Abstract:To study the seismic performance of precast shear walls, in which the vertical rebar is connected by a new type of hybrid connection (the vertical steel bars in the edge members are connected by composite straight-thread sleeve and the vertical distributed steel bars are connected by anchoring closed loop reinforcement) and the surface of the precast shear wall is full-length shear resistant groove, quasi-static tests of two precast reinforced concrete shear walls and a cast-in-place reinforced concrete wall were carried out. The shear span ratio of the specimens was 1.91, which was designed according to strong shear weak bending. Results show that the failure mode of the precast shear wall was normal section compression failure, which was consistent with the design. There were no visible cracks in the sleeve and no slip of the connecting steel bar at the end of the test. The seismic performance of the precast wall specimens was basically the same as that of the cast-in-place wall specimen, which meets the requirements of the current code. The normal section compressive bearing capacity of the precast wall specimens was not less than 1.1 times the calculated value of that of the cast-in-place wall specimen, and thus the current specification formula can be used to calculate the normal section compressive bearing capacity of precast shear walls. When the inter-story drift ratio was 1/100, the deformation of the precast wall specimens was basically the same as that of the cast-in-place wall specimen, and the vertical deformation of the section basically agreed with the plane section assumption. There was no obvious slip of the horizontal surface and vertical surface and no obvious opening of the vertical surface, indicating good integrity between the precast and post-cast parts of the wall.

Abstract:To investigate the seismic performance of prefabricated steel frame structure with ribbed thin walls, a full-scale two-story two-span structural model was designed and constructed, and shaking table tests of 12 test groups were performed. Three seismic waves, i.e., El Centro wave, Taft wave, and an artificial wave, were selected for the tests. Peak ground acceleration (PGA) increased from 0.07 g to 1.2 g gradually during the tests process. Failure characteristics, dynamic characteristics, and seismic responses of the model were studied under different working conditions. Results show that the model was in elastic state under frequent ground motion of 8 degrees, and the damage of the model mainly occurred in the ribbed thin walls, while the strain of the steel frame was low under rare ground motion of 8 degrees. With the increase of PGA, the thin walls cracked gradually, and the lateral stiffness of the model degraded gradually, while minor damage occurred in the steel frame. When PGA reached 1.2 g, the lateral stiffness of the model decreased by 37.8% in X direction and 33.6% in Y direction. During the whole tests process, the damping ratio of the model increased gradually between 4.29% and 7.19%, and the acceleration magnification factor of the model was between 0.93 and 2.46, which was obviously lower than that of traditional rigid structure. The maximum inter-story drift ratios of the model under frequent ground motion and rare ground motion of 8 degrees were 1/868 and 1/220, which meet the drift ratio limits according to the design code. The maximum inter-story drift ratio of the model under very rare ground motion of 9 degrees was 1/71, and there was no danger of collapse, indicating good seismic performance. Hence, prefabricated steel frame structure with ribbed thin walls can be applied in high seismic fortification intensity regions.

Abstract:To investigate the distribution of radial expansion compressive stress along the direction of hole depth in static crushing process and explore the influences of hole diameter, hole depth, and constraint degree on radial expansion compressive stress, the radial expansion compressive stress of 21 specimens during static crushing process was measured by means of resistance strain gauge method by using seamless steel pipe to simulate drilling. According to the theory of hoop of concrete filled steel tubular, the ratio of sectional area of steel tube to sectional area of static crushing agent was defined as constraint degree. Experimental results show that the radial expansion compressive stress produced by static crushing agent was not uniformly distributed along the direction of the hole depth, and the expansion compressive stress in the middle and bottom of the hole was generally greater than that in the top. With the increase of the hole size, the difference between the radial expansion compressive stress at the bottom of the hole and the upper part was more significant, the expansion compressive stress at the middle and bottom of the hole in the same time became larger, and more crushing agent overflowed the orifice. Hole depth and constraint degree had little influence on the expansion compressive stress, and the expansion compressive stress generated by static crushing agent in the middle and bottom of the hole increased with the increase of hole depth or constraint degree. The selection of suitable hole depth and larger hole size is more conducive to static crushing agents for developing their expansion potential.

