Narrow Results

In this paper, we propose an adaptive continuous trajectory recognition method for badminton sports scenarios. First, we adopt an integrated pre-training model with high generality and expressive ability as the infrastructure. By pre-training large-scale badminton game data, the model can learn the features of badminton sports from rich and diverse trajectory data. Second, to adapt to the trajectory differences in different scenarios, we introduce an adaptive mechanism. By extracting and encoding features from real-time acquired badminton motion trajectories, the model can adjust its own parameters and weight assignments according to the changes in the current scene to optimally recognize and track the motion targets. The algorithm also employs the Kalman filter aggregation Enhanced Correlation Coefficient (ECC) method of the motion model to improve the prediction accuracy. Finally, a series of experiments and comparisons are conducted to validate the effectiveness of the approach. The results show that our proposed adaptive continuous trajectory recognition method for badminton achieves better performance in different scenarios and various complexities, and has higher accuracy and robustness than the traditional method. The FDA-SSD model operates at 28.7fps, which is about 16.5% faster than the traditional SSD. In the target tracking experiments, the target tracking mean squared error is about 1.0%. In the target tracking experiments based on the FDA and geometrically constrained localization method, the root-mean-square error of the target tracking is less than 4.72cm.

In recent years, image denoising methods based on total variational regularization have attracted extensive attention. However, the traditional total variational regularization method is an approximate solution based on convex method, and does not consider the particularity of the region with rich details. In this paper, the adaptive total-variation and nonconvex low-rank model for image denoising is proposed, which is a hybrid regularization model. First, the image is decomposed into sparse terms and low rank terms, and then the total variational regularization is used to denoise. At the same time, an adaptive coefficient based on gradient is constructed to adaptively judge the flat area and detail texture area, slow down the denoising intensity of detail area, and then play the role of preserving detail information. Finally, by constructing a nonconvex function, the optimal solution of the function is obtained by using the alternating minimization method. This method not only effectively removes the image noise, but also retains the detailed information of the image. The experimental results show the effectiveness of the proposed model, and SNR and SSIM of the denoised image are improved.

We describe a hybrid pattern-matching algorithm that works on both regular and indeterminate strings. This algorithm is inspired by the recently proposed hybrid algorithm FJS and its indeterminate successor. However, as discussed in this paper, because of the special properties of indeterminate strings, it is not straightforward to directly migrate FJS to an indeterminate version. Our new algorithm combines two fast pattern-matching algorithms, ShiftAnd and BMS (the Sunday variant of the Boyer-Moore algorithm), and is highly adaptive to the nature of the text being processed. It avoids using the border array, therefore avoids some of the cases that are awkward for indeterminate strings. Although not always the fastest in individual test cases, our new algorithm is superior in overall performance to its two component algorithms — perhaps a general advantage of hybrid algorithms.

In this paper, we present a fully distributed random walk based clustering algorithm intended to work on dynamic networks of arbitrary topologies.

A bounded-size core is built through a random walks based procedure. Its neighboring nodes that do not belong to any cluster are recruited by the core as ordinary nodes. Using cores allow us to formulate constraints on the clustering and fulfill them on any topology.

The correctness and termination of our algorithm are proven. We also prove that when two clusters are adjacent, at least one of them has a complete core (i.e. a core with the maximum size allowed by the parameter).

One of the important advantages of our mobility-adaptive algorithm is that the re-clustering is local: the management of the connections or disconnections of links and reorganization of nodes affect only the clusters in which they are, possibly adjacent clusters, and at worst, the ordinary nodes of the clusters adjacent to the neighboring clusters. This allows us to bound the diameter of the portion of the network that is affected by a topological change.

Despite the rule that cyclists must ride on the right half of the road is written into the state vehicle code, the phenomenon of riding against the bicycle flow is still serious. To investigate the effect of bicycles going in the wrong direction, a Bi-Directional Adaptive EBCA model is developed in this paper. The phase transition F-J as well as the phase transition F-S-F are suggested by observing the spatial-temporal pattern. The deterministic case that the linear relationship between the average flow rate and the bicycle number disappears when the average density exceeds a particular value is shown. Under the stochastic case, the impacts of the avoiding probability Ps and the returning probability Pr on the traffic system are analyzed. The results of the simulation are in good agreement with the realistic bicycle flow.

