74results about How to "Accuracy unchanged" patented technology

The invention relates to a method and a device for analyzing the reliability of an integrated circuit. In the analytical method, a unit circuit delayed aging stochastic analysis reference model in consideration with both negative bias temperature instability (NBTI) effect and process parameter perturbation is established, a scaling function and an equivalent aging time concept are provided to solve the delayed statistical distribution of a unit circuit under the actual work environment quickly from the reference model, and the pre-clipping process of the circuit is provided to reduce the complexity of reliable analysis. The device of the invention comprises an input unit, an output unit, a program storage unit, an external bus, a memory, a storage administration unit, an input/output bridging unit, a system bus and a processor. In the method and the device, the effect of the process parameter perturbation, the NBTI effect and the work environment of the circuit on reliability are considered simultaneously, and the complexity of the reliable analysis can be reduced effectively by utilizing the scaling function, equivalent aging time and the pre-clipping technology so as to realize the quick analysis on the reliability of super-large-scale integrated circuits in consideration with process deviation.

The invention discloses an electric operator for a general-purpose mechanical automatic transmission clutch, which comprises a servo motor, a shell, a coupling, a connector, a front end flange, a back end flange, a bearing, a hinge pin, a ball screw, a pressure sensor, a spacing sensor and a displacement sensor, wherein the front end flange fixes the shell on the clutch bracket; the motor is fixed on the shell; the output shaft of the motor is connected with the screw stem of the ball screw through the coupling; the screw stem drives a screw stem shaft sleeve through balls to drive a push rod to move; and the other end of the push rod is connected with the pressure sensor, thereby finally acting on the driving lever of the clutch, so that the driving lever drives the clutch to move. The invention has the advantages of simple structure, high transmission efficiency, accurate control and the like, is suitable for clutches with different rigidities and movement strokes, and can monitor the movement state of the clutch from different angles. The invention is mainly used in electric operators for clutches in experimental study.

The invention discloses a signal reproduction type integrated circuit board failure detection device which is composed of two portions including a hardware portion and a circuit portion. The hardware portion comprises a housing, alternating current connection wires, a power switch, a USB interface, a SD card interface, a circuit board interface to be detected, an integrated board, keys, a display screen, and a power indicator light. The circuit portion comprises a power circuit, a multi-channel signal acquisition and reproduction circuit and a master control circuit. The master control circuit comprises a single chip microcomputer, a display device, a USB interface, a storage device and the like. The multi-channel signal acquisition and reproduction circuit comprises a single chip microcomputer, an A/D conversion and D/A conversion circuit, a signal amplification circuit and the like. The device is used to acquire signals of a standard circuit board, and the acquire input signal is reproduced onto a corresponding input channel of a failure integrated circuit board, and at the same time, the input/output signal conditions of the failure integrated circuit board are acquired, so that from the input/output signal abnormal conditions, the position of a failure can be judged without knowing the internal structure of the integrated circuit board, so the maintenance speed can be greatly quickened, and the maintenance difficulty can be simplified.

A device and method for measuring ice thickness on a first surface of a construction element. The device includes a sensor arrangement arranged in connection with the construction element and arranged to provide signals related to an impedance. A processing unit is coupled to the sensor arrangement and arranged to determine the presence of ice based on the provided impedance signals. A reference sensor arrangement is arranged to generate reference impedance measurement signals. The processing unit is further arranged to determine the ice thickness based on determinations of a relation between the impedance signals determined by the sensor arrangement and the reference impedance values generated by the reference sensor arrangement.

The invention discloses a behavior recognition method based on continuous frame images. The behavior recognition method comprises the following steps: utilizing a target detection algorithm YOLO to carry out regression on a BBOX frame of a pedestrian; cutting the regressed BBOX frame into a picture according to coordinates, adjusting the size of the picture, and storing the picture by adopting a list; and sending the pictures to a trained behavior recognition network to obtain a pedestrian recognition result. Aiming at the problem of insufficient expression ability caused by only using a target detection algorithm to detect pedestrians, the invention provides an identification method combining pedestrian detection with behavior identification. A behavior recognition network based on a 3D-RDBC module is provided, and the accuracy is basically unchanged while the parameter quantity is reduced.