Abstract:The research on longitudinal plate stiffened joints used in UHV long-span transmission towers has been lagging behind. The effect of stiffening on the bearing capacity of joints has not been considered in design process, and there is no explicit calculation method in the national code. Therefore, emphasis was placed on the systematic research of elastic-plastic ultimate bearing capacity of K-type longitudinal plate stiffened tubular joints. The ultimate bearing capacity of the joints was analyzed combined with the variation of geometric parameters, and the failure modes and failure mechanism of the joints were further explored based on the expansion process of the plastic zone of the joints. Effect of each parameter on the ultimate bearing capacity of joints was summarized from the design point of view. Through multiple linear regression, the calculation formula of the ultimate bearing capacity of joints was obtained, and a method of modifying the bearing capacity of longitudinal stiffened joints was proposed. Compared with the results of numerical analysis, the simplified formula had good accuracy, which can effectively simplify the calculation of the bearing capacity of longitudinal stiffened tubular joints and guide the joint design of long-span transmission tower projects.

Abstract:To solve the problem of precision of the current elastic-plastic models in masonry structures and the need for seismic fragility analysis, the equivalent diagonal brace model was introduced into the finite element model of masonry structures. According to the lateral stiffness characteristics of solid wall, window wall, and door wall, the corresponding equivalent diagonal brace width calculation formula was proposed. Based on the equivalent diagonal brace model, the elastic-plastic finite element models of the three types of masonry walls were established and analyzed under low-cycle reciprocating loading. The accuracy and practicability of the models were verified by the test results. On this basis, a typical three-story masonry structure finite element model was established. Following the incremental dynamic analysis method, seismic fragility analysis was carried out by taking the maximum inter-story displacement angle and elastic-plastic energy dissipation difference as damage parameters. The fragility curves of the masonry structure model of transverse wall and that of both longitudinal and transverse walls were compared and analyzed. Results show that the traditional inter-story displacement angle could not fully represent the seismic performance of the longitudinal wall, and the damage parameters based on energy could more accurately reflect the evolution process of structural damage and the performance of the longitudinal wall. The finite element model of masonry structures with appropriate precision should be established according to the needs of fragility evaluation and calculation efficiency.

Abstract:To study the fatigue cracking performance of prestressed reactive powder concrete-normal concrete (RPC-NC) composite beams, four identical model beams were designed and produced by taking the 32-m span T-girder of Chinese railway as the prototype. Static test and constant amplitude fatigue test were carried out on the beams. The number of cracks, crack widths, and crack spacing of the model beams at different cycles were analyzed. Based on the plane-sections hypothesis, the magnified factor of strain in the rebars at crack sections was considered under cyclic loads. Tensile property of RPC and bonding property between RPC and rebar under fatigue loads were discussed, and calculation methods for crack spacing and crack width of RPC-NC composite beams at stable stage of fatigue were deduced based on the bond-slip relationship. For the convenience of design and application, by taking account of the influence of RPC tensile property and the amplification effect caused by cyclic loads, the formula applicable to the crack widths of prestressed RPC-NC composite beams was obtained on the basis of the formula provided by the railway codes of China. The comparison between experimental and calculation results obtained by the two methods show that the maximum crack width of prestressed RPC-NC composite beams can be accurately calculated.

Abstract:To achieve consistent collapse risk, risk-targeted ground motion parameters were investigated. First, based on the ArcGIS Engine software, a discrete algorithm was used to analyze the probability of seismic hazard in Xian region. Then, by considering the seismic fragility of the structure, the risk-targeted ground motion parameters of very rare earthquake, maximum considered earthquake, and design basis earthquake for each control point were obtained by risk integral based on the annual exceeding probability curves obtained from seismic hazard analysis and the ground motion decision parameters obtained from existing research results. The risk-targeted ground motion parameters were denoted as PGA_RV, PGA_RM, and PGA_RD, respectively. The relation between the risk coefficient R_c (defined as the ratio of PGA_RM to the ground motion intensity corresponding to maximum considered earthquake) and K₁ (defined as the ratio of PGA_RV to PGA_RD) and K₂ (defined as the ratio of PGA_RM to PGA_RD) were calculated. Finally, based on the parameter effect analysis, the effect of βR (i.e., logarithmic standard deviation of structural fragility) on PGA_RM, R_c, K₁, and K₂ was investigated with its value chosen to be 0.5,0.6,0.7,0.8,0.9, and 1.0, respectively. Results show that the collapse probability did not increase with the increase of the ground motion intensity corresponding to very rare earthquake. Parameters PGA_RM, R_c, K₁, and K₂ all increased with the increase of βR. The larger βR was, the greater the effect on PGA_RM and R_c was, while the effect on K₁ and K₂ remained unchanged with increasing βR. Compared with PGA_RM and R_c, K₁ and K₂ were more sensitive to βR, and K₁ was the most sensitive to βR among all the parameters. The risk-targeted ground motion parameters obtained in Xian region can provide references for seismic design.