This paper studies the generalized synchronization of hyperchaos systems, and a new method, by which adaptive generalized synchronization of chaotic systems with a kind of linear and nonlinear relationship between the drive and response systems can be achieved, is proposed. This new method has more extensive application scope. Based on the Lyapunov stability theory, the correctness of the proposed scheme is strictly demonstrated. It is also illustrated by applications to hyperchaotic Chen system and hyperchaotic Lorenz system and the simulation results show the effectiveness of the proposed scheme.

In this paper we propose a second order hybrid reconstruction strategy for the fully adaptive multi-resolution scheme based on the second order finite volume method. The fully adaptive multi-resolution scheme is an adaptive grid method for solving the hyperbolic conservation laws. To improve its robustness, the third order central reconstruction is replaced by the second order hybrid reconstruction, i.e. the anti-ENO (essentially non-oscillatory) reconstruction and the ENO reconstruction have been adopted for the multi-resolution analysis and the inverse multi-resolution analysis, respectively. Several numerical examples indicate that this new hybrid reconstruction strategy is much less sensitive to the tolerance than the central reconstruction.

This work presents a general framework of finite-difference hybrid scheme which contains a linear central scheme and a nonl-inear WENO scheme. A new optimal-designed shock sensor is used to distinguish the smoothness of flowfield and a binary-type weighting function is used to switch sub-schemes rationally. Based on the above improvements, the effects of different combinations of each component within the hybrid scheme are characterized in linear advection equation and Euler equations. The maximum reference threshold values are provided. Extensive test cases indicate the hybrid scheme’s numerical robustness, low-dissipation, and superior computational efficiency. Specifically, benefited from the high-resolution shock sensor which can accurately perceive shocks without excessive misidentifications, the hybrid scheme can achieve non-oscillatory solutions, and resolve more vortices in smooth regions compared to the original shock-capturing scheme. Meanwhile, the superiority of the hybrid scheme is further confirmed in the Reynolds-averaged Navier–Stokes equations/Lager Eddy Simulations (RANS/LES) for the DLR scramjet combustor case with viscous terms and/or sub-grid scale models are used. The present hybrid framework can be easily implemented within the existing numerical simulation code framework.

To reduce the calculation cost and improve the accuracy of flow field prediction, an adaptive proper orthogonal decomposition (APOD) surrogate model based on K-means clustering algorithm was proposed to reconstruct the flow field of impeller. The experiment samples were designed by introducing the perturbation of the blade control parameters such as blade wrap angle and blade angle of outlet. K-means clustering algorithm was used to classify the sample blade shapes, and find out the cluster of the objective blade. The snapshot set, which consisted of the blade shape and the flow field data of impeller, can be described as a linear combination of orthogonal basis by POD method. The radial basis function (RBF) was used to fit the orthogonal basis coefficients of the objective blade, and then the flow field of objective impeller was reconstructed. The traditional fixed sample POD (FPOD) method and the proposed APOD method were used to reconstruct the flow field in impeller, respectively, and the prediction results of the two methods were compared and analyzed. The results show that the proposed APOD method could quickly and accurately reconstruct the objective flow field. The flow field prediction accuracy of the APOD method is significantly higher than the FPOD method, and the calculation time for the flow field prediction is less than 1/360 of the CFD.

Partial discharge (PD) detection is an effective means to find high-voltage cable defects. However, various interference signals affect the PD signal in practical applications, resulting in wrong judgment. In order to improve the accuracy of PD detection of high voltage cables, an adaptive fuzzy C-means (FCM) clustering was proposed to identify PD signals. The adaptive threshold pulse extraction algorithm based on fixed interval width was employed. The threshold was changed adaptively to extract the effective PD pulse waveform according to the change of the background noise and the degree of PD. Then PD pulse features were analyzed in time and frequency domain by employing the equivalent time frequency analysis method. The adaptive fuzzy clustering algorithm was used to classify the signals. The phase distribution concentration of all kinds of pulse signal was calculated. The phase standard deviation of various types of pulse was taken as the index that measures the concentration of density to distinguish between PD and interference signals. Results show that the adaptive FCM clustering algorithm, compared with the traditional method, can not only identify the PD signal accurately, but also be conscious of the PD category. The PD recognition method proposed in this paper has strong applicability and high accuracy, which is particularly suitable for application in the field of engineering.

