58results about How to "Small step size" patented technology

An implantable pulse generator includes a current steering capability that allows a clinician or patient to quickly determine a desired electrode stimulation pattern, including which electrodes of a group of electrodes within an electrode array should receive a stimulation current, including the amplitude, width and pulse repetition rate of such current. Movement of the selected group of electrodes is facilitated through the use of remotely generated directional signals, generated by a pointing device, such as a joystick. As movement of the selected group of electrodes occurs, current redistribution amongst the various electrode contacts takes place. The redistribution of stimulus amplitudes utilizes re-normalization of amplitudes so that the perceptual level remains fairly constant. This prevents the resulting paresthesia from falling below the perceptual threshold or above the comfort threshold.

A system and method for locating regions in a target image that match a template image with respect to color and pattern information. The method may comprise performing a first-pass search using color information obtained in a color characterization analysis of the template image in order to find a plurality of color match candidate locations. For each color match candidate location, a region proximal to the location may then be searched in detail, based on pattern information obtained in a pattern analysis of the template image.

The invention discloses a video-based detection and recognition system and a method of vehicles. The system comprises a target characteristic database module, a moving target detection module and a vehicle recognition module. A vehicle sample target characteristic database is built to store data by virtue of the target characteristic database module. After a moving target zone is detected by the moving target detection module, the vehicle recognition module builds a search window and recognizes vehicles. The system and the method resolve the technical problems of poor vehicle distinguishing capability and low distinguishing accuracy of the video-based vehicle detection under the condition of obvious background movement, vehicle adhesion, vehicle occlusion and the like in the prior art. The video-based detection and recognition system and the method of the vehicles have the advantages of excellent vehicle distinguishing capability and high accuracy.

The invention discloses a detection method for a variable step length LMS (Least Mean Square) harmonic current based on versiera. The method comprises the steps of: sampling, estimating an active current of a fundamental wave, estimating a general harmonic current, updating a weight coefficient, updating the step length and iterating. According to the orthogonal characteristics of a trigonometric function, the current error output by a filter and an instantaneous cross-correlation time mean value are subtracted from a load current to eliminate interference of the harmonic current and the active current of the fundamental wave on update of the step length. Furthermore, update of the step length is controlled by an improved versiera function. Meanwhile, in the updating process of the weight coefficient, variation of the weight coefficient is imported so that the steady state error is very small and the calculated amount is smaller. The detection method for the variable step length LMS harmonic current based on versiera provided by the invention is suitable for occasion where the load current has saltus and value of the optimal parameter provided by the invention is less dependent on the load current.

A system for displaying an image is disclosed. A user input subsystem (1) is arranged for enabling a user to indicate at least one point of a region of interest of an image (5). A zoom subsystem (2) is arranged for performing a zoom-in operation by filling a viewport with successively smaller portions of the image (5), wherein the successively smaller portions are selected such that the region of interest is shown at a decreasing distance from a center of the viewport. The user input subsystem (1) is arranged for enabling the user to control the zoom-in operation by indicating, after the viewport has been filled with one or more of the successively smaller portions of the image, whether further zooming is desired for the, already indicated, at least one point of a region of interest.

The invention provides a phase locked loop having a modulator which is based on a ΣΔ fractional N phase locked loop. In the forward path of the PLL, the output of the oscillator has an additional frequency divider which provides the output frequency of the PLL in a plurality of different phases. A multiplexer which is connected upstream of the multimodulus divider in the PLL's feedback path and which is actuated by the ΣΔ modulator, like the divider, selects the respective desired phase. This allows the minimum step size of the division factors to be reduced to values of less than 1 relative to the output frequency, which significantly reduces the quantization noise. The PLL bandwidth may therefore advantageously be the same size as the modulation bandwidth.

The invention discloses a mechanical sealing device for magnetofluid lubricating and an adaptive sealing control method, and belongs to the technical field of sealing. The mechanical seal is characterized by comprising a movable ring and a fixed ring, wherein a micron-level deep spiral groove is formed in one end face of the movable ring, and the fixed ring is matched with the movable ring. Under a working state, magnetofluid is injected between the end faces of the movable ring and the fixed ring through a through hole, which is close to the outer diameter, of the fixed ring; under the action of the spiral groove, the magnetofluid flows to the inner diameter direction from the outer diameter of a sealing end face, and a hydrodynamic effect is generated; pressure on a magnetofluid film is increased, and the two end faces are pushed away, so that lubrication on the fluid on the end faces is realized. An adaptive sealing property control system judges whether the sealing property of the mechanical seal meets the requirement by comparing the pressure of a liquid film with the pressure of a sealed medium; if the rotating speed of the movable ring or the pressure of the sealed medium is changed, the system can adjust the voltage of a magnetic field generator according to a preset strategy to adjust the pressure of the liquid film and correspondingly adjust the input flow of ferrofluid, so that the sealing effect of the mechanical seal is guaranteed.

