34 results about "Adaptive integration" patented technology

The invention discloses a video target tracking method which is based on multi-cue integration and particle filtration and a device thereof. The method carries out the self-adaptive integration according to three cues of color, edge and feature points to obtain target observation information, then the particle filtration technology is used for updating the target state, and the specific steps comprise: A. a target template is extracted and target parameters are initialized; B. dynamic prediction is carried out on a particle set according to a motion model; C. the particle weight is updated according to multi-cue integration information; D. the motion state of a target is updated. The video target tracking method and the device can improve the tracking effect and the tracking stability of the moving target under the complex environment and the tracked target can comprise human heads, pedestrians, cars, etc.

Particular techniques for improving the scalability of concurrent programs (e.g., lock-based applications) may be effective in some environments and for some workloads, but not others. The systems described herein may automatically choose appropriate ones of these techniques to apply when executing lock-based applications at runtime, based on observations of the application in the current environment and with the current workload. In one example, two techniques for improving lock scalability (e.g., transactional lock elision using hardware transactional memory, and optimistic software techniques) may be integrated together. A lightweight runtime library built for this purpose may adapt its approach to managing concurrency by dynamically selecting one or more of these techniques (at different times) during execution of a given application. In this Adaptive Lock Elision approach, the techniques may be selected (based on pluggable policies) at runtime to achieve good performance on different platforms and for different workloads.

The invention discloses a SVC sliding mode control method based on integration adaptive backstepping; the control method comprises the following steps: building a SVC equipment single-machine infinite bus system dynamic mathematics model, designing a controller according to an integration adaptive backstepping method, firstly designing a state variable tracking error, adding an integration item into the error item, simultaneously considering system parameter nondeterminacy, building a Liapunov function, online adaptively processing the uncertain parameters, finally adding a Terminal slide mode surface, obtaining a SVC integration adaptive backstepping stable controller, and finishing SVC sliding mode control based on integration adaptive backstepping. The method uses the adaptive integration method to eliminate errors, adds the Terminal slide mode control, so the system has stronger robustness on interference and system perturbation; the method can fast damp power concussions, can keep the machine end voltage, and can improve the power system transient stability.

A method and apparatus to estimate the channel fade (both the amplitude gain / loss and the phase rotation) to assist the receiver to detect and recover the transmitted signal employs a continuous pilot signal such as the pilot code channel or pilot symbols. The channel estimator uses the same scrambling pattern of pilot channel and coherently integrates the continuous pilot signal to yield a channel estimate. The present invention employs adaptive integration duration to yield a channel estimate. The integration duration of the pilot signal for channel estimation is adaptive and proportional to the Doppler period. The Doppler period is proportional to the inverse of Doppler frequency and is an indicator of how fast the channel changes.

The invention discloses an objective evaluation method for the full reference image quality, which comprises the steps of applying a BP neural network to image quality evaluation, designing a visual multi-channel multi-algorithm adaptive integration BP neural network image quality prediction model, inputting a distorted image into the BP neural network based on visual multi-channel evaluation results of various objective evaluation algorithms, performing supervised learning and training on the BP neural network by taking the score of a human eye subjective test result as a training objective,then predicting and outputting objective evaluation results of the various objective evaluation algorithms, and performing adaptive integration on the objective evaluation results of the various algorithms to obtain final objective evaluation for the quality of the distorted image. The method disclosed by the invention comprehensively improves the level of various indexes of the evaluation methodsuch as PSNR, SSIM and SVD, exceeds the latest evaluation methods such as visual feature perception processing and visual psychological derivation integration, and has better evaluation stability.

An adaptive integration activity management framework for on demand business process integration provides a mechanism to enable easy integration of legacy and new applications. The framework minimizes the effort need to integrate a new application into an existing business process environment such that the new activity is a “plug-in” into an action manager by implementing a standard adaptation layer. Activity integration is implemented in the principle of “on-demand” because it is invoked as required, so the communication and collaboration between partners become much more flexible.

The invention relates to an industrial rubber compound Mooney viscosity soft measurement method based on integrated immediate learning. The method is used for online prediction of Mooney viscosity in the industrial rubber compound process. According to the method, Gaussian process regression (GPR) is used as a local modeling technology; in combination with similarity disturbance and a multi-modal disturbance strategy of an input sample, the diversity of instant learning is excited; secondly, an instant learning base model meeting diversity and accuracy is constructed on the basis of evolutionary multi-objective optimization, and finally, fusion of the instant learning base model is realized and a final Mooney viscosity predicted value is obtained by introducing an adaptive integration strategy of a limited mixing mechanism. According to the method, the problems that the cost is increased and the product quality is difficult to improve due to acquisition lag of the Mooney viscosity value in the rubber mixing process are solved, and high-precision online real-time acquisition of the Mooney viscosity is realized.

