38 results about "Exponential form" patented technology

The invention discloses an Ad Hoc network MAC (Media Access Control) protocol realizing method based on STDMA (Self Organizing Time Division Multiple Access), solving the problem that a current Ad Hoc network MAC protocol realizing method has poor applicability. The technical scheme of the invention is as follows: a complete data link layer protocol frame is divided into three parts, namely a control frame, a broadcast frame and an information frame, wherein the control frame is in one-to-one correspondence to nodes and is used for applying an information timeslot when transmitting data. The broadcast frame is used for processing application conflict. The information frame is used for distributing the number of reference timeslots according to an exponential form of one 2n, two 2n-1, four 2n-2, etc., and distributing a timeslot number to be suitable for businesses with different QoS (Quality of System) requirements. The timeslots are organized in a massive manner so as to meet the QoS requirements of different businesses. According to the Ad Hoc network MAC protocol realizing method based on the STDMA, the topology suitability of the Ad Hoc network is improved, the characteristics of low expense and high time slot utilization rate are realized, and great support can be provided for sound and video communication.

The invention discloses a device for realizing conversion between space waves and artificial terahertz surface plasmon polariton (SPP) waves. The device comprises a medium substrate and two metal sheets fixedly connected to media to form a mirror-image relation. The metal sheets are provided with grooves arranged periodically. The depths of the grooves gradually increase from the two ends to the middle parts of the metal sheets and keep constant after being increased to a specific value. The widths of the grooves are equal. A gap is formed between one end of one metal sheet and one end of the other metal sheet. The gap is widened from the rear end to the front end in an exponential form until the width of the gap exceeds the width of each metal sheet. The device can realize effective conversion between the terahertz space waves and the SPP waves within a wide frequency range and a large angle range on the basis of novel artificial electromagnetic materials, and has the advantages of being small in size, low in weight, capable of being bent and the like, thereby having good application prospect in terahertz functional devices.

The invention provides an elevation correction method for critical coronal voltage of hardware fittings of high-elevation transmission lines. The elevation correction method is based on data of corona discharge characteristic of hardware fittings at simulated elevation in tests carried out in an ultra-high-voltage alternating-current test base environment and climate laboratory, and the correction method realizes exponential-form fitting, and can be applied to elevation correction calculation for design and optimal research for the high-elevation 330-kV alternating-current transmission lines. The elevation correction method has the advantages that calculation for elevation correction factors for the hardware fittings of the 330-kV alternating-current transmission lines in high-elevation areas is provided, equipment size and design basis meeting actual running requirements are provided, and simultaneously, fundamental data and reference for optimization and improvement operation of the hardware fittings are provided.

The invention provides a fundus image lesion degree identification and visualization system based on a convolutional neural network. The system is characterized by comprising a data inter-class exponential form sampling module in a training stage, a uniform inter-sample sampling module in the same class in the training stage, an image preprocessing module in the training stage, an image preprocessing module in a test stage and a fundus image prediction module in the test stage. The sampling method provided by the invention not only is simple to implement, but also is much faster in sampling speed. According to the neural network structure provided by the invention, a detection result is better than that of an existing neural network structure, and a discriminative region can be visualizedunder the condition that only an image-level label is used.

The invention discloses a sparse code multiple access detection method with low complexity and belongs to the field of signal detection of a wireless communication system. Classification is carried out by utilizing channel coefficient module values of the adjacent edges, namely the edge and the edge with the larger module value in the adjacent edges of the edge are selected to be in the same class to participate in information updating; other edges with the smaller module values are classified to the same class and are approximate to noise by utilizing a Gaussian approximation principle. According to the classification, an original sparse factor graph becomes a dynamic and sparser factor graph. Meanwhile, each iteration process utilizes mean and variance of the approximate noise in the last iteration process to carry out information feedback in order to make up for the information loss caused by the edge which does not participate in the information updating process. In comparison with the computation complexity in the background technology, the computation complexity is reduced in an exponential form under the condition of little BER performance loss. The result proves that the computation complexity and the BER performance are well balanced.