The invention discloses a gesture recognition model training method, a gesture recognition method and a gesture recognition system. The training method comprises the steps of collecting a gesture picture sample in a plurality of scenes, and performing random cutting on the gesture sample to obtain a new gesture sample, and taking the gesture picture sample and the new gesture sample as a sample set; constructing a light YOLO network, and training the Light YOLO network by utilizing a sample set, and a first light YOLO network is obtained; a selective-dropout layer is added behind each layer ofconvolution layer of a first light YOLO network, and a second light YOLO network is obtained. Training and convergence are carried out on the second light YOLO network by means of the sample set, andthen channel cutting is conducted so as to obtain a gesture recognition model. According to the invention, the detection performance of the network with smaller resolution is improved; the gesture recognition method disclosed by the invention is high in accuracy and good in real-time performance; meanwhile, the system can directly obtain the recognition result from the picture, and end-to-end optimization can be carried out.

The present invention provides a method and apparatus for calibrating a black level in an imager to reduce flicker noise. A first and a second range is set. The first range corresponds to a range that is larger than a noise level at a highest PGA gain. The second range corresponds to a range that is smaller than a level which reduces an ADC dynamic range too much due to a large black level. The ranges may be adjusted to changes in an imager in real time. When the black level is within the second range a determination is made as to whether the black level has been calibrated before. When it has, the DAC output is held constant. Otherwise the DAC code is adjusted such that the black level is moved toward the first range. A small step size is used in adjusting the DAC code in order to reduce flicker. Step sizes may be adjusted according to the black level in relation to the ranges. A small step size is used when the black level is within the second range. A large step size is used when the black level is beyond the second range.

The invention discloses a method for determining the optimal offset angle of a non-sun-oriented solar cell array, and the method mainly comprises the steps: obtaining a calculation input condition, such as an orbit height H and an orbit incident angle range; calculating a current offset angle and a ground shadow field angle under a current orbit incident angle; calculating the output energy of thesolar cell array under the current bias angle and the current orbit incident angle by taking the earth shadow field angle as input; and traversing all bias angles and all orbit incident angles, and evaluating the output energy of the solar cell array to obtain an optimal bias angle. The method is used for selecting the optimal bias installation angle of the satellite solar cell array to obtain the comprehensively optimal solar cell array output energy, traversing all bias angles and the solar cell array output energy under all orbit incident angles, and evaluating the calculation result of the solar cell array output energy to realize the selection of the optimal bias angle.

A method and apparatus for emulating stream clock signal in asynchronous data transmission. The inventive subject matter proposes a system consisting of a transmitter module, a receiver module, and a link or network in between. A scheme to generate the emulated stream clock across a wide frequency range is also proposed with the property of controllable deviation from the original stream frequency to meet jitter requirement and fast frequency convergence (minimal number of converging steps). The scheme includes an optional first step to derive a frequency estimation of the stream clock and a second step of continuous adjusting the emulated clock frequency to keep the average frequency equals that of the original stream clock.

The invention relates to an independent solar power generation controller, and the controller is used for a solar power generation system, and controls a solar power supply device to work in a state of maximum power, so as to outputs a corresponding control signal and an output voltage to a storage battery according to the state of a solar cell panel. The controller comprises a main circuit, a sampling circuit, a microprocessing control circuit, an MOS tube drive circuit, and a power circuit. The microprocessing control circuit is connected with a main circuit through the MOS tube drive circuit, and the MOS tube drive circuit comprises a backflow prevention MOS tube. When the input end voltage of the microprocessing control circuit is lower than an output end voltage, the MOS tube drive circuit is controlled to open the backflow prevention MOS tube NMOS-in for backflow prevention. When the input end voltage of the microprocessing control circuit is greater than the output end voltage, the MOS tube drive circuit is connected, and the microprocessing control circuit is enabled to control the main circuit for charging the storage battery.