The physical properties of water lead to attenuation of light that travels through the water channel. The attenuation is dependent on the color spectrum wavelength, that results in low contrast and color cast in image acquisition. Several methods have been proposed to handle these problems, such as Linear Stretching, Histogram Equalization (HE) and their variants. Considering the advantages of HE and Linear Stretching, this paper presents a new Adaptive Linear Stretch method (ALS) which can efficiently improve the subjective impression of the traditional Linear Stretching and keep the computational cost low at the same time. To achieve adaptability, the adaptable threshold is deduced from the histogram of image. Performance analysis reveals that the proposed method significantly enhances the image contrast, reduces the color cast and meanwhile, keeps the computational consumption low.

Building structure and other factors lead to the performance deterioration of global postioning system (GPS) positioning systems indoors. An adaptive model for Bluetooth-based indoor positioning is proposed in this paper, targeting at the complex indoor environment, to improve the performance of Bluetooth-oriented indoor positioning systems. More accurate Received Signal Strength Indicator (RSSI) calibration which is optimized via Gaussian filtering, together with the environment-dependent attenuation coefficient optimization, results in a more precise hybrid model in the complicated indoor environment. Experiment results show that the difference between the estimated results and the measured samples is less than 0.25m as the target node and reference node is less than 1.5m far from each other. As the distance increases to more than 1.5m, the relative difference between the estimated values and the measured ones decreases to 7.8% at most, satisfying the requirements for indoor positioning applications.

The recognition rate of computer vision algorithms is highly dependent on the image quality. To enhance the visual quality of the images captured under high-dynamic range (HDR) scenes, we propose an efficient and adaptive tone mapping algorithm based on guided image filter (GIF). The HDR image is compressed adaptively according to its average luminance. Then we decompose it into a base layer and a detail layer using the guided image filter. We improve the base layer and enhance the detail layer simultaneously, and combine the two layers to get the final low-dynamic range (LDR) image. Since the parameters are linked with image statistics, they adaptively fit to various kinds of images. The objective evaluation results on HDR image sets demonstrate the superiority of our proposed algorithm. Meanwhile, the result of our algorithm can reduce the halo artifacts and preserve more detail by subjective observation.

Adversarial training is by far one of the most effective methods to improve the robustness of deep neural networks against adversarial examples. However, the trade-off between robustness and accuracy is still a challenge in adversarial training. Previous methods used adversarial examples with a fixed perturbation budget or specific perturbation budgets for each example, which is inefficient in improving the trade-off and lacks the ability to control the trade-off flexibly. In this paper, we show that the largest element of logit, $z_{\mathrm{\text{max}}}$, can roughly represent the minimum distance between an example and its neighboring decision boundary. Thus, we propose group adaptive adversarial training (GAAT) that divides the training dataset into several groups based on $z_{\mathrm{\text{max}}}$ and develops a binary search algorithm to determine the group perturbation budgets for each group. Using the group perturbation budgets to perform adversarial training can fine-tune the trade-off between robustness and accuracy. Extensive experiments conducted on CIFAR-10 and ImageNet-30 show that our GAAT can achieve a more perfect trade-off than TRADES, MMA, and MART.

Aiming at the situation that small unmanned surface vehicle (USV) encounters unknown disturbance during low speed sailing, a course controller with finite time stability is designed. To solve this problem, we construct an undisturbed ideal navigation model which simply meets the stability requirements, and constructs an adaptive sliding mode surface. The control under finite time approach law is also introduced. The model under perturbation can land on the sliding mode surface in finite time and then synchronize with the ideal navigation model. The adaptive control was applied in the implementation of power control for the thruster structure, so as to ensure the tracking of the desired course within the finite time, and satisfy the needs for the stable system performance. Lyapunov direct method is used to strictly prove that the designed controller can ensure the system which converges to the steady state value in a given time period. Simulation results show that the designed adaptive finite-time controller can ensure the stable course tracking of the USV with thruster structure at low speed, and meet the requirements of the course robustness of the USV under dynamic conditions.

A detailed controller design for indirect constant power regulation of high intensity discharge (HID) electronic ballast and the experimental implementation of an adaptive passivity-based constant power (PBC) controller has been extensively studied through simulation and the findings are reported in this paper. An indirect method to regulate the inductor current to ensure the lamp constant output power is proposed in order to overcome the difficulties in measuring the output power of the HID ballast. The controller is derived using passivity theory which guarantees global stability and asymptotic convergence of all state errors. The simulation and controller design are based on an average model using the Euler–Lagrange (EL) equations of the system. As the lamp resistance will inevitably change with ageing, an adaptive method is used to compensate for the lamp performance as the lamp ages. This equips the controller the power to adapt to load variations and there is no need to tune the design parameters values each time manually. Another interconnection and damping assignment (IDA)-PBC method is analyzed and simulation results are provided for comparison. Computer simulation and hardware implementation are carried out to verify the system model and to demonstrate the controller is robust.