The invention relates to a modeling method for a GIS disconnecting switch high frequency voltaic arc resistance model, the method holds that a high frequency voltaic arc consists of three stages, namely pre-breakdown, arcing and arc blowout; voltaic arc resistance decreases according to an exponential form in the pre-breakdown stage, is a stable value in the arcing stage, and increases according to the exponential law when entering the arc blowout stage. By SF6 gap high frequency arcing tests, the effect law of a stable arcing resistance, a current peak, air pressure and a voltaic arc length on the voltaic arc resistance is obtained so as to build a steady state arcing resistance math model; by GIS disconnecting switch hot-line operation tests, the time varying law and parameter of the voltaic arc resistance are obtained so as to build a double exponential time varying math model of the GIS disconnecting switch voltage arc resistance; and by carrying out emulation calculation on very fast transient overvoltage generated by GIS disconnecting switch hot-line operation, in comparison with a GIS disconnecting switch test result, the effectiveness of the GIS disconnecting switch high frequency voltaic arc resistance match model is verified, and the scientific nature and accuracy degree of the very fast transient overvoltage are obviously improved.

The invention discloses a hydrocylinder-type anti-rolling system used for a crane, which is structurally characterized by comprising a hanger, a hoisting drum and a trolley frame, wherein the hoisting drum is connected with the hanger after passing through a fixed pulley by a hoisting steel wire rope and is connected with an anti-rolling drum arranged on the trolley frame by a chain transmission device; the anti-rolling drum is connected with a fixed pulley set arranged on the trolley frame by an anti-rolling steel wire rope (5. a. b. c. d); then, the anti-rolling steel wire rope (5. a. b. c. d) is crosswise connected with the fixed pulley set arranged on the hanger after passing through the fixed pulley set; finally, the anti-rolling steel wire rope (5. a. b. c. d) is connected with a piston rod of an anti-rolling oil cylinder (2. A. B. C. D) after passing through the fixed pulley; and a hydraulic power unit is connected with the anti-rolling oil cylinder (2. A. B. C. D) by an oil channel pipe. The invention has the advantages that the system has simple structure, lower manufacture cost, easy realization, good anti-rolling effect and the like; the hoisting swing angle is attenuated according to an exponential form; when the structural parameters of the system are determined, the hydrocylinder-type anti-rolling system has a structural parameter matching hoisting weight with the best anti-rolling effect (the shortest attenuation time).

The invention provides a micro-flow detector based on an SERS (Surface Enhanced Raman Scattering) principle and a preparation method thereof. The micro-flow detector comprises at least one micro-flow channel structure, and each micro-flow channel structure comprises a base and a metal optical grating, wherein the base comprises an upper surface and a groove; the groove is recessed down from the upper surface; a detection area is formed in the groove; the metal optical grating is formed in the detection area of the groove and used for defining a plurality of micro-flow channels of the groove; substances to be detected flow through the micro-flow channels. The metal optical grating is integrated in the micro-flow structure, so that the micro-flow detector based on the SERS principle is used for enhancing raman signals of molecules of the substances to be detected in an exponential form, and the increase amplitude is further higher than the sensitivity of detection with a metal optical grating based on the SERS principle, so that the micro-flow detector based on the SERS principle is more suitable for detection on molecule signals with ultra-low concentration. In addition, the micro-flow detector is good in controllability and is small in size and convenient to carry.

The invention discloses a new method for applying the boron neutron lifetime well logging technology to detect the quality of pile foundation post pressure grouting. The new method comprises the following steps that firstly, boric acid used for capturing thermal neutrons is added in cement grout; secondly, an underground neutron generator and a gamma ray detector are placed in the bottom of a detection pipe reserved in a pile foundation, the underground neutron generator starts to repeatedly transmit high-energy pulse neutrons, the high-energy pulse neutrons are gradually elevated upwards, the gamma ray detector synchronously records the relative change of the number of thermal neutrons in media, the number of the thermal neutrons is attenuated in the exponential form at the measuring stage, and a base line is measured; fourthly, grouting is conducted; and fifthly, after grouting, the detecting detection manner is the same as the detection manner in the third step, and a boron water curve is measured. According to the new method, the base line measured before grouting and the boron water curve measured after grouting are compared, the upward movement height and the distribution position of the post-pressure grout can be visually recognized, and accordingly the distribution situation of the post-pressure grout in a pile end soil layer and a pile periphery soil body is judged.

The invention discloses a laminated board structural damage simulating method based on an elastobrittle damage mode, and relates to the technical field of analysis of composite material laminated boards. The laminated board structural damage simulating method based on the elastobrittle damage mode comprises the following steps: step one, dividing a laminated board into a plurality of entity unit bodies in the thickness direction, taking each unit body as an analyzed object, and describing damage features of a single layer of composite material by symmetric second-order damage tensor; introducing a damage mode of a material integration point into numerical calculation, introducing an equivalent displacement concept, and taking a non-linear evolution law in an exponential form as a damage evolution model of the material integration point; calculating a damage variable evolution law according to the equivalent displacement concept and the non-linear evolution law which are introduced in the step two; step four, calculating damage rigidity of a single layer of unit body; and step five, constructing a constitutive relation between the laminated board and a damaged fiber reinforced composite material. The laminated board structural damage simulating method based on the elastobrittle damage mode has the advantage that dependence to tests and experiences of researchers is avoided.