Abstract:In order to study the differences in ground motion selection schemes at home and abroad, code for seismic design of buildings (GB 50011—2010) and minimum design loads and associated criteria for buildings and other structures (ASCE/SEI 7[KG-2mm]-16) were taken as examples, and the corresponding selection results were compared. Through the conversion of design spectrum parameters, ground motion records were selected at the same seismic fortification level under two working conditions according to the methods in two codes, which were then used as dynamic time history input for 4-story and 12-story concrete frame structures to compare the story displacement angle and story shear force. Results show that the selection and amplitude scaling based on a single direction according to the Chinese seismic code may result in a low matching degree in the other direction and a large dispersion of the calculation results. The American seismic code adopts the maximum direction spectrum, which has clearer physical meaning. In addition, the Chinese code only specifies the relative error of the recorded response spectrum, while the American code specifies the lower limit of the spectrum. Although the Chinese code avoids the conservativeness of the results, it may lead to significant disparities between one or more records and the target spectrum, which may cause the dispersion of the results to be uncontrollable. Hence, it is suggested to combine the advantages of the two methods to control the relative error and the lower error limit.

Abstract:In power plant, the novel desulfurization and exhaust smoke integrated steel tower is a complex high thin-walled structure. With the top section suddenly becomes thinner, the steel tower is prone to vibration under wind loads. To reduce the wind-induced vibration of the structure, a tuned mass damper (TMD) vibration reduction scheme suitable for desulfurization and exhaust smoke integrated steel tower was proposed. Based on the fluid-structure interaction theory, structure field and flow field models were established. Wind vibration responses of the structure with or without TMD were calculated by using a numerical wind tunnel method, and the method was verified by on-site wind vibration monitoring. Results show that the vibration modes of the integrated steel tower were mainly local or overall translation, rotation, buckling and torsion. By installing a TMD on top of the tower, the peak value of the displacement dynamic response was reduced by 41% and the RMS value was reduced by 75%. Spectrum analysis results show that the low-frequency component of the structure was excited after installing TMD, and the dynamic response of the main frequency was greatly reduced. Vibration reduction effect was obvious for the desulfurization and exhaust smoke integrated steel tower with TMD, which provides a reference for wind-induced vibration control of similar structures.

Abstract:To investigate the influence of the structure of legs on walking gait patterns of bipedal inverted pendulum, a bipedal model with two-segment legs was proposed. The relationship between the reaction force on the leg and the length of the leg was determined. Functions of motion and the simulation model of the proposed bipedal model were built. Periodic walking gait patterns were solved with Poincare return map and Newton-Raphson algorithm. Influences of model parameters including mechanical energy, attack angle, knee rest angle, and joint stiffness on gait patterns were analyzed. Analytical results show that the initial stiffness of the two-segment leg was large, while it gradually decreased with the increase of the deformation of the leg, indicating that the two-segment leg may soften with deformation. The proposed bipedal model could achieve various walking gait patterns in a certain range of parameters. It could also cover the range of the pacing rate of normal human walking, if the values of the parameters were well selected. Compared with models with linear spring legs, the proposed model was closer to the body structure of real human, and the physical meanings of its parameters were clearer, which lays foundation for experimental verification and parameters calibration of the model.