Particular techniques for improving the scalability of concurrent programs (e.g., lock-based applications) may be effective in some environments and for some workloads, but not others. The systems described herein may automatically choose appropriate ones of these techniques to apply when executing lock-based applications at runtime, based on observations of the application in the current environment and with the current workload. In one example, two techniques for improving lock scalability (e.g., transactional lock elision using hardware transactional memory, and optimistic software techniques) may be integrated together. A lightweight runtime library built for this purpose may adapt its approach to managing concurrency by dynamically selecting one or more of these techniques (at different times) during execution of a given application. In this Adaptive Lock Elision approach, the techniques may be selected (based on pluggable policies) at runtime to achieve good performance on different platforms and for different workloads.

The invention discloses a target tracking method based on deep convolution feature adaptive integration. The method comprises the steps of extracting deep convolution features; calculating a kernel correlation filter; updating the integrated vector of the current frame by using an integrated vector updating formula; predicting a target position of the current frame image by using a self-adaptive integrated calculation formula; updating the deep convolution feature of the current frame by using a deep convolution feature updating formula; and taking the target center position of the current frame as the center position of the to-be-tracked target when the iteration of the video image sequence containing the to-be-tracked target is ended. By integrating the features, the defect that a tracker in the prior art cannot fully utilize information contained in different channel target features is overcome, so that the position of the target to be tracked is more accurately acquired in the target tracking process, and the accuracy and reliability of target tracking are enhanced.

The invention discloses a gas adaptive integration valve with double air sources, which is composed of a valve body (1), wherein one end of the valve body (1) is provided with a low-pressure air source air inlet (3) and a high-pressure air source air inlet (4); a self-sealing valve group (10) is installed in each of the low-pressure air source air inlet (3) and the high-pressure air source air inlet (4); a self-regulating voltage-stabilizing valve (5), an ODS (Ozone Depleting Substances) air passage selection component (6) and a main air passage selection component (7) are installed in the valve body (1); the self-regulating voltage-stabilizing valve (5), the ODS air passage selection component (6) and the main air passage selection component (7) respectively abut against the corresponding pressure regulating reed by the respective pressure regulating component; the pressure regulating reed is abutted with a regulating pulling rod (13); the regulating pulling rod (13) is provided witha convex-concave structure which causes the pressure regulating reed to act; two ends of the regulating pulling rod (13) are supported in chutes or holes on the valve body (1); and the valve body (1)is also provided with a locating slot (15) for locating an operating lever (14) on the regulating pulling rod (13). The gas adaptive integration valve with double air sources has simple structure, a manual selection switch is omitted, safety is improved, and the production cost is lowered.

The invention relates to a MIMO-SCFDE adaptive transmission scheme based on model-driven deep learning. The invention establishes an adaptive transmission model based on the MIMO‑SCFDE system. AMNet and ADNet are used to replace the signal modulation and modulation identification parts in the traditional system respectively. AMNet adopts a combined network with 2D CNN, LSTM and FC-DNN as sub-networks to form an integrated neural network model, and adjusts the modulation mode of the sending end according to the channel conditions of the receiving end, and inputs the feature information extracted from the received signal into multiple sub-networks , and realize the conversion of features and optimal modulation schemes according to the network parameters obtained through training. At the same time, the received power under different path delays is selected as the adaptive factor to realize the adaptive integration of the results of each sub-network. According to the cyclic spectrum, ADNet has the advantage of accurately detecting the signal type under low signal-to-noise ratio, and based on the complexity of the cyclic spectrogram, it completes the adaptive selection of the modulation recognition scheme. This system is more suitable for the performance requirements of the 5G communication system.

The invention provides an embedded infrared image super-frame processing method based on adaptive integration time, the purpose of which is to utilize the advantages of the infrared image super-frame processing method to improve the detection sensitivity and dynamic range of the infrared imaging system, and effectively solve the problem of traditional super-frame processing In the method, the integration time of acquiring subframes cannot be adaptively adjusted according to the radiation temperature of the target scene, so as to better guide the detector to work in the middle area of ​​the entire response range. The invention includes two stages of off-line calibration and on-line processing. The invention also provides correction measures for the image non-uniformity problem caused by the adjusted integration time of the detector. The invention is an infrared image super-frame processing method based on an embedded platform, does not need any special device in the realization process, and can realize the real-time engineering application of the method.

The invention proposes an adaptive integral backstepping control method for anti-load disturbance of PMSM used in elevators. Compare the motor speed w with the given motor speed w * By comparison, the speed error e is obtained; according to the feedback motor speed w and electromagnetic torque T e Estimated load torque τ L ;The estimated load torque τ L and the rotational speed error e are input into the adaptive integral backstepping controller for adjustment to eliminate the steady-state error of the speed so as to obtain the virtual control quantity of the q-axis of the stator current in the rotating coordinate system d-q and use the excitation current component as a reference value to input the current In the ring, and the stator current i after the coordinate transformation d Make a difference to get the d-axis stator current error e d ;According to the q-axis stator current error e q and d-axis stator current error e d Calculate the control voltage u d and u q ; will u d and u q After inverting through the park, it is input to the SVPWM pulse width modulation module to generate the pulse signal required to drive the inverter, thereby driving the motor to run. The invention improves the robustness of the electric system of the elevator, and enhances the comfort and speed of the elevator.

