186 results about "Signal fusion" patented technology

The invention discloses a circuit breaker fault diagnosis method for identifying a circuit breaker fault diagnosis based on sound and vibration signal fusion processing. There are four states: stuck fault, base loose fault, and refusal to move fault. (2) Step 2, use the db3 wavelet base to decompose the acoustic and vibration signals into three layers, and reconstruct the signals on the eight nodes in the third layer; perform the Hilbert transform on the reconstructed signals of the eight nodes , to obtain the modulus envelope. (3) The envelope signal is divided into 3 sections according to the principle of before opening, during opening and after opening. According to the principle of equal energy, it is divided into 3 sections before opening, 9 sections during opening, and after opening It is divided into 3 sections, and the energy entropy of the wavelet packet square difference of each state of the circuit breaker is extracted. (4) Input the feature vector matrix of each signal into the support vector machine and use the "one vs one" voting strategy to obtain the basic probability distribution. (5) Using improved evidence theory to fuse multi-sensor information to identify circuit breaker fault types.

The invention discloses a self-balance control method of a movable type inverted pendulum system. State signals of the movable type inverted pendulum system are collected, the filter processing data method with the combination of a wavelet topology network and a regulator is adopted, a signal fusion technology is used for assistance, input data, middle data and output data of each filter node where effective values of the wavelet network enter are continuously adjusted, the controlled system control number output by the system is completed, the stability and the reliability of work of an inverted pendulum are improved, and the driving and riding stability, safety and comfort of an electric riding tool based on the inverted pendulum system are improved. The invention discloses a self-balance vehicle intelligent control system. By the adoption of a filter processing unit and a signal fusion unit, the capacity of combined control and information feedback of an electrical system and a mechanical system are improved, tiny disturbance generated in the movement process of a self-balance vehicle is reduced, the intelligent degree is improved, the self-balance vehicle intelligent control system is convenient and easy to use, comfortable, stable and capable of reducing cost and improving the cost performance.

The invention relates to a fusion processing of sampled signal from multiple sections, which is a frequency-domain analysis method suitable for signal frequency estimation. The method can compensate the influence on the signal processing results due to signal frequency difference in different sections and phase discontinuousness by setting different frequency-domain analysis parameters and phase parameter compensating factor, so as to achieve relevant information fusion and useful information extraction to obtain higher frequency estimation accuracy. The method of the invention has wide application range and is suitable for the situations of unknown initial phase of sampled signal from each section, unknown time interval between adjacent sampling sections, unknown variation of sampled signal in the time interval of adjacent sampling sections, or different sampling frequency of different sampling processes; and can achieve high-accuracy processing of signals with different frequencies sampled from multiple sections, and meet the frequency estimation requirement under the situation of signals with the same frequency sampled from multiple sections, signals with demultiplied frequencies sampled from multiple sections, or signals with split frequencies sampled from multiple sections. The method of the inventioni provides theoretical derivation process, verifies and compares experimental results, and performs real-time computing capacity analysis.

The invention creates and provides a UAV landing navigation method and device based on GPS and image vision fusion, and the method comprises the following steps of placing a ground end horizontally onthe ground, and using a wireless link to feed the azimuth angle omega and position coordinates G0 of the ground end back to an airborne end; forming a point cloud by recording the GPS value at the ground end in a recent certain time interval, fitting the contour of the point cloud with a circular curve, and noting that the coordinate value of the center of the fitted circle is G0 and the fit radius is R. The UAV landing navigation method and device based on GPS and image vision fusion provided by the invention create and realize a signal source-level image visual signal and GPS signal fusion,directly output a "GPS-like signal" which is a fusion of GPS beacon values, have stronger module portability, versatility and universality, achieve double calibration and complementary correction with the help of perspective transformation of image identification itself and feedback signals from several inclination sensors, and greatly improve the accuracy of navigation.