Abstract:To accurately assess the random fatigue damage of bimodal and trimodal Gaussian process in frequency domain, a cross-mode coupling analysis method is proposed in this paper. Different from the conventional probabilistic method, the proposed method was developed based on spectral discretization method. The response spectrum was discretized into a large number of infinitesimal frequency bands, and the fatigue damage with respective to each band could be computed individually, while a coupling coefficient ξ was introduced to account for the cross-mode coupling between any two frequency components. By using the combination rule developed in this paper, the fatigue damages caused by the infinitesimal frequency bands and their couplings were assembled to obtain the total fatigue damage. Through comprehensive case studies, by taking the results of rainflow counting (RFC) method as reference and comparing with several main frequency-domain methods, it was proved that the proposed cross-mode coupling method is more accurate and robust.

Abstract:To investigate the vortex-induced vibration characteristics and coupling mechanisms of two tandem square cylinders with typical spacing ratio, vortex-induced vibration of two tandem square cylinders was numerically simulated at Reynolds number Re=150 with two spacing ratios of 2.0 and 4.0, and variations of the vibration responses with reduced velocity were studied. The flow pattern and the evolution of the flow field around the cylinders were discussed, and the energy input mechanism for the downstream cylinder was emphatically analyzed. The vibrations of the two cylinders were mainly in the transverse direction for the two spacing ratios. When the spacing ratio was 2.0, the “soft-lock-in” phenomenon occurred in the co-shedding regime for both cylinders (with the lock-in frequency much less than 1.0 in the vibration lock-in region). The maximum transverse amplitudes of the two cylinders both appeared in the vibration lock-in region, and the maximum amplitude of the upstream cylinder was larger. In the vibration lock-in region, due to the abrupt increase of the mean drag force of the upstream cylinder, the mean distance between the two cylinders severely decreased, which disturbed the energy input of the lift force to the vibration of the downstream cylinder. Thus, the vortex in the wake region of the downstream cylinder became unstable. When the spacing ratio was 4.0, the “lock-in” phenomenon was only observed for the downstream cylinder in the reattachment regime. The maximum transverse amplitudes of the two cylinders appeared outside the vibration lock-in region, and the maximum amplitude of the upstream cylinder was much smaller than that of the downstream cylinder. Caused by the mixing of the vortex generated by the two cylinders, the energy input of the lift force for the downstream cylinder was enhanced, enlarging its transverse amplitude. The transverse distance between the two cylinders increased accordingly, which resulted in the parallel vortex street mode of the co-shedding regime in the wake region of the downstream cylinder. In addition, the transverse amplitudes of the downstream cylinder had obvious extreme values in the reattachment regime, and the corresponding wake modes were all parallel vortex street mode of the reattachment regime.

Abstract:Drilled and tapped high-strength steel plates are possible replacements of nuts and can be combined with high-strength bolts to form new assemblies that are used as blind bolts for steel tube column frame joints. In order to study the tensile capacity and failure mechanism of steel plate composite bolts, on the basis of our experimental research results, 64 specimens were tested under monotonic tensile load by ABAQUS software. The thicknesses of four types of steel plates (i.e., Q345B, 45#, Q460C, and Q690D) and the diameters of five specifications of high-strength bolts (i.e., M16, M20, M24, M27, and M30) in grade 10.9 were investigated. Results show that the Q345B, Q460C, and Q690D steel plates can replace the 45# steel plate. The tensile value of the steel plate composite bolts can be designed as 70% of the ultimate tensile load of the screw thread under reliable anchored. For thin steel plates, the bolt was pulled out, the steel plate thread experienced shear failure, the out-of-plane deformation of the steel plate was larger than 1/200 of the support distance, and the screw maximum stress located in the anchoring area. For thick steel plates, the bolt experienced tensile fracture, and the screw maximum stress located in the structural area. The positions of the maximum stress of the steel plates in the two failure modes were both at the first thread. In order to avoid the steel plate thread failure in the steel plate composite bolt, it is suggested that the thicknesses of Q345B, Q460C, and Q690D steel plates should be no less than 1.15d, 1.10d, and 0.95d, respectively. Based on the elastic mechanics and thread-related research, the formula for tensile capacity of high-strength steel plate composite bolts was derived. The calculated results were in good agreement with the finite element results. The research results can provide a reference for the design of high-strength steel plate composite bolts.