The invention relates to a high resolution reconfigurable optical analog-to-digital conversion method and device, in particular to an optical analog-to-digital conversion device with the sampling rate capable of being reconfigured according to any power of 2. The device comprises a pulse laser, a reconfigurable repetition frequency multiplication module, an electro-optic signal sampling module and a reconfigurable signal quantization and reconstruction module. According to the device, a high-stability passively mode-locked pulse laser is adopted as a system optical clock, and a heavy frequency multiplication module based on a multi-channel multiplexing technology is used, so that high speed optical clock signals can be effectively acquired. A switching device is used for changing channel numbers of the heavy frequency multiplication module and a back-end parallel photoelectric conversion and quantization module, so that multiplication of the light sampling rate is realized according to reconfigurability of power of 2, time delay and range matching are not required to be readjusted after marked, and the switching speed of the system in working states is guaranteed by quick response of the switching device. Besides, high speed optical sampling of electric signals is realized by the aid of an ultra broadband electro-optic modulator (or an electro-optic sampling gate), and finally, high speed optical analog-to-digital conversion is realized by the aid of parallel photoelectric conversion, parallel electric quantization and parallel data processing.

The invention discloses a wind speed sensor and a detection method for detecting cotton heterosexual fibers. A central bearing is connected to a wind direction cup device through sliding balls, and several wind direction cups are horizontally distributed on the outside of the wind direction cup device along the circumferential direction; the middle section of the center bearing is coaxial Fixed gear, the top of each tooth of the gear is fixedly pasted with a magnetic steel sheet; there is an inductive switch element on the front end of the Hall sensor, the inductive switch element is located near the magnetic steel sheet and the two are on the same level, and the Hall sensor is connected internally The photoelectric coupler and the wind cup rotate with the wind to drive the central bearing and the gear to rotate synchronously, and the Hall effect is generated between the magnetic steel sheet and the Hall sensor switch; then the frequency of the pulse signal is measured by the multi-cycle synchronization method through the single-chip microcomputer, The pulse signal is proportional to the signal of the speed of the wind cup, so that the output signal can be measured with equal precision in the whole frequency range, the measurement error is reduced, and the measurement accuracy of the frequency of the pulse signal is improved.

The invention discloses a multi-terminal model compression method and device based on knowledge federation, a task prediction method and device and electronic equipment. The multi-terminal model compression method comprises the following steps: aggregating N-th trained local models reported by a plurality of participants to obtain a global to-be-compressed model, wherein N is greater than or equalto 1; compressing the global to-be-compressed model based on a preset performance index by adopting a common data set to obtain a global compression model, wherein the preset performance index is used for representing a performance index of the global compression model during prediction, and the common data set is obtained by performing data enhancement on the data of the plurality of participants; and sending the global compression model to the plurality of participants to carry out N + 1 rounds of training.

The invention belongs to the technical field of space volume rate or loading rate accurate measurement, and discloses a laser measuring device and a measuring method. The device and the method mainlysolve the technical problem that the measuring efficiency is low when a self-developed laser measuring device is used for measuring the volume rate of the space or the loading rate of the carriage ofa truck. The laser measuring device comprises a laser emitting mechanism, a vertical rotating mechanism, a horizontal rotating mechanism and a bracket. The vertical rotating mechanism, the horizontalrotating mechanism and the laser emitting mechanism are arranged on the bracket. The vertical rotating mechanism is connected with the laser emitting mechanism and drives the laser emitting mechanismto rotate. The horizontal rotating mechanism is connected with the bracket and drives the bracket to rotate. The measuring working efficiency of the laser measuring device can be effectively improved.Meanwhile, the laser measuring device can also be installed on the top of the measured space to carry out the up-down measurement operation. The dead zone of the measured space is reduced.