The invention provides a rheological model suitable for medium-and-long-term settlement deformation behaviors of a rockfill dam. The model aims to solve the problems that existing three-parameter, six-parameter and seven-parameter index rheological models each only have one exponential form for rheological convergence rate adjustment of the rockfill dam, and settlement change rules of the rockfill dam in the early stage and late stage of water storage cannot be reasonably reflected easily; a nine-parameter power function rheological model can well reflect the settlement change rules of the rockfill dam, but mathematical operation is complex. The combination index rheological model suitable for the medium-and-long-term settlement deformation behaviors of the rockfill dam is provided, and a finite element computational formula of the combination index rheological model is deduced. The rheological model can well reflect rheological deformation rules of the rockfill dam, and mathematical operation is convenient.

The invention relates to a distributed pulse consistency control method and system for a time-varying and time-lag multi-agent, and the method comprises the following steps: constructing a time-varying and time-lag nonlinear multi-intelligent model containing a plurality of independent agents, and obtaining a target state model based on the time-varying and time-lag nonlinear multi-intelligent model; constructing a distributed pulse containment controller based on a pulse containment control method; defining an error vector and establishing a corresponding error system model according to the time-varying time-delay nonlinear multi-intelligent model of the intelligent agent and the distributed pulse containment controller; constructing a Lyapunov function based on the error system model and proving that a derivative value of the Lyapunov function has an upper bound; solving the relation of Lyapunov function values at the pulse generation time; and performing upper bound analysis by using a pulse comparison principle and a parameter variation method to obtain that the non-linear multi-agent system with time-varying and time-lag can realize consistency. The consistency error of the system is converged to zero in an exponential form, and the accuracy is high.

An intelligent selection and representation method for significant digits of a significant figure operational result includes the steps: firstly, searching each character in a numeric expression from left to right, acquiring adding and subtracting items with '+' or '-' serving as a breakpoint, acquiring multiplying and dividing items in each adding and subtracting item with '*' or '/' serving as a breakpoint, and acquiring power base items of each multiplying and dividing item with '^' serving as a breakpoint; secondly, performing trigonometric function and logarithm operation, selecting the significant digits of the operational result, and utilizing a selection result to substitute for an original trigonometric function or logarithm function expression; thirdly, counting significant digits of multiplying and dividing data and power base data in each adding and subtracting item, wherein the minimum digits are the significant digits of a result of the data after multiplication, division and power operation, and the operational expression only including addition and subtraction is obtained after multiplication, division and power operation; and fourthly, unifying the exponential form of scientific notation of all the adding and subtracting data, counting error digits of all the data, and obtaining error digits of significant figures. The intelligent selection and representation method adopts intelligent selection, is effectively applicable to occasions for processing massive data and with random digits, and is high in applicability.

The present invention relates to a method for applying exponential form brightness transition in image splicing syncretizing of projector. The method comprises the following steps: overlapping the same parts; gradually transiting the brightness of same parts; and completing the syncretizing splicing. Compared with the prior art, the method of the invention has the advantages that the method is more satisfied for the visual effect of human eye and hardly strip effect exists in brightness transition; and adjustment is more convenient as only two parameters exist.

The invention discloses a low voltage low power consumption amplifier based on log domain non-linear transmission function. The amplifier comprises a compression module and a non-linear conversion module; two output ends of the compression module are respectively connected with two input ends of the non-linear conversion module; the compression module comprises a PMOS pair and a NMOS pair; the PMOS pair is used for converting, in exponential form, a first differential input signal into a first voltage signal; the NMOS pair is used for converting, in exponential form, a second differential input signal into a second voltage signal; the non-linear conversion module comprises a mirror current source and conversion submodule; the mirror current source is used for generating current signals based on the inputted first voltage signal and second voltage signal; the conversion submodule is used for converting the current signals into voltage signals based on a non-linear conversion function. The inputted differential signals are converted into the voltage signals in exponential form, thereby finishing the compression conversion, executing the non-linear function, making the output expansion transistor work in a saturated area, increasing the output amplitude, and synchronously increasing the linearity and the common mode rejection ratio.

