Abstract:

Abstract: In order to solve the problems of complex algorithm of traditional leader-follower method in formation control of multi-robot system and difficulty in completing circular formation of multi-robot system with common formation control law, the formation problem of multi-robot system was transformed into tracking control problem among robots by improving the traditional leader-follower method, and a velocity compensation algorithm based on position information for multi-robot formation was proposed in this study. The formation control model of robot with velocity compensation algorithm is established, and the formation control law was designed based on the pose error between the following robot and the virtual robot, and it is proved theoretically that the proposed control law could complete the multi-robot formation task. Then, on the basis of studying the multi-robot formation problem, the obstacle avoidance problem in the multi-robot formation process is further studied. The classical artificial potential field method was introduced, and the artificial potential field method was combined with the speed compensation algorithm of this study. The combined algorithm could enable the multi-robot system to maintain formation operation, and not only preventing the robots in the system from colliding with each other, but also adaptively avoiding obstacles in the surrounding environment. The results showed that multi-robots could not only complete the formation task efficiently but also successfully complete the obstacle avoidance task when encountering obstacles. Finally, the proposed algorithm was verified by experiments on multi-robot simulation and physical platform. The algorithm could reduce the number of calling parameters, simplified the formation algorithm, and improve the formation efficiency.

Abstract: In order to ensure cyberspace to provide more reliable information, fraud detection became more important. Existing methods only considered the static dense sub-graphs formed between user comments when detecting fraudulent users, while ignored the abnormal behavior of users during the comments, which led to reduced accuracy. Meanwhile, further manual verification was often required to verify the reliability of the results in practice. For this problem, this paper proposed the CPOFD method, which used a new measure “comparative equivocation”. This measure mainly included topological connection information to detect fraud groups in a more aggregated manner. Specifically, this metric emphasized the dynamic comparison between fraudsters and normal users, so that the algorithm could effectively detect the fraudster′s abnormal behavior in terms of topological connections, timestamps, and scoring information. At the same time, this method combined the clustering algorithm based on the density sub-graphs and the decision tree classification algorithm to group users in the social network effectively, and used the simulated annealing algorithm to optimize the pruning when classifying the clusters, so as to find the approximate optimum solution more concisely and quickly. The time complexity of the algorithm was linear to the number of fraudsters, and it had high scalability. In experiments based on the Yelp dataset, the accuracy of the CPOFD method for fraudulent public opinion detection reached more than 98%, which verified the effectiveness of the CPOFD method.

Abstract: In order to solve the underestimation problem of the twin delayed deep deterministic policy gradient algorithm in the reinforcement learning actor-critic framework, deep determinstic policy gradient based on mean of multiple estimators(DDPG-MME) was proposed. The method contained one actor and k(k > 3)critics, and the minimum of the output values of two critics and the mean value of the remaining (k-2) critics was calculated first, and then the average of the two values as the final value was taken to calculate the TD error. Finally, we update the critic network based on the TD error, and the actor network is updated based on the value of the first critic. The weighting operation of the method could alleviate the underestimation problem of the twin delayed deep deterministic policy gradient algorithm and reduces the estimation variance to a certain extent to achieve more accurate Q-value estimation. The expectation and variance of the estimation error of our method, deep deterministic policy gradient was analyzed theoretically, and twin delayed deep deterministic policy gradient, and the accuracy and stability of the method was demonstrated. And the experimental results in four MuJoCo continuous control environments, such as Reacher-v2, HalfCheetah-v2, InvertedPendulum-v2 and InvertedDoublePendulum-v2, showed the superior final performance of the deep deterministic policy gradient based on mean of multiple estimators algorithm over TD3 and DDPG, and the results showed that the final performance and stability of our algorithm were significantly better than the comparison algorithms under the same hyperparameters (network structure, reward function, environment parameters, batch size, learning rate, optimizer and discount factor) settings as the comparison algorithms.