The invention discloses an intelligent street lamp control system based on target recognition. The system comprises a photosensitive sensor, a camera provided with a light supplementing module, an information processing module and a street lamp control module. The photosensitive sensor is used for detecting intra-regional light intensity, the camera is used for collecting image information when the light intensity is lower than a preset value, the signal processing module processes and recognizes the collected image information and sends a control instruction to a power supply controller, and the street lamp control module is used for connecting or disconnecting a street lamp power supply. The design is carried out mainly for arterial streets with relatively small pedestrian volume and traffic flow, the start triggering technology based on fusion between target recognition and photosensitive sensor signals is adopted, electric energy can be saved to the maximum limit, and the lamp service life is prolonged.

The invention provides a multi-joint combined control system and method for an exoskeleton robot. The control system inputs the real-time feedback signals of a whole-sole pressure sensing system and alimb gesture sensing system into a "multi-signal fusion and decoupling model", the model finishes the fusion and the decoupling of various signals, and outputs various categories of joint motor control quantities to control a motor to operate; a neural network is used for building the "multi-signal fusion and decoupling model", and through a quantity of training, a mature application model with multi-signal fusion and decoupling is generated; according to a human-computer interaction sensing system in a control system, various categories of joint motor control quantities output by the model are corrected in time to perform an optimization assistance function so as to achieve an optimal wearing effect of the exoskeleton robot; and on an aspect of joint motor control, the real-time currentclosed loop control and position (joint angle) closed loop control of the motor are used for realizing the quick and accurate control of the motor.

The invention discloses a monitoring system of an embedded intelligent bearing. The monitoring system comprises a bearing body, a sensing element installed in the bearing body, a signal processing module electrically connected with the sensing element and a fusion analysis module used for sensing and decision making. The invention further discloses a monitoring method of the embedded intelligent bearing monitoring system. The monitoring method specifically comprises the following steps that 1, the signal processing module receives signals from the sensing element; 2, a fusion analysis module extracts related characteristic indexes, divides the bearing operation state into three states, and judges the operation state of the bearing through the extracted fusion indexes. The method is advantaged in that vibration and temperature signals of the bearing can be sensed automatically, the fault degree of the bearing can be judged, the sensing element is installed in the bearing, precision of measured data is higher than that of a previous measurement mode, and a working state of the bearing can be judged more comprehensively and accurately in a multi-feature-index fusion mode.

The invention belongs to the technical field of fault diagnosis, and particularly relates to a fault monitoring device and diagnosis method based on multi-source signal fusion, and the device comprises a vibration signal collection unit A, a current signal collection unit B, and a signal processing unit C; the signal acquisition units A and B use sensors to acquire vibration and current data of the seawater cooling pump; the signal processing unit C carries out fast Fourier transform (FFT) processing on the signals to obtain frequency domain information, and the fault monitoring device can monitor the time-frequency domain states of vibration signals and current signals of the sea water pump in real time and is used as data preparation for fault diagnosis. According to the seawater pump fault diagnosis method based on multi-source information fusion, the migration kernel locality preserving projection algorithm is used as a model, multi-source signal information obtained by a fault monitoring device is used as input, Acc and F1-score are used as evaluation indexes, and a trained classifier is used for classifying and identifying unknown fault data, so that fault information can bequickly obtained, the generalization ability is enhanced, and the diagnosis accuracy is improved.

The invention relates to a tool condition monitoring and identifying method based on a signal fusion and multi-fractal spectrum algorithm. The tool condition monitoring and identifying method based onthe signal fusion and multi-fractal spectrum algorithm comprises the following steps of (1) in the cutting process, acquiring a cutting force signal and a vibrating signal; (2) denoising the cuttingforce signal and the vibrating signal acquired in the step 1; (3) for a denoised signal sequence, analyzing multi-fractal characteristics of the signals, searching the relation between the signals andthe tool wear through the multi-fractal characteristics, extracting relevant characteristic vectors from a multi-fractal spectrum obtained through calculation, and representing the relation between the signals and the tool wear through the characteristic vectors; and (4) merging the characteristic vectors extracted in the step 3 into a characteristic matrix, using as an input parameter variation,building a support vector machine model for tool wear condition monitoring, and utilizing the optimized support vector machine model for diagnosing the unknown tool condition. By adopting the methodprovided by the invention, the high identification rate of the tool condition is achieved.