The method is suitable for the technical field of model construction, and provides a model construction method, a model construction device, a terminal device and a medium. The model construction method comprises the following steps: determining a convolution layer queue from a trained face detection model; then, based on the channel weight of each feature channel of each convolution layer in theconvolution layer queue and a preset channel deletion proportion, deleting the feature channels in the convolution layer queue; arranging the convolution layers according to the arrangement sequence of the convolution layers in the convolution layer queue; performing channel deletion on each convolution layer in sequence. According to the method, the convolution layer is not deleted in the new face detection model, and only part of channels in the convolution layer are deleted, so that the resource occupancy rate when the new face detection model is operated can be reduced under the conditionof ensuring that the face detection accuracy is not changed, and meanwhile, the face image recognition efficiency is improved.

The invention discloses a mass traffic data knowledge mining and parallel processing method. The method is characterized in that an LSTM model is stored in N computing servers of a distributed cluster, a training data set is divided into N sets which are respectively input into the N computing servers, and the computing servers are trained simultaneously, so that the training time of the LSTM model can be shortened. A parameter matrix is transmitted to a parameter server after a calculation server finishes one-time forward propagation and back propagation, and the parameter matrix is decomposed into a form of multiplying two matrixes by adopting a matrix decomposition method, so that the total quantity of parameters is reduced, and the communication time consumption is reduced. A matrix compression rate is set, the matrix is compressed by using an adaptive threshold filtering method, so that the number of parameters is reduced again, and the communication time consumption is reduced again. An error matrix before and after matrix compression is calculated and is transmitted the error matrix to the next round of training, the matrix after error elimination is calculated by using the parameter matrix and the error matrix at the beginning of the next round of training, and then training is performed, so that the error caused by matrix compression can be made up, and the precision of the LSTM model is ensured.

The invention provides a sensor detection circuit, a drive control circuit board, electronic equipment and an air conditioner, and the sensor detection circuit comprises a sensor, a first sampling circuit connected with the input voltage end of the sensor and used for collecting the input voltage of the sensor, a second sampling circuit connected with the output voltage end of the sensor and used for collecting the output voltage of the sensor, and a control circuit connected with the output end of the first sampling circuit and the output end of the second sampling circuit and used for outputting a detection result according to the collected output voltage of the sensor and the collected input voltage of the sensor. Compared with an existing technical scheme, the sensor detection circuit can improve the detection precision of the detection result output by the sensor; therefore, the sensor detection circuit can adapt to a use scene with a higher detection result precision requirement.

The present invention provides an interdigital instrument terminal bus bar, which comprises a stator fixed with the bus bar, a strip finger fixed on one side of the stator and fixing holes on the stator. Wherein, fixing holes are arranged at one end of the finger. The finger is higher than a right angle of a corner while the stator is fixed. With a simple and rational structure, the present invention can not only effectively improve measuring precision of electric energy meters of different COS Phi but also eliminate impact amongst users of electric energy meters.

The invention relates to a double-disc type lubricating oil traction force testing machine, which comprises a motor I, a motor II, a coupling I, a coupling II, a lower disc shaft, an upper disc shaft, a lower disc, an upper disc, a hydraulic loading mechanism and a force testing mechanism. With the testing machine of the present invention, real working conditions of the lubricating oil under an elastohydrodynamic lubrication state can be simulated, a rolling movement between the corresponding transmission contact members is accompanied with a sliding movement, continuous loading can be achieved, and the lubricating oil traction force under an elastohydrodynamic lubrication state can be accurately measured.

The invention relates to the technical field of integrated circuits, and in particular relates to a photoresist reworking etching process. According to the photoresist reworking etching process disclosed by the invention, in an etching process manufacturing procedure, a photoresist needing to be reworked and a non-nitrogen bottom anti-reflecting layer are sequentially removed, a new non-nitrogen bottom anti-reflecting layer is re-deposited, and the photoresist is continuously coated for performing a subsequent photoetching process so as to complete a photoresist reworking process. Therefore, during the photoetching process, preparation of a protection layer on the non-nitrogen bottom anti-reflecting layer for preventing optical parameters or surface properties of the non-nitrogen bottom anti-reflecting layer from being influenced during subsequent photoetching reworking is eliminated, so that the flow of the whole photoetching process is simplified; meanwhile, by the adoption of the photoresist reworking etching process, the accuracy of a product subjected to photoresist reworking in a subsequent photoetching process remains unchanged, and the performance and yield of the product are further improved.