolated image regions.

Dynamic Analysis of Towed and Variable Length Cable Systems

Towed cable systems are frequently used in marine measurements where the length of the towed cable varies during launch and recovery. In this paper a novel method for modeling variable length cable systems is introduced based on the finite segment formulation. The variable length of the towed cable is described by changing the length of the segment near the towing point and by increasing or decreasing the number of the discrete segments of the cable. In this way, the elastic effects of the cable can be easily handled since geometry and material properties of each segment are kept constant. Experimental results show that the dynamic behavior of the towed cable is consistent between the model and the physical cable. Results show that the model provides numerical efficiency and simulation accuracy for the variable length towed system.

Effective Virtual Reality Based Building Navigation Using Dynamic Loading and Path Optimization

Although computer capabilities have been improved significantly, a large-scale virtual reality （VR） system demands much more in terms of memory and computation than the current computer systems can offer. This paper discusses two important issues related to VR performance and applications in building navigation. These are dynamic loading of models based on cell segmentation for the optimal VR operation, and the route optimization based on path planning for easy navigation. The VR model of engineering and information technology complex （EITC） building at the University of Manitoba is built as an example to show the feasibility of the proposed methods. The reality, enhanced by three-dimensional （3D） real-time interactivity and visualization, leads navigators into a state of the virtual building immersion.

Parallel-Dynamic Interpolation Algorithm of Sea Surface Height for Future 2D Altimetry Mapping of Sea Surface Height

The sea surface height data volume of the future wide-swath two-dimensional(2D)altimetric satellite is thousands of times greater than that of nadir altimetric satellites.The time complexity of the 2D altimetry mapping reaches O(n^(3)).It is challenging to map the global grid products of future 2D altimetric satellites.In this study,to improve the efficiency of global data mapping,a new algorithm called parallel-dynamic interpolation(PA-DI)was designed.Through the use of 2D data segmentation and fine-grained data mosaic methods,the parallel along-track DI processes were accelerated,and a fast and efficient spatial-temporal high-resolution and low-error enhanced mapping method was obtained.As determined from a comparison of the single-threaded DI with the PA-DI,the new algorithm optimized the time complexity from O(n^(3))to O(n^(3)/KL),which improved the mapping efficiency and achieved the expected results.According to the test results of the observing system simulation experiments,the PA-DI algorithm may provide an efficient and reliable method for future wide-swath 2D altimetric satellite mapping.

Spectral-domain optical coherence tomography dynamic changes and steroid response in multiple evanescent white dot syndrome

Dear Editor,Multiple evanescent white dot syndrome （MEWDS） was first described in 1984 as a rare, acute, unilateral,multifocal retinochoroidal disorder, typically affecting young myopic women. Previous studies with fluorescein angiography （FA） and electrophysiology suggested that MEWDS to be a disease in the retinal pigment epithelium （RPE） or outer retina, while recent studies with spectral- domain optical coherence tomography （SD-OCT） suggested it may be an outer retinal disease due to observation of hyperreflective material in outer retina and subtle disruptionsof the ellipsoid zone without RPE disruption.

Analysis of the Dynamic Response of a Shield Tunnel in Soft Soil Under a Metro-Train Vibrating Load

This paper concerns the impact of an operating metro train on the structure of a shield tunnel lining and its soft foundation. An elastoplastic 3D dynamic finite difference model was established by using the FLAC3D numerical soft- ware. By fully considering the joints, the A-B-K segments and the soft stratum, the dynamic response of the shield tunnel buried in thick, soft soil under the vibrating load induced by a metro train was numerically simulated. The simulation result, for which the joint was considered, was compared with the result when the joint was not considered. The results show that an operating metro train induces a significant dynamic response in the structure of the lining of the shield tunnel and its soft foundation. The severe dynamic response zones of the lining structure are largely distributed in the range of the lower half of the segment-ring and the nearer to the bottom of the segment-ring, the more severe the response. Of two horizontally symmetric, corresponding places on the segment lining, the one near the joint is more severe in its dynamic response than that of the one far from the joint; the nearer the zone of the foundation soil to the lower half of the seg- ment-ring, the more severe the dynamic response. The maximum shear strain of the foundation soil takes place near the joint between two normal segments at the bottom. The dynamic response influenced by joints is more severe than the response not influenced by joints, showing that the non-joint assumption is somewhat impractical.