The invention discloses a device and a method for realizing model reference adaptive control on operation speed of a combine harvester and mainly relates to the field of agricultural machinery. The device is characterized by consisting of a controller, a touch screen, a sensor system, a manual / automatic selector switch and an execution system. The sensor system is used for acquiring signals of a threshing cylinder, a header auger, and a conveying trough, an input amount signal and a loss amount signal and the like of the combine harvester; the manual / automatic selector switch is used for switching between a manual state and an automatic state; the execution system consists of an amplifier, a speed control valve and an electronic control handle; and during work, the controller receives the signals acquired by the sensor system, and the operation speed of the combine harvester is adaptively controlled by a model reference adaptive control algorithm. In the control process, the running state of the combine harvest is displayed by the touch screen, and if the combine harvest is failed, an audible and visual alarm system on the touch screen can perform audible and visual alarm.

The invention discloses an electroencephalogram signal unmanned platform intelligent control method based on a deep convolutional adversarial network. The method comprises the steps that a terminal carries out the noise removal of a collected electroencephalogram signal, and obtains a denoised electroencephalogram signal; performing deep feature extraction on the denoised electroencephalogram signal through a capsule network to obtain a deep feature signal; fusing the deep feature signal and the electroencephalogram signal and then carrying out classification and recognition to determine a corresponding control instruction signal; and the terminal performs offline and online test verification on the unmanned platform, and after verification succeeds, the unmanned platform receives and executes the control instruction signal sent by the terminal. According to the method, the existing noise data are integrated into the one-dimensional electroencephalogram signal training network, the mathematical model is simplified, the problem of insufficient noise training data is solved, the one-dimensional prediction signal is reconstructed by using the auto-encoder architecture, the attention mechanism is used for feature selection, and the calculation efficiency is improved.

The invention discloses a high-speed target detection method based on P / D band radar signal fusion processing. The method comprises the following steps: a transmitted waveform of a P-band working radar is selected as a pseudo-random binary phase shift keying pulse signal and a transmitted waveform of a D-band working radar is selected as a linear frequency modulated continuous wave signal; transmitted waveform parameters of the pseudo-random binary phase shift keying pulse of the P-band radar and the linear frequency modulated continuous wave of the D-band radar are determined; the corresponding working waveforms of the P-band and D-band are transmitted and received cooperatively; echoes of the transmitted signals of the P-band and D-band radars are processed; and the signal processing results of the two waveforms are fused to obtain the distance and velocity information of a target. According to the invention, the problem that velocity detection ambiguity of the high-speed target is likely to occur under the D-band radar is solved; and the accuracy of target ranging and velocity detection is improved.

The invention relates to the information technology application field, and particularly relates to a method for full digital real-time multiple signal fusion, and the method is mainly applied tothe monitoring command scheduling of urban traffic and an event dealing handling system. The method, by means of circumference radiation, displays the global position system (GPS), videos, flow and static resources in a map and performs four core components processing of adaptation and receptio , conversion, real-time communication and fusion of a geographic information system (GIS). This method can perform real-time unification processing of monitoring command scheduling of urban traffic and an event handling system. This method is used for solving the technical problems of multiple signal related data processing such as a large quantity of information, images ,videos, etc., in the monitoring command scheduling of urban traffic. The advantages of this method are that data processing and business logic are processed by a larger server terminus, so that a technology-supported platform which is real-time, multiple-media and united is provided, and a goal of transmitting a large quantity of data with multiple types between clients and servers is achieved .

The invention discloses a distributed multi-channel signal acquisition system for human lower-limb motion intention recognition. The signal acquisition system adopts a distributed structure and comprises a PC, a host module, a CAN bus and slave modules, wherein the PC is connected with the host module through an RS232 interface; the host module and the slave modules are connected through the CAN bus; the slave modules are in charge of signal acquisition and signal processing; the host module is in charge of multi-source signal fusion and data communication control; and a thin film pressure sensor is adopted to acquire the pressure of human feet, and a magnetic angle sensor is adopted to acquire hip joint, knee joint and ankle joint angle signals of human lower limbs. According to the distributed multi-channel signal acquisition system, as the scattered and mutually independent slave modules are adopted, the lead wire lengths of the sensors can be shortened, and electromagnetic interference on the sensors from the outside is reduced; the modules communicate with each other through the CAN bus with very high reliability; and a high-speed ARM single-chip microcomputer serves as a main controller, so that the advantages of small size, high reliability and strong anti-interference ability are achieved.