Abstract: In view of the problem that the tamper detection algorithm based on SIFT feature matching had， such as unclear location of tamper region and interference of SIFT mismatch on tamper region resolution, a method of image tamper location based on mathematical morphology and feature matching was proposed. The method detects the extreme value of the image in the scale space, locate the key points of the image features, determine the gradient direction of the image, and form the feature point descriptor. Then based on the Euclidean distance, the feature descriptors were matched by two matching blocks, the tamper region is roughly located, and the redundant matching block was removed by mathematical morphology. SIFT feature point of suspected region pairs are extracted, the feature points were matched after PCA dimensionality reduction of SIFT feature matrix, and the matched feature points were purified by using the correlation coefficient. Finally, according to the distribution of matching feature points, the tampered area is relocated to accurately obtain the copy-move tamper area. In the experiment, a group of representative images were selected and detected after adding copy-move, rotation and zooming attacks. The results showed that the prorosed method could locate the location of the copy-move tamper area more accurately, and had higher accuracy in the image copy-move tamper detection.

Abstract: As one of the machine learning methods, twin parametric insensitive support vector regression (TPISVR) had a simple mathematical model and good learning performance. It was especially suitable for solving data regression problems with structural heteroscedasticity noise. However, the training speed of TPISVR was low, and the training efficiency needs to be improved. The traditional algorithm of TPISVR could be reduced to solve two quadratic programming problems with inequality constraints by transforming dual problems. However, this method of solving quadratic programming problems with large number of samples would be restricted by time and memory, which was the key to the low training efficiency of TPISVR. In this study, the positive sign function was introduced to transform the two quadratic programming problems of TPISVR into two non-differentiable unconstrained optimization problems. Secondly, CHKS smooth function and regular term were introduced to regularize TPISVR model, and smooth approximation was made to the non-differentiable unconstrained optimization problem, so as to transform the non-differentiable model into a differentiable unconstrained optimization problem. The new model was solved by Newton Armijo method with fast convergence speed, and a smooth twin parameterized insensitive support vector regression machine (STPISVR) was proposed. Finally, it was proved theoretically that STPISVR model was convergent and had arbitrary order smoothness; In order to verify the effectiveness and feasibility of the algorithm, simulation experiments were carried out on the artificial data set and UCI data set commonly used in machine learning. The experimental results showed that compared with other machine learning methods, STPISVR achieved higher training efficiency without reducing the accuracy.

Abstract: To solve the problems of insufficient multiscale feature extraction and object edge blur in image-based salient object detection, an end-to-end salient object detection model was proposed based on attention embedding pyramid feature and stepped edge optimization. Firstly, the attention embedded dense atrous Pyramid Module (AEDAPM) composed of multiple dilated convolutions was designed to obtain rich and effective multi-level multi-scale features without reducing the feature resolution； Secondly, in order to solve the problem of blurring the edges of salient objects, a stepped edge optimization module (SEOM) is proposed, which gradually supplements spatial detail information during the process of feature restoration resolution, so that the salient objects detected by the model could have clear edge contours. The method in this paper was compared with 12 state-of-the-art saliency methods under 3 common indicators on 5 public datasets, such as DUTS-TE, ECSSD, DUT-OMRON, HKU-IS, and PASCAL-S. The experimental results show that the method proposed in this paper can obtain more accurate and clearer saliency results. In addition, the ablation study also fully proved the effectiveness of the AEDAPM and the SEOM proposed in this study.

Abstract: Recommendation methods of deep learning-based on review text mainly means to describe the feature information of users and items by terms of review texts, by rating relationship between users and items to improve the recommendation performance. Existing studies ignore the interpretable contribution of sentimental features on the rating prediction. To solve this problem, by incorporating the roles of review text and sentimental polarity orientation in the embeddings of users and items, respectively, a sentimental attention recommendation method was proposed based on review text (IncorRAS-Rec); Firstly, CNN (convolutional neural network) was used to handle review sets for users and items, represent the review features of users and items, and obtain relevant users features and items features; Then, by combining users′ rating preference for items, users and items embedding with reviews′ sentimental features were learned. Secondly, by aggregating reviews′ relevant sentimental feature information for users and items in terms of attention mechanism, the embeddings of users and items were learned; Finally, the users ratings on the items were predicted based on users and items embeddings together with their bias information. The experimental comparison and analysis were carried out on public Amazon datasets, to evaluates the effectiveness of the model performance. Experimental results showed that the proposed IncorRAS-Rec model not only could outperform other traditional methods in terms of RMSE(Root mean square error) and MAE(Mean absolute error) metrics, but also implement the explanatory role of sentimental features in rating prediction.