Abstract:To improve the anti-fading ability of the classical single carrier system (SC) and multi-carrier system (MC) without changing the carrier scheme, a double-component combination anti-fading scheme based on generalized weighted fractional Fourier transform (GWFRFT) is proposed in this paper, taking the advantage of the high flexibility of the generalized hybrid carrier (GHC) system. As for the SC system, the proposed double-component scheme contains time domain component and time domain reverse component, which enhances the ability to resist time-selective fading; while as for the MC system, the scheme contains frequency domain component and frequency domain reverse component, which possesses higher anti-frequency-selective fading ability. To fully dig out the potential of double-component combined signal form, mechanism of the characteristic advantage is analyzed. Results show that the power allocation between two components and the fading independence of the same symbol in the two components are the key factors affecting the performance of the double-component combination signal. In view of this, equal power allocation scheme and half-block inversion scheme are proposed to further optimize the performance of the double-component combination signal. In that case, generation method of double-component combined signal is proposed and the implementation complexity as well as spectral characteristics are also analyzed. Simulation results illustrate the improvement on anti-fading ability of proposed scheme without occupying extra time and frequency resources, compared with that of classic single-carrier and multi-carrier system. In addition, equal power allocation scheme and half-block inversion scheme enable a further improvement on the anti-fading ability.

Abstract:Ground target tracking experiments using unmanned aerial vehicle (UAV) are difficult to verify and the cost is high. To solve the problem, a simulation platform based on visual simulation software and MATLAB/Simulink was designed and implemented. First, vehicle was selected as ground target. The 3D models of UAV and vehicle were imported into the visual simulation software, and virtual gimbal and camera were set up. Next, the relative motions of the UAV, vehicle, and gimbal were used to simulate scene changes during the tracking process, and the gimbal control algorithm based on motion compensation ensured that the virtual camera always points at the target. Then, the virtual camera was used to capture the image of the target. The image tracking algorithm could track the target in the image and use the target image model to calculate the target position in the world coordinate system. Finally, according to the target position, the reference point guidance method was used to generate the desired roll angle command to guide the UAV to circle around the target. Closed-loop communication between visual simulation software, image tracking algorithm, and MATLAB/Simulink was carried out with UDP and shared memory. In addition, a practical and reliable calibration method was proposed to calibrate the internal matrix of the virtual camera in the visual simulation software. Simulation results demonstrate that the platform can simulate the tracking of ground target by using UAV, and the results have high reference value for engineering. The research in this paper provides a good simulation environment for target tracking experiments and can reduce the cost of experiment effectively.

Abstract:At the cost of high complexity, the existing spatial error concealment algorithms may slightly improve the recovery quality by iterative approximation, and some algorithms are optimized only for certain loss patterns. To achieve a trade-off among performance indices, this paper proposes a non-iterative shrinkage multi-directional (NSM) prediction algorithm aiming at handling various types of loss patterns. For the error concealment of the current missing block, the proposed NSM algorithm firstly adopts an isotropic gradient detector to learn the local feature information of the current extrapolation region. Based on a basic concealment unit with 16-pixel neighbor and eight prediction directions, the multi-directional predictor recovers each pixel of the missing block one by one in a shrinkage filling order, and adjusts the weighting coefficients of the predictor according to the availability of adjacent pixels. During the shrinkage filling of a block, different pixel groups are recovered group by group according to their neighbor-level availabilities, and different missing pixels in a pixel group are predicted one by one according to a priori filling rule. Thus, non-iterative reconstruction with low complexity can be realized. Compared with other spatial error concealment algorithms, experimental results show that the proposed NSM algorithm achieved better overall reconstruction performance under various loss patterns, and realized a competitive performance trade-off among versatility, computational complexity, and recovery quality.

Abstract:Single target tracking is one of the most challenging application scenarios in the field of computer vision. To solve the problems of occlusion, object deformation, and motion blur during tracking, a training method was proposed to train single target tracking network based on generated diverse positive instances, and the problem of scarcity of various training samples was also mitigated. Specifically, during the offline stage, a variational autoencoder (VAE) was employed to encode original samples into latent space. Then, the hard positive data was generated via sampling variables in latent space to improve the diversity of the training data, and a training dataset was constructed by combining the generated data and the original samples. Besides, for the target template and the negative and positive samples of the training sequences, a probability triple loss function was utilized to train the tracking network. The relation between the positive and negative samples was investigated to improve the discriminative power of the tracking network. During the test stage, the pretrained Siamese neural network (SNN) was used to track the target, and the position of the target at the moment could be determined by correlating the target template and the search area. The experiment shows that the proposed algorithm improved the robustness and accuracy of SNN tracking in the cases of interference of similar objects and deformation, fast motion, rotation, motion blur, and occlusion of target during movement, and achieved real-time tracking performance.