The invention provides a rolling bearing fault detection method based on acoustic vibration signal fusion, which comprises the following steps of: firstly, extracting a vibration signal and a sound signal of a rolling bearing to be detected, and establishing a sliding rectangular window function of the sound signal; secondly, by starting from a first signal point of a to-be-detected vibration signal, establishing a fusion signal after each time of movement by moving a rectangular window, solving a root-mean-square value and a mean-flexible value of the fusion signal to obtain a mean-slip value, and finding out an optimal fusion signal; and finally diagnosing a fault of a bearing to be detected by judging a value range of an approximate fault characteristic frequency where the optimal fusion signal is located. According to the invention, the problems that the vibration sensor arrangement is limited, the amplitude at the fault characteristic frequency is low and the like are solved, meanwhile, the problem that the signal-to-noise ratio is low due to the fact that sound signals are affected by background noise is solved, weak faults of a rolling bearing can be accurately recognized, the diagnosis accuracy and diagnosis efficiency are improved, and a good effect is achieved in rolling bearing state monitoring.

The invention discloses an electroencephalogram and electrocardiogram-based fatigue detection method of an electrocardiogram sensor embedded into a steering wheel. The method performs fatigue detection by constructing a product fuzzy convolutional network. The method specifically comprises the following steps: S1, acquiring electrocardiogram data through an electrocardiogram detection chip, and acquiring electroencephalogram time series data by using an electroencephalograph; S2, processing the electroencephalogram time series data by adopting a fuzzy neural network which contains a laminationand has feedback, and acquiring electroencephalogram characteristic; S3, establishing a depth feature extraction network based on a one-dimensional convolutional neural network framework to extract fatigue features of the electrocardiogram data, and generating an electrocardiogram feature sequence; S4, designing a fusion network, inputting the electrocardiogram characteristic sequence and the electroencephalogram characteristic at the same time, fusing the two signals together, and giving a prediction value; and S5, performing optimizing by using an adaptive moment estimation algorithm, and training a network model. The method can reduce the noise and improve the detection precision, the limitation of the fuzzy neural network on the feature dimension of the input data is reduced by introducing the lamination, and the accuracy of the classification result is improved.

The invention discloses a compressed sensing sampling reconstruction method and system based on a linear sampling network and a generative adversarial residual network. The method comprises the steps:obtaining a training image, and segmenting the training image into a plurality of image blocks through segmentation processing; constructing a linear sampling network to measure the image blocks to obtain measurement values corresponding to the image blocks; in the generative adversarial residual network, carrying out linear mapping processing on measurement values of all image blocks through a full connection layer to obtain an initial reconstruction result; inputting the initial reconstruction result into a residual error network, and training to obtain residual error information; performing signal fusion on the initial reconstruction result and the residual error information to obtain a generation result of the generator; jointly inputting a generation result of the generator and the original image block into a discriminator for judgment; and calculating a loss function, and performing iterative training on the linear sampling network and the generative adversarial residual networkto obtain a final image reconstruction result. The method can effectively improve the reconstruction effect at a low sampling rate.

The invention aims at overcoming shortages in the prior art, and provides a metering valve state detection method and a metering valve state detection system based on multi-signal fusion. According to the metering valve state detection method based on multi-signal fusion, the following steps are executed when a valve is in an action state: carrying out mechanical in-place detection; carrying out locked-rotor voltage detection after the mechanical in-place detection; carrying out valve in-place information judgment according to a mechanical in-place detection result and a locked-rotor voltage detection result; judging that the valve is normally installed in place at this moment when both a mechanical in-place signal and a locked-rotor voltage signal are detected; and carrying out valve in-place fault information judgment when any of the mechanical in-place signal and the locked-rotor voltage signal is not detected. According to the metering valve state detection method and the metering valve state detection system based on multi-signal fusion, the two in-place detection manners are combined, and detection can be continued after one in-place detection is failed, so that the problem that detection cannot be continued after a single in-place detection manner is failed in the prior art is overcome, the stability of the in-place detection is improved, and the accuracy rate of the in-place detection is increased.