Abstract: In order to study the crack propagation problem, the smooth finite element method (S-FEM) was combined with the extended finite element method (X-FEM) to form the smooth-extended finite element method (S-XFEM) algorithm and program.Firstly, based on the theory of S-XFEM, the calculation method of crack stress intensity factor was formed, and the general program suitable for general finite element method was written to analyze the unidirectional tensile model of an infinite plate with a central crack. By changing the mesh size and crack angle to analyze the influence of crack stress intensity factor, the stress intensity factor values of type I, type II and type I, II composite cracks obtained by the algorithm were compared with the theoretical values. With the decrease of mesh size, the stress intensity factor values of type I and type II obtained by S-XFEM approached to the exact solution continuously. With the increase of crack degree, the stress intensity factor value of type I increased continuously, and the stress intensity factor of type II increased first and then decreased with the angle.The larger the angle was, the closer the crack was to the type I crack. Therefore, the stress intensity factor calculated by the S-XFEM had high accuracy and good adaptability.Secondly, on this basis, the crack propagation criterion was introduced, and the MATLAB programming was used to study the crack propagation path of four-point bending experimental model of the plate with side cracks and the beam with holes. The results showed that the crack propagation path obtained by this example was consistent with that obtained by references.Finally, in this paper, S-XFEM was used to study the crack propagation，compared to using other methods for crack propagation research, it was not necessary to re-divide or encrypt the grid in the area near the crack,which would make the analysis process more simple, improve the computational efficiency.

Abstract: In order to improve the prediction accuracy of brake judder by brake dynamic model, a brake dynamic model was established, in which the contact form between components and the displacement input of model were more in line with actual situation. Firstly, the finite element model of ventilated disc brake was established by using finite element software, and the disc thickness variation caused by friction heat was calculated by using thermal-structure coupling method. Then, an eight-degree-of-freedom dynamic model of brake system was established in the contact form of face to face between disc and block. At the same time, the initial disc thickness variation of brake disc with judder fault was measured by experiment. The initial thickness variation was superimposed with thickness variation obtained through thermal-structural coupling simulation, and the superimposed displacement was taken as displacement of brake block in dynamic model. The brake pressure variation and brake torque variation was predicted by dynamic simulation. Finally, the reliability of model was verified by experiments. The results showed that the error between simulation result and test result of brake pressure variation is 13.18%, and the error between simulation result and test result of brake torque variation was 11.15%.It was concluded that the superposition input of initial disc thickness variation and thickness variation generated by thermal deformation of disc could make the brake dynamic model have a good accuracy in predicting brake judder, and the brake surface contact model established could effectively predict brake judder.

Abstract: A 40Cr/graphite spherical plain bearing was developed to meet the requirement that PTFE and copper parts were not allowed to be used in self-lubricating spherical plain bearings for a nuclear power damper. In order to study the influence of the swing condition of the hydraulic damper on the spherical plain bearing, 50 000 swing friction tests were carried out on the spherical plain bearing under the radial load of 10 kN, the swing frequency of 1 Hz and the swing amplitude of ±5° by using the self-developed spherical plain bearing load simulator. Then the tribological properties of graphite and 40Cr were analyzed by using the friction torque, radial wear and friction temperature recorded during the test. Finally, the wear mechanism of the spherical plain bearing under test conditions was discussed by the microstructure of graphite inner ring obtained by double beam scanning electron microscope. The results showed that the wear quantity, friction coefficient and temperature increased greatly in the running-in stage, and then transition to the stable wear stage. At the later stage of the stable wear stage, they all showed a slight upward trend. During the whole swing wear process, the bearing moved well without jamming. After the test, the radial wear amount of graphite inner ring was 112 μm, and the mass decreased by 0.02 g. The wear mechanism of graphite inner ring is different at different contact positions. The core bearing area is mainly abrasive wear, fatigue wear and adhesive wear, while the non-core bearing area is mainly abrasive wear.