The invention relates to an artificial intelligence method and device for judging the slope collapse rockfall distance during an earthquake. According to the method, firstly, sample data are collected, then, and a DWKNN algorithm is used for predicting the rockfall distance through the collected sample data, wherein the weight in the DWKNN algorithm is in an exponential form. According to the scheme of the invention, the method can achieve the higher robustness for the slope collapse rockfall distance, and is suitable for the prediction of a small sample. Moreover, according to the embodiment of the invention, the weight calculation method in the KNN algorithm step is improved into the exponential form, so that the accuracy of the prediction result can be improved to a great extent.

The invention provides a crowd evacuation method in a low-visibility environment. The crowd evacuation method comprises the following steps: firstly, establishing an initial model of social force by virtue of a Newton second law; secondly, defining an acting force in an exponential form, representing the attraction of the leader to a common individual within the influence range, and selecting theexpected speed direction of the individual according to the familiarity degree of the individual to the surrounding environment; and finally, adding panic factors and visual influence factors, and thevisibility range, and studying the number and positions of leaders and the influence of the panic factors on the crowd evacuation efficiency. The model can effectively improve the evacuation efficiency and reduce crowd congestion and trampling phenomena, and has practical significance.

The invention discloses a pulse waveform transformation method which is used for trapezoidal forming of non-standard negative exponential nuclear pulse signals. The method comprises: firstly, expressing a non-standard negative exponential nuclear pulse signal with a slowed rising edge as a double-exponential form; secondly, discretizing the double-exponential pulse signal, and performing Z transformation; then, carrying out Z transformation of trapezoidal pulses, solving a transfer function of trapezoidal forming, and establishing a recursion formula of trapezoidal forming, wherein related parameters in the recursion formula are obtained through a genetic algorithm. According to the method, trapezoidal forming of the nuclear pulse signal with the rising edge becoming slow is achieved to acertain extent, the symmetry of trapezoidal pulses is improved, accurate acquisition of pulse amplitudes is facilitated, improvement of the system energy resolution and increase of the counting rate are facilitated, and the method has great significance in improving the accuracy of radioactivity measurement.

The invention relates to a time-varying weighting algorithm based on an exponential form, and aims to solve the problem that airborne bistatic radar clutter distance direction distribution is not stable, and clutter suppression performance of a space-time adaptive processing (STAP) technology is reduced. The airborne bistatic radar STAP algorithm based on exponential form time-varying weighting comprises the steps of establishing a universal clutter model of the airborne bistatic radar; obtaining a bistatic radar target signal S; acquiring an echo signal X (l) of airborne bistatic radar clutter plus noise as a training sample; supposing that the optimal weight vector of the airborne bistatic radar STAP is an exponential function about the distance unit l through matrix transformation, performing exponential form expansion on training sample data to obtain an expanded training sample in which alpha represents an attenuation coefficient, and mu1 and mu2 are normalization coefficients; performing autocorrelation processing on the expanded training sample to obtain an expanded clutter covariance matrix estimated value, calculating an optimal weight vector of the STAP processor as a guide vector of the expanded target signal, and acting on a to-be-detected unit to obtain clutter suppressed output;

The invention discloses a complex network construction method with a twin peak effect characteristic. The method mainly comprises the steps of determining an initial network, setting a link edge amount increase mode, determining a newly increased node amount, setting a maximum link edge amount and the like. The degree distribution of a complex network which is obtained by adopting the method presents the twin peak effect characteristic; a twin peak effect derives from a mode that a Logistic equation is used as a node selection strategy to cause node degrees to be increased in a Logistic way, which is different from a mode that a roulette or other mode is commonly used as the node selection strategy to cause the node degrees to be increased in an exponential form; in addition, the movement, the scaling and other control of two peaks of the degree distribution of the complex network are realized by setting the link edge amount increase mode and the maximum link edge amount. The complex network which is obtained by the construction method is different from a classical random network, a small-world network, a scale-free network and a self-similarity network, and the economic and social polarization phenomena in the true life can be better explained.

The present invention provides an improved method and system for a deformable activation function, and a storage device. Based on parameter learning, a deformable exponential function is introduced tooptimize the exponential form in the existing activation function. At the same time, a scale parameter is introduced, and the scale parameter is updated by the gradient during a training process. Byintroducing the idea of "parameterization" into the activation function, the performance of the activation function is improved, thereby improving the training efficiency and performance of the network; the method is suitable for the construction of any neural network, and has good scalability without platforms.