The invention discloses a small navigation receiver applicable to a low-earth-orbit satellite. The small navigation receiver has the advantages that integral design integrating multiple modules such as a channel module and a passage module; the FPGA+DSP architecture of a traditional navigation receiver is broken through, core component's dependence on import is eliminated, a radio frequency chip and base band SoC chip group is applied in an in-orbit manner, and satellite borne navigation production core function chips and domestic design are achieved; a navigation equipment space operation environment is combined with an orbit determination combining dynamics and kinematics to provide a multi-navigation-system real-time orbit determination method, and broadcast ephemeris orbit errors adopta random walk strategy to perform parameterization separation; on the basis of the anti-error estimation principle, extended Kalman filter is used to estimate the optimal motion state of the low-earth-orbit satellite, and spatial information such as high-precision real-time positions and speed is provided for a satellite platform and load; the small navigation receiver is high in reliability, lightweight, especially applicable to small satellite environments and capable of satisfying the requirements of high-reliability, high-integrity and small-size satellite systems.

A zero-norm set membership affine projection adaptive echo cancellation method based on weight vector reuse comprises the following steps: A, remote signal sampling; B echo estimation: obtaining an output value y(n) at a current time n through an adaptive filter based on an input vector X(n) of the adaptive filter at the current time n, wherein y(n) is the estimated value of echo, and y(n)=X<T>(n)*w(n); C, echo cancellation: subtracting the output value y(n) of the adaptive filter at the current time n from a near-end signal d(n) with echo at the current time n picked up by a near-end microphone to get a residual signal e(n), wherein e(n)=d(n)-y(n), and transmitting the residual signal back to a remote end; D, filter tap weight vector updating: getting a tap weight vector w(n+1) (Please see the description for the expression) of the filter at next time n+1 using a zero-norm set membership affine projection method based on weight vector reuse; and E, letting n=n+1, and repeating the steps B, C and D until the end of the call. The method has the advantages of good cancellation effect for acoustic echo of a communication system, high convergence rate, and small steady-state error.

An electrostatic drive 150 having a plurality of mover electrodes 152 operatively secured to a mover 154, and a plurality of stator electrodes 156 operatively secured to a stator 158. The mover 154 and stator 158 are configured to move relative to each other via electrostatic force generated between the mover electrodes 152 and the stator electrodes 156. The electrostatic drive 160 includes a driver configured to place the stator electrodes 158 in any of a number of sequential voltage states, each being defined by a combination of LO and HI voltage levels at the individual stator electrodes 156. Transition from one voltage state to a sequentially adjacent voltage state produces a step size of relative movement between the mover 154 and stator 158. For each of the sequential voltage states, the driver 160 is further configured to selectively vary voltage applied at one of the stator electrodes 156 to an amount between the LO and the HI voltage levels, in order to produce a proportionally smaller step size.

The application discloses an asynchronous motor restart frequency sweep method and device. The method comprises the following steps: current amplitude closed loop control is established; when a stator is subjected to bidirectional full frequency searching operation, voltage values that are actually output are compared with N voltage comparison threshold values in a voltage comparison threshold value group from small to large in a one by one manner, a maximum voltage comparison threshold value is a product of a voltage frequency characteristic gradient and a frequency value in search processes, first N-1 voltage comparison values have maximum voltage comparison threshold values which are orderly M1 times to MN-1 times as much as of the first N-1 voltage comparison values, N is a positive integer greater than 1, and M1 to MN-1 are orderly positive numbers that orderly increase and are less than 1; when the voltage values that are actually output are greater than the voltage comparison threshold values being compared, step length of current frequency search is reduced, and otherwise the step length of current frequency search is maintained; when the voltage values that are actually output are not smaller than the maximum voltage comparison threshold values, that searched frequency is rotor synchronization frequency is determined, a frequency converter can be prevented from overflowing, and an aim of frequency sweep acceleration can be attained.

The invention discloses a satellite channel multi-mode blind equalization algorithm, and the algorithm comprises the steps: firstly carrying out the equalization of a satellite signal, generating a discrete time signal, and calculating a judgment error absolute value of the discrete time signal; secondly processing the judgment error absolute value through employing a VS-CMA algorithm when the judgment error absolute value is greater than a preset maximum judgment error threshold value; thirdly carrying out the processing of the judgment error absolute value through employing a DD-LMS algorithm when the judgment error absolute value is less than a preset minimum judgment error threshold value; finally carrying out the processing of the judgment error absolute value through employing the VS-CMA algorithm and the DD-LMS algorithm when the judgment error absolute value is greater than the preset minimum judgment error threshold value and less than the preset maximum judgment error threshold value. The method introduces a nonlinear residual error function to control the step of the CMA algorithm, so as to achieve the variable-step CMA algorithm. The method provided by the invention is smaller in residual error, is higher in convergence speed, enables intersymbol interference to be smaller, and is strong in tracking capability of time-variable channel.