Abstract: In order to study the influence of pitting faults on the contact force characteristics and load sharing characteristics of double-ring herringbone planetary transmission system, a virtual prototype model of the system was established based on SolidWorks, KISSsoft and ADAMS when the sun wheel was healthy and different types of pitting faults existed. and the numerical solution of the contact behavior of the system was solved by GSTIFF integral dynamic solver. On this basis, the characteristics of the contact force in time domain and frequency domain were studied. The load sharing characteristics of the system were analyzed according to the contact force of the gear. The results show that when the system is fault-free, the contact force of the gear was relatively stable and showed periodic change, with a certain time-varying characteristics. The system has good load sharing performance, and the fundamental frequency was consistent with the theoretical meshing frequency of the system. There are harmonic frequencies in the frequency domain diagram, and there were obvious side bands on both sides of the frequency. In the case of pitting failure, the periodic peak appears in the time domain diagram of contact force. With the increase of pitting quantity and pitting depth, the peak value and load sharing coefficient of contact force increase, the load sharing performance became worse, and the side bands on both sides of fundamental frequency and harmonic frequency became more and more dense. The influence of circular pitting on load sharing performance of the system was greater than that of rectangular pitting, and the influence of the increase of pitting quantity on load sharing performance of the system was greater than that of the increase of pitting depth. The results could provide a basis for gear dynamics and fault diagnosis research.

Abstract: In order to investigate the supporting capacity of the new prefabricated recyclable double-row pile supporting structure,the ABAQUS finite element calculation software was used to analyze and study the excavation process of the foundation pit in the urban village reconstruction plan in a certain area of Zhengzhou City, Henan Province. The corresponding structural parameters were used to construct a new supporting system of front-row piles, back-row piles, connecting beams and plates, a new three-dimensional double-row pile support model was established for the foundation pit, and the excavation process of the foundation pit and support process of the supporting structure were simulated. The Mises stress and displacement deformation of the supporting pile and connecting beam was calculated in the structure. Based on the test results, it is shown that as the excavation depth of the foundation pit increased, the front-row piles and the back-row piles were similar in the deformation form of the new fabricated recyclable support structure, the upper part of the pile was relatively large in force and deformation, and the maximum deformation of the pile would move down with the excavation depth, the connecting beam could better coordinate the deformation of the front-row piles and the back-row piles during the excavation process, and effectively transfer the force of the front-row piles to the back-row piles for common force. During the excavation of the foundation pit up to 12 m, the maximum deformation of the connecting beam was less than 5 mm. At the same time, the establishment process of the foundation pit support model was introduced in detail in terms of the selection of model parameters, the establishment of analysis steps and the setting of boundary conditions, which would provide a reference for further improving the model in the future.

Abstract: In order to study the variation of larger unbalanced earth pressure of backfill behind abutment caused by vehicle load, several reciprocating low-cycle pseudo-static tests on interaction of soil-abutment-steel H-pile with or without considering the effect of unbalanced earth pressure of backfill behind abutment were carried out. A comparative study was performed to find magnitude, distribution law and calculation method of larger unbalanced earth pressure of backfill behind abutment. The test results indicated that under the test conditions of this study, with the increase of displacement load, the distribution law of earth pressure for the abutment-steel H-pile interaction under backfill behind abutment (AHP model)changed from “triangle“distribution to “trapezoidal” distribution. The distribution law of earth pressure for the abutment-steel H-pile interaction under larger unbalanced earth pressure of backfill behind abutment (LAHP model) was a "trapezoidal" distribution. The earth pressure coefficient for LAHP model was larger than that of AHP model under the reciprocating displacement loading. The existing calculation method of earth pressure coefficient K was not applicable to IAJBs. The earth pressure coefficient of AHP and LAHP models were 12.34 and 16.16 respectively. They were significantly larger than active earth pressure coefficient (Kac=0.24) and earth pressure coefficient at rest (K0=0.43) from the code of General Specifications for Design of Highway Bridges and Culverts (JTG D60—2015) as well as the Coulomb theory(KpC=9.96) when δ/h reached 0.016 (δ/h=0.016). In addition, the existing calculation method of earth pressure coefficient K was also larger than the existing theoretical calculation methods(the Rankine theory KPR=3.69). In addition, the relevant research in this study could provide reference for the design and calculation and the formulation of relevant specifications in IAJBs