A novel near-threshold voltage startup monolithic boost converter is presented in this paper using an adaptive sleeping time control (ASTC) scheme for low-power applications. The proposed ASTC scheme can promote the power efficiency of the current-mode boost converter under light load by automatically adjusting the sleep time of the converter, and the converter's quiescent current drops down to 4$\mu$A during the sleeping period. In addition, a new soft-start method is introduced to make the boost converter start up with a near-threshold input voltage. The proposed boost converter was fabricated in a standard 0.18$\mu$m CMOS process and occupies a small chip area of 0.50$\mathrm{\text{mm}}\times 1.22$mm. Experimental results show that the boost converter achieves the minimum 0.5-V startup voltage when the output voltage is set to 1.8V. After startup, the input voltage range can be expanded from 0.3V to 1.5V with a switching frequency of 1MHz. In addition, a peak efficiency of 94% and a minimum efficiency of 81% are measured at the 1.5-V input voltage as the load current ranges from 0.1mA to 100mA.

Recently, deep neural networks have achieved remarkable progress in class balancing instance segmentation. However, most applications in the real world have a long-tailed distribution, i.e., limited training examples in the majority of classes. The long-tailed challenge leads to a catastrophic drop in instance segmentation because the gradient of the head classes suppresses the gradient of the tail classes, leading to a bias towards the major classes. We propose LiCAM, a novel framework for long-tailed segmentation. It features an adaptive loss function named Moac Loss, which is adjustable during the training according to the monitored classification accuracy. LiCAM also cooperates with an oversampling technique named RFS, which alleviates the severe imbalance between head and tail classes. We conducted extensive experiments on the LVIS v1 dataset to evaluate LiCAM. With a coherent end-to-end training pipeline, LiCAM significantly outperforms other baselines.

This paper presents in-depth research and analysis of the optimization method of tourist traffic routes using the adaptive cuckoo algorithm. The traditional cuckoo algorithm has the disadvantages of slow convergence speed and easy falling into local extremes. This paper proposes the iterative adaptive CHSACS algorithm based on iteration. Based on the original CS algorithm, a dynamic adaptive step control amount and a segmented weighted position update formula are introduced to give a class of improved CS algorithms to coordinate the problem of local search and global search of the CS algorithm and speed up the convergence speed at the later stage. To address the problem that the out-of-bounds nests interfere with the convergence of the algorithm, a memory strategy is introduced to relocate the out-of-bounds nests in the search space to improve the stability of the algorithm. Experiments are conducted on the iterative adaptive-based conductive CHSACS algorithm with test function sets. Compared with the original CS algorithm, and ACO algorithm, the CHSACS algorithm has faster convergence, higher search accuracy and a better ability to avoid local optima when dealing with continuous function optimization problems. For the dynamic travel path problem of travel time optimization, the road travel time and sightseeing time of different periods are predicted based on historical data, and the most traveled distance of travel time between two sights is dynamically searched by using the search mechanism based on time windows. The iterative adaptive CHSACS algorithm is used to continuously find the tourist traffic route with the shortest travel time travel path from all combinations of tour sequences. This paper combines the characteristics of bus vehicle scheduling itself, takes into account the interests of both bus companies and tourists, establishes a bus vehicle scheduling model with the departure interval as the independent variable, introduces this hybrid cuckoo algorithm into bus scheduling, verifies the scientific and feasibility of the algorithm through examples, and provides a new idea for solving bus scheduling optimization problems.

We consider the coupling of two nonidentical dynamical systems using an adaptive feedback linearization controller to achieve partial synchronization between the two systems. In addition we consider the case where an additional feedback signal exists between the two systems, which leads to bidirectional coupling. We demonstrate the stability of the adaptive controller, and use the example of coupling a Chua system with a Lorenz system, both exhibiting chaotic motion, as an example of the coupling technique. A feedback linearization controller is used to show the difference between unidirectional and bidirectional coupling. We observe that the adaptive controller converges to the feedback linearization controller in the steady state for the Chua–Lorenz example. Finally we comment on how this type of partial synchronization technique can be applied to modeling systems of coupled nonlinear subsystems. We show how such modeling can be achieved where the dynamics of one system is known only via experimental time series measurements.