The invention discloses a method for obtaining electromagnetic scattering properties based on subregion adaptive integration. First, geometric modeling is performed on a radar conductor object by means of CAD software for cutting the whole object into multiple subregions and virtual surfaces among subregions are kept; secondly, all subregion surfaces are scattered with triangles and subregion grid information after dissection is output; Descartes grids are established in subregions respectively again, induced current of subregion surfaces is obtained by means of adaptive integration method and coupling among subregions; finally, far region scattered field is calculated by means of the induced current of subregion surfaces and further radar scattering cross section of the conductor object is obtained. Auxiliary point current amount is reduced on the premise that accurate calculation results are guaranteed, and memory requirement is reduced and convergence rate and simulation efficiency are increased.

A nuclear magnetic resonance gyro self-adaptive dynamic and static closed-loop control method, including gyro dynamic work and static work adaptive discrimination strategy, dynamic work closed-loop parameter adaptive adjustment method, static work closed-loop parameter adaptive adjustment method and self-adaptive integral method. The invention performs mean square statistics on the long-term input closed-loop error, and then divides the dynamic working state and the static working state of the gyroscope through the setting of the soft threshold, which has a high degree of confidence and is beneficial to the gyroscope closed-loop output in different situations. More precise closed-loop control. In addition, through the dynamic adjustment of the closed-loop coefficient, the static and dynamic response ability is improved. Finally, the present invention performs selective input control on the integral quantity of the closed-loop control, which can play the role of the integral term in the closed-loop control, and can effectively avoid the influence of long-term noise accumulation on the drift of the gyro output.

The invention discloses an infrared focal plane array non-uniformity correction method adapted to the dynamic adjustment of integration time based on a neural network, which belongs to the field of non-uniformity correction of an infrared system. The method includes: calibrating at multiple working points of the system, calculating Average gray-scale response; use mathematical regression to calculate the interpolation curve of average gray-scale response and pixel gray-scale response under different integration times; build and train a forward neural network for non-uniformity correction; use the trained neural network to perform Non-uniformity correction, the integration time can be adjusted dynamically during the correction process. The present invention overcomes the disadvantages of the previous infrared system, such as few steps of integration time, large adjustment span, and troublesome switching, makes the adjustment of integration time more delicate and flexible, expands the adaptability of the infrared system to scenes with a large dynamic temperature range, and is beneficial to infrared systems. The system performs intelligent imaging on the target.

The invention discloses an intelligent driving vehicle environment adaptive merging method under an urban environment. The method comprises steps: an initial state vector is extracted; an action variable is calculated according to a greedy strategy, a merging scene is updated while a merging action is executed, if the action variable adopts a random action, a merging gap and a merging action are selected with a uniform probability, if an intelligent method is adopted, candidate gaps comprise a front vehicle, a following vehicle and a merging vehicle, the maximum action value functions of all candidate gaps are compared, the maximum value function is selected, the gap and the action corresponding to the maximum value are picked out, and a target merging gap and an intelligent merging actionare returned; the state vector at a next moment is sensed; a reward value is calculated according to the environmental feedback information; the initial state vector, the action variable, the state vector at the next moment and the reward value are saved to a sample set, and after enough samples are obtained, evaluation and improvement are carried out according to an LSQ method; and the above steps are repeated until merging succeeds. The sample set and the learning time are lower than a Q learning algorithm, and the success rate is high.

The invention discloses a reduced matrix construction method for accelerating the iterative solution of the characteristic basis function method, which can effectively accelerate the iterative solution speed of the characteristic basis function method. Firstly, the excitation source is compressed using the singular value decomposition technique, and the compressed new excitation source is obtained, which is defined as voltage basis functions (VBFs); secondly, under the new excitation source, the characteristic basis function of each subdomain ( CBFs); Finally, VCBFs and CBFs are used as test functions and basis functions to construct a reduced matrix, and the main diagonal sub-matrixes of the obtained reduced matrix are unit matrices, and the condition number of the reduced matrix is ​​optimized, which can speed up the iterative solution of the reduced matrix equation. The invention provides a new method for the iterative solution of the eigenbasis function method. At the same time, the invention can be combined with algorithms such as multi-layer fast multipole method, adaptive integration method, pre-correction-fast Fourier transform method, etc. Efficiency of function method in analyzing electromagnetic scattering characteristics of electrically large targets.