Abstract: The current research on the schedule delay risk of large public projects mainly focused on identifying the risk factors of schedule delay and analyzing the interaction between the risk factors, there was a lack of quantitative analysis of the sensitivity and importance of risk factors and the intensity of influence between risk factors. In this study, Bayesian network and interpretive structure model were used in the assessment of schedule risk of large and complex public projects. Based on the collection of 7 deferred risk factor systems derived from real cases, 24 major deferred risk factors were refined, then the explanatory structure model was used to process the risk factors into four different levels, and the hierarchical structure diagram was transformed into a Bayesian network. Finally, according to the established Bayesian network, the risk factors were collected and evaluated based on actual project data. The collected data were imported into GeNIe, the parameters of the Bayesian network model was learnt, and obtain the probability distribution of the risk level of schedule delay based on the sample data.And using Bayesian network reverse reasoning, sensitivity analysis, and influence intensity analysis were used, the key factors and sensitive factors that lead to schedule delays in large and complex public projects are clarified, and a scientific and effective theoretical basis and control tools were offered for schedule risk management.

Abstract: Indium tin oxide(ITO) material powder, as a key factor affecting the preparation of target, played an important role in the quality of the prepared target. In order to improve the suppression performance of nano indium oxide and tin oxide mixed powder, to improve the quality of ITO raw material powder, in this experiment, nano-indium oxide and tin oxide powder were mixed with dispersant and binder, and then ball milled, and then pressure spray granulation technology was used to prepare near spherical indium oxide tin powder. The effects of slurry viscosity, binder content, slurry solid content and inlet air temperature on the surface morphology, bulk density, tap density and particle size distribution of indium tin oxide powders were studied. The results showed that when the mass fraction of binder polyvinyl alcohol (PVA) was 1%, the solid content of indium tin oxide slurry was 50%, and the inlet air temperature of pressure spray granulator was 200 ℃, the indium tin oxide granulating powder with good performance could be obtained. Spray granulation process can effectively improve the bulk density and tap density of nano indium oxide and tin oxide mixed powder, and improve its pressing performance, which would be favorable for subsequent sintering to obtain high density ITO target.

Abstract: In order to enhance the storage stability and road performance of SBS modified asphalt, C9 petroleum resin, a cheap and excellent by-product of petroleum cracking, was compounded with SBS modifier to prepare high-performance modified asphalt. The effects of C9 petroleum resin with different dosage on the physical properties, storage stability, high temperature performance, low temperature performance and fatigue performance of the composite modified asphalt were investigated. It was found that the addition of C9 petroleum resin could reduced the softening point difference ΔTR&B and the distance between the main curves of rheological properties, improve the storage stability of modified asphalt. Due to its high softening point, the asphalt modified by C9 petroleum resin and SBS had excellent high temperature deformation resistance. Under the same pressure and temperature, the Jnr of the modified asphalt was significantly reduced than that without C9 petroleum resin. Proper amount of C9 petroleum resin was helpful to improve the low temperature crack resistance and fatigue resistance of modified asphalt. However, owing to the brittleness of C9 petroleum resin at low temperature, its dosage should 6%. The storage stability, high temperature deformation resistance, low temperature crack resistance and fatigue resistance of SBS modified asphalt were significantly enhanced by adding C9 petroleum resin.