Abstract:To improve the accuracy and stability of satellite clock bias (SCB) prediction, an adaptive SCB prediction method was proposed based on the combination of particle swarm optimization (PSO) and weighted grey regression with index function and linear function approximation. First, considering the phenomenon of frequent SCB clock jumps, the clock jump data was detected and excluded through median absolute deviation (MAD) before modeling, and the piecewise linear interpolation method was used to complement the missing clock bias data. Then, in order to deal with the system noise of the SCB data, the three-point smoothing method was used to smooth the clock bias data. After the processing, an SCB prediction model based on the combination of exponential function and linear function approximation for weighted grey regression was established. Aiming at the problem that the increasing precision factor in the model was difficult to determine, a PSO algorithm was used to adaptively optimize the accuracy increasing factor. Finally, a 6-hour forecast test was performed by adopting the post-accurate SCB product released on the IGS server and combining with four typical trends. Experimental results show that the prediction performance of the proposed method was significantly better than those of other commonly used models. Compared with the quadratic polynomial prediction model (QPM), grey prediction model (GM (1,1)), modified exponential curve method (MECM), and autoregressive moving average model (ARMA), the 6-hour average forecast accuracy (RMS) and stability (Range) of the proposed method were increased by 79.10%, 44.00%, 80.70%, 32.30%, and 63.10%, 29.80%, 77.60%, 26.30%, respectively.

Abstract:In the process of panoramic video target tracking, the target deformation and scale changes caused by light change, interference of similar background, and object moving may result in target drift or missing, leading to low success rate and poor robustness. To address these issues, a target tracking method based on long short-term memory (LSTM) network and improved Real-Time MDNet (RT-MDNet) network was proposed. First, shallow convolution neural network was utilized to extract features, and adaptive RoIAlign was adopted to reduce pixel loss in the convolution process. Then, the weight of the last layer of the full connection layers was updated online by utilizing the target features to achieve foreground background separation and extract the target area. Lastly, the scale of the target box was selected adaptively by means of LSTM, and the target position information was thus obtained. Experimental results show that monocular vision algorithm could hardly adapt to the scale change and background change when applied in panoramic dataset, while the proposed method that utilizes 3-layer LSTM network to construct scale prediction module could effectively solve these problems. The algorithm can efficiently deal with the situations of small target, target occlusion, and cross motion of multiple targets in target tracking while maintaining accuracy, achieving better visual effect and higher overlap rate score.

Abstract:Integrated circuit chips work in complicated electromagnetic environments, which is susceptible to soft errors caused by high-energy particles. In a chip, the memory accounts for more than half of the total area, making it important to improve the reliability of the processor by hardening the memory. Therefore, a dual modular redundancy hamming (DMRH) code is designed and proposed in this paper, which can mitigate one-bit and two-bit upset in memory. First, logic optimization was carried out in the hamming code encoder to reduce the delay of the circuit, and the parity generated by this module was processed with dual modular redundancy technology, which was used as the output of the DMRH encoder. Then, the combinations of each parity and original code were processed according to the hamming decoding rules, and the revised data and two-bit upset flag were obtained. Through analysis, it was found that when the two-bit upset did not happen in the original code at the same time, the correct output could be obtained according to the two-bit upset flag. Finally, the layout segmentation technology was used to suppress the two-bit upset in the original code, which further improved the reliability of the memory. In this study, three types of DMRH codes with word lengths of 4,8, and 11 were realized. Compared with other correction codes, results show that the circuit delay of the codes obtained in this study was 85%, 89%, and 96% of the eight-bit hamming code, which was lower than the BCH codes.