The invention discloses a shrinkage variable step length sub-band echo cancellation method, and the method comprises the steps: A, signal sampling and processing: segmenting an input vector U(n) into I far-end sub-band vectors Ui(n) through a first analysis filter, enabling a near-end signal d(n) picked by a near-end microphone to be segmented into I near-end sub-band signals di(n) through a second analysis filter, carrying out the I extraction of a far-end input sub-band vector Ui(n) through an extractor, obtaining far-end input sub-band extraction vectors Ui(k), and, similarly, carrying out the I extraction of the near-end sub-band signals di(n) through the extractor, and obtaining near-end sub-band extraction signals di(K); B, filter outputting: enabling the far-end input sub-band extraction vectors Ui(k) to pass through a sub-band filter in an adaptive echo cancellation filter, obtaining output sub-band signals yi(k); C, echo cancellation; D, the updating of a weight coefficient vector: obtaining an updating step length [Mu]i(k) of the weight coefficient vector of the sub-band filter based on a denoising near-end signal square estimation quantity; E, the supposing: k is equal to (k+1), and the repeated execution of steps A, B, C and D till the conversation is ended. The method can obtain a higher rate of convergence and a lower stable error.

An exemplary embodiment of an echo cancellation circuit is provided, for use in a voice interaction device simultaneously outputting a remote signal while receiving a local signal. The local signal may comprise an echo generated from the remote signal. A first filter learns the remote signal at a first speed to generate a first coefficient set, and filters the local signal by the first coefficient set to generate a first filter output. A second filter learns the remote signal at a second speed to generate a second coefficient set, and filters the local signal by the second coefficient set to generate a second filter output. A third filter comprises a third coefficient set, canceling the echo from the local signal to generate a third filter output as an echo cancellation result. The controller updates the third coefficient set based on the first, second and third filter outputs.

A method and a device relating to a scheduling mechanism in a base station in a WCDMA system are disclosed. The mechanism enables the base station to rapidly adapt to users momentary traffic demands and to interference variations, a dynamical adjustable margin is proposed in this invention. The adjustment is based on the RoT measurement in a cell. To fully use the resource in a cell, the RoT can be targeted to be as closer to the RoTmax as possible however without exceeding the limit. The scheduler is preferably set to schedule as high a rate or as many users as possible to fill up the available RoT. The margin which is reserved for the neighboring cell load and external interference is decreased step by step as long as the RoT measurement is below a threshold under RoTmax. Whenever the RoT measurement exceeds the threshold, the margin is increased by one step.

The invention discloses a short circuit current calibration circuit of a spark test device. The circuit comprises a calibration selection switch connected to a calibration power supply and a calibration parameter selection circuit connected to the calibration selection switch. The calibration parameter selection circuit comprises N calibration resistors which are connected to contact points of the calibration selection switch respectively. A calibration switch is connected to each calibration resistor. Two ends of the calibration switch are taken as calibration current output terminals. And the N is greater than 4. Compared to the prior art, in an embodiment of the invention, more short circuit current calibration parameters can be provided so that a step length among the short circuit current calibration parameters is shortened and calibration precision is increased.

A digital event generator includes a counter configured to provide at least one count value based on a clock signal, and a comparator configured to evaluate a first portion of a first count value to detect a near occurrence of an event, in response to a detection of a near occurrence of an event, evaluate a second portion of a second count value, and provide the event signal based on the evaluation and digital event time information. A switched mode energy converter uses said digital event generator.

The invention discloses a conduction sensitivity measurement method and device, a storage medium and a comprehensive measurement terminal. The method comprises the following steps: determining original measurement parameters and parameter values of an intelligent terminal; determining initial signal intensity according to the original measurement parameters and the parameter values thereof; and taking the initial signal intensity as initial measurement intensity, testing the bit error rate of the intelligent terminal by adopting a plurality of test signal intensities so as to obtain a test signal intensity critical value, and taking the test signal intensity critical value as the conduction sensitivity. According to the scheme, the signal intensity measured for the first time is closer toa critical value, so that the measurement frequency is reduced, the measurement time consumption is reduced, and the measurement efficiency is improved.

A tape supply-side hub, a tape take-up-side hub and a magnetic tape wound onto the tape supply-side hub are included inside a mainframe case. A leader tape fixed onto the tape take-up-side hub and one end of the magnetic tape are connected via a splicing tape. The other end of the magnetic tape is wound onto the tape supply-side hub, and the magnetic tape and the tape supply-side hub are directly connected with adhesive agents. A wound tape layer wounded onto the tape supply-side hub in the fully wound state is composed of a non-recording layer on the winding leader side composed of a specified amount of the wound tape layer and a recording layer continuing to the non-recording layer. The feeding limit position of the magnetic tape in the tape supply-side hub is set to be in the vicinity of a boundary between the non-recording layer and the recording layer.