A framework of region-based dynamic image fusion

A new framework of region-based dynamic image fusion is proposed. First, the technique of target detection is applied to dynamic images (image sequences) to segment images into different targets and background regions. Then different fusion rules are employed in different regions so that the target information is preserved as much as possible. In addition, steerable non-separable wavelet frame transform is used in the process of multi-resolution analysis, so the system achieves favorable characters of orientation and invariant shift. Compared with other image fusion methods, experimental results showed that the proposed method has better capabilities of target recognition and preserves clear background information.

一种室内动态场景下RGB-DSLAM的前端优化方法

针对动态物体会导致位姿估计误差过大的问题,提出一种适用于动态场景下的RGB-D SLAM算法。首先,将目标检测的结果根据物体是否具有自主移动能力划分为动态物体与潜在动态物体,并使用大津法在对应的深度图中完成检测框内区域的前、背景分割;然后,结合检测框的位置坐标与深度信息确定潜在动态物体与动态物体在三维空间内是否关联,初步筛选出具有一定运动概率的潜在动态物体;最后,通过运动一致性检测算法来最终确定是否将潜在动态物体作为动态物体处理。实验部分选用公开的TUM数据集,结果表明,该算法在保持实时性的同时具有较高的定位精度和鲁棒性。

Visual Motion Segmentation in Crowd Videos Based on Spatial-Angular Stacked Sparse Autoencoders

Visual motion segmentation(VMS)is an important and key part of many intelligent crowd systems.It can be used to figure out the flow behavior through a crowd and to spot unusual life-threatening incidents like crowd stampedes and crashes,which pose a serious risk to public safety and have resulted in numerous fatalities over the past few decades.Trajectory clustering has become one of the most popular methods in VMS.However,complex data,such as a large number of samples and parameters,makes it difficult for trajectory clustering to work well with accurate motion segmentation results.This study introduces a spatial-angular stacked sparse autoencoder model(SA-SSAE)with l2-regularization and softmax,a powerful deep learning method for visual motion segmentation to cluster similar motion patterns that belong to the same cluster.The proposed model can extract meaningful high-level features using only spatial-angular features obtained from refined tracklets(a.k.a‘trajectories’).We adopt l2-regularization and sparsity regularization,which can learn sparse representations of features,to guarantee the sparsity of the autoencoders.We employ the softmax layer to map the data points into accurate cluster representations.One of the best advantages of the SA-SSAE framework is it can manage VMS even when individuals move around randomly.This framework helps cluster the motion patterns effectively with higher accuracy.We put forward a new dataset with itsmanual ground truth,including 21 crowd videos.Experiments conducted on two crowd benchmarks demonstrate that the proposed model can more accurately group trajectories than the traditional clustering approaches used in previous studies.The proposed SA-SSAE framework achieved a 0.11 improvement in accuracy and a 0.13 improvement in the F-measure compared with the best current method using the CUHK dataset.

Activity Tendency and Dynamic Characteristics of Shanxi Fault Zone

The tendency and dynamic characteristics of horizontal movement along the Shanxi fault zone have been analyzed using the data obtained from 6 repeated measurements (1996～2001) in the GPS monitoring network arranged along the Shanxi fault zone. The results indicate: (1) the tendentious activity of the present stage is characterized by a W trending movement along the northern segment of the zone, an E trending movement along the southern segment and counter clockwise differential activity on the whole, but the intensity of the tendentious activity is not high. The tendentious differential movement is only about 3 mm/a in the direction perpendicular to the fault zone from the south to the north, and its stretch in the SN direction is only 1 mm/a and mainly occurs along the north segment of the fault; (2) The azimuth of the principal compressive stress field reflected by the tendentious movement is 72°; (3) The property of annual activity is not the same, even contrary to one another or deviates from the tendentious activity. Therefore, the parameters of the strain field derived from them dont reflect the physical characteristics of the basic stress field. (4) The high frequency movement (yearly) does not only exist but is also complicated by an intensity several times higher than that of the tendentious movement; (5) Obvious differential movements, including strike slip, can not be seen in either in secular activity or annual activity on both sides of any fault. The tendentious movement not only verifies the conjecture of “strong in the south and weak in north”, which is the basic feature forcing the western boundary of the North China area, but it also extends to the hinterland of North China. The fact that there is no obvious differential activity on both sides of the fault might indicate that the differential activity among the intraplate blocks is completed by gradual variation in a certain space, rather than the abrupt change bordered by a fault or narrow stripe zone. The obvious dynamic activity might indicate: (1) there is stress disturbance in the basic stress field; (2) the inhomogeneous or non synchronous variation that appeared in the regional stress and strain fields was due to the different physical property of the medium; (3)the response occurred because of a variety of external variations. The movement in 2001 shows that the Daixian county and its adjacent area might be the boundary segment for the relative differential activity. More attention should be paid here.