Abstract:To improve the limited response ability of wheeled mobile robots (WMRs) and ensure that WMRs track the given reference trajectories rapidly, this paper utilizes information preview (i.e. future reference information) to design a preview control strategy and driving voltage of wheeled mobile robot (WMR). First, a virtual controller was proposed for the kinematic model to ensure that the WMR tracks the given reference trajectory asymptotically. Then, a linear discrete-time state-space model of the proposed virtual controller was established, and the tracking problem was thus converted into a linear quadratic regulation problem for systems with given input on the basis of the dynamic model of the WMR and the linear model of the virtual controller. Finally, the feedback gain of the optimal preview control strategy was provided based on the solution of the Riccati equation. The proposed strategy for virtual controller modeling could be adopted for any discrete signal, which can facilitate signal processing and application. In addition, due to the utilization of the reference information in advance, the proposed strategy could realize rapid tracking with given trajectories. Simulations for sinusoid-type reference trajectories were carried out, and results show that the velocity and trajectory tracking errors of the WMR tended to zero rapidly, which verifies that the proposed preview control strategy is feasible.

Abstract:Paraphrase identification is widely used in question answering system, plagiarism detection, and personalized recommendation. Since the existing paraphrase identification techniques are lack of effective feature extraction mechanism, a new paraphrase model was proposed. Different from previous works which normally adopt the “encoding-matching” mode, the proposed model adopts the “encoding-matching-extraction” mode by adding feature extraction layer to better acquire classification information. The proposed model is consisted of six layers: input layer, embedding layer, encoding layer, matching layer, feature extraction layer, and output layer. The encoding layer utilizes contextual bi-directional long short-term memory network (BiLSTM) with self-attention to encode context of sentences, which can make full use of contextual information in both forward and reverse directions of a sentence. The matching layer uses several matrix operations to get sentence pair matching information from different angles. The extraction layer chooses Xception as the feature extractor to better extract classification information from the matching results. Moreover, this paper combines GloVe word vectors, character vectors, and additional feature vectors as the final embeddings, which carries richer information than ordinary pretrained embeddings. Results show that the proposed model achieved competitive results on two public datasets: Quora Question Pairs (as a representative of large datasets) and SemEval-2015 PIT (as a representative of small and medium datasets).

Abstract:To solve the problems that the basic salp swarm algorithm (SSA) converges slowly and is easy to fall into the local optimum, a new modified SSA based on refracted opposition-based learning (ROBL) and adaptive control factor (RCSSA) was proposed. First, the ROBL mechanism was used to calculate the refracted opposite solution of individual solution, which greatly improved the convergence accuracy and speed of the algorithm. Then, the adaptive control factor of the leader in SSA was introduced into the position update of the follower, which could effectively control the entire search process and increase the local exploitation ability. To verify the optimization performance of the proposed algorithm, seven unimodal, 16 multimodal benchmark functions, and one engineering design problem were utilized to investigate the algorithm. In the experiment, two SSAs improved by single strategy were introduced to verify the proposed algorithm, and five state-of-the-art intelligent optimization algorithms such as whale optimization algorithm were added to further verify the superiority of the algorithm. Research results show that the addition of ROBL mechanism and adaptive control factor could effectively enhance the exploitation and exploration abilities of the basic SSA for both low-dimensional and high-dimensional benchmark optimization problems, and the optimization performance of RCSSA was better than most other intelligent algorithms.

Abstract:Compared with the common color video, infrared video is easily affected by the surrounding environment. In infrared pedestrian tracking, the appearance contour and gray distribution of the pedestrian target often have great changes, which lead to the difficulty of tracking. To solve this problem, this paper proposes a VPSiamRPN (Video prediction with Siamese Region Proposal Network) infrared pedestrian target tracking system. Aiming at the factors that seriously affect the performance of infrared pedestrian tracking (such as target deformation, target occlusion, and background clutter), the image prediction function of PredNet (Deep Predictive Coding Networks for Video Prediction and Unsupervised) was designed and applied to SiamRPN (Siamese Region Proposal Network) to improve the similarity between the tracking template and the detected target, so as to improve the tracking ability to the infrared pedestrian target. Nine comparative experiments were carried out by changing the number of layers of the network, the number of target images and frames used for prediction, and the tracking strategy of the network. On PTB-TIR dataset, experimental results show that the success plots and precision of theinfrared target recognition in thermal crossover, intensity change, occlusion, scale variation, and other attributes were much higher than those of SiamRPN, indicating good performance of infrared pedestrian tracking, which will have broad application prospects in this field.