The invention discloses an inductance-capacitance type numerically controlled oscillator with high tuning precision. A resonant circuit comprises an inductor L, a capacitor C, a first control unit circuit, ..., and a nth control unit circuit, wherein the first control unit circuit is controlled by a first control bit D1 and consists of P type field effect transistors PM11, PM12, PM13 and PM14, the nth control unit circuit is controlled by a nth control bit Dn and consists of P type field effect transistors PMn1, PMn2, PMn3 and PMn4, and the inductor L, the capacitor C, the first control unit circuit, ..., and the nth control unit circuit are connected in parallel to form a control unit array; a first input end of the capacitor L is connected with a first input end of the capacitor C to be taken as a first input end (1) of the resonant circuit, and a second input end of the inductor L is connected with a second input end of the capacitor C to be taken as a second input end (2) of the resonant circuit; and each control unit is formed by the reverse parallel connection of PMOS (Positive Channel Metal Oxide Semiconductor) tube pairs with different sizes, the small capacitance difference can be obtained by the size difference of the PMOS tubes, and the fine frequency tuning step can be obtained, namely, the high frequency tuning precision is realized.

The present invention provides a method for low profile acoustic floor isolation. An isolator is partially but substantially recessed in an acoustic floor and a leveling mechanism at least partially recessed in the acoustic floor is connected to the acoustic isolator to adjust the height of the acoustic floor when the floor is in place.

The invention relates to a complex image Canny edge detection method based on double sparse decomposition. According to the method, through structured NSCT decomposition, non-stable change high-frequency component data difficult to express in an image is separated, and a matrix of the high-frequency component data is sparse. Therefore, K-SVD dictionary learning is only carried out on a high-frequency component obtained through NSCT decomposition, a specific base vector of the image can be fused into a dictionary to form complementation with NSCT decomposition, the calculation amount of the algorithm can be greatly reduced, and the efficiency of the algorithm is improved; according to the method, a double sparse dictionary learning method is adopted, and part of stably changing high and low frequency information is possibly discarded while specific feature information of an image is reserved, so that the richness of dictionary atomic data is reduced. Therefore, a K-SVD learning dictionary and a DCT dictionary are combined, the deficiency of the learning dictionary in the aspect of stably changing a base vector is made up, and the sparse representation capability of the mixed sparse dictionary is improved.

The invention relates to a pan-tilt control module for a cloud terminal. The pan-tilt control module includes a vertical pan-tilt control circuit and a horizontal pan-tilt control circuit. The motor has a small step size of 0.5 degrees and a high degree of precision.

The invention creatively combines an improved seagull algorithm with an unmanned aerial vehicle group, and discloses a multi-unmanned aerial vehicle gas leakage source positioning method based on the improved seagull algorithm. Compared with a traditional traceability method depending on a fixed monitoring station or a ground mobile robot, the method takes the unmanned aerial vehicle cluster as a traceability platform of gas leakage, and has the advantages of low cost, wide search range, high flexibility, high traceability efficiency and the like. Compared with a standard seagull algorithm, the gas leakage source positioning method based on the improved seagull algorithm has the following innovativeness: the value range of a balance operator B is reduced, the step length is greatly reduced, and the global search capability is greatly improved; the spiral radius r is set to be a fixed value, it is guaranteed that no disordered moving direction occurs in the iteration process of an individual, and the stability of the algorithm is improved; the foraging behavior is introduced, so that other individuals can explore themselves while following the optimal individual, and local optimum is avoided.

The invention discloses an anti-shock automobile safety airbag stiffness test device comprising an L-shaped operation platform (1). Two fixing blocks (2) are arranged up and down on the vertical side of the operation platform (1), and the fixing blocks (2) include an upper fixing block (3) and a lower fixing block (4). The upper fixing block (3) is provided with a smooth through hole for a screw (5) to pass through. The upper surface of the lower fixing block (4) is provided with a circular groove. The screw (5) is in threaded connection with a moving block (6), and the moving block (6) is connected with an upper extruding board (8) in the horizontal direction. The top end of the vertical side is connected with a top board (15). A telescopic rod (16) is vertically arranged on the lower surface of the top board (15). The upper surface of the upper extruding board (8) is provided with a groove (17). According to the anti-shock automobile safety airbag stiffness test device provided by the invention, the test result is real and reliable, and the upper extruding board has an anti-shock function.