The invention discloses a whole-ship mobile phone communication system supporting public network mobile phone multi-operator signal fusion. The invention belongs to the technical field of ship electronic information. According to the invention, a multi-operator digital wireless distribution system is adopted to realize fusion access with indoor distribution equipment, so that the equipment deployment cost is reduced; under the condition that signals of the open sea public network mobile phone base station are not covered, the shipborne cellular mobile communication base station achieves publicnetwork mobile phone signal network access through shipborne satellite communication channel transfer. Public network mobile phone signal network access is realized by the public network mobile phonebase station and the ship outdoor antenna within the signal coverage range of the offshore public network mobile phone base station, the supported public network mobile phone terminal types are expanded, and richer application capability is provided for offshore ship public network mobile phone users.

The invention discloses a Bluetooth signal positioning method and a multi-signal fusion positioning method and system, and the Bluetooth signal positioning method comprises the steps: solving a distance from a beacon emission source based on a segmented attenuation model, determining an optimal point and controlling an advancing distance, and enabling a Bluetooth signal to have a small error when the fluctuation of the Bluetooth signal is large, calculating and judging the pace and angle of travel, and realizing real-time follow-up and automatically adjusting the target position to achieve precise positioning. The multi-signal fusion positioning method comprises the steps of, based on a Bluetooth signal positioning method, carrying out fusion positioning on personnel of Bluetooth, GPS, UWB and 5G base stations according to priorities of positioning signals, therefore, the practicability is high, and different positioning schemes can be adopted according to different environments.

A probabilistic motion model to calculate the motion parameters of a user's hand held device using a noisy low cost Inertial Measurement Unit (IMU) sensor. Also described is a novel technique to reduce the bias noise present in the aforesaid IMU sensor signal, which results in a better performance of the motion model. The system utilizes a Particle Filter (PF) loop, which fuses radio signal data accumulated from Bluetooth Low Energy (BLE) beacons using BLE receiver with the signal from IMU sensor to perform localization and tracking. The Particle Filter loop operates based on a Sequential Monte Carlo technique well known to persons of ordinary skill in the art. The described approach provides a solution for both the noise problem in the IMU sensor and a motion model utilized in the Particle Filter loop, which provides better performance despite the noisy IMU sensor.

The invention provides an ultra-long-distance double-fiber interconnected multistage optical fiber time frequency transmission system. The device is provided with one or more transmission units, the transmission unit comprises a plurality of optical fiber single-channel time frequency high-precision transmission devices, a plurality of time frequency signal comprehensive devices based on multi-source time frequency signal fusion and optical fibers connected with the devices. The system performs cross transmission in different links of time-frequency signal transmission by increasing the configuration number of devices, so that the time-frequency transmission system operates stably.

The invention discloses a recording method and a recording device of electronic equipment, belongs to the technical field of communication, and can solve the problem of improving the recording quality. The electronic equipment comprises M microphones, each microphone is connected with a first voice path and a second voice path, and the scheme comprises the following steps: acquiring a shooting focal length of a camera in a video shooting state; if the shooting focal length is changed from the initial focal length to the target focal length, determining a target gain according to the initial focal length, the target focal length and the initial gain, and adjusting the gain of a second voice path connected with the ith microphone to the target gain, the initial gain being the gain of a firstvoice path connected with the ith microphone; performing signal enhancement processing on the voice signal output by the first voice path connected with the ith microphone and the voice signal outputby the second voice path to obtain an ith voice enhancement signal; and performing signal fusion processing on the M voice enhancement signals to obtain a first recording signal.

The embodiment of the invention provides a cooperative detection method and device for multi-radar space registration based on signal fusion. The method comprises the following steps: determining relative position information of a detection target and each cooperative detection radar; detecting the relative position information of the target and each cooperative detection radar, and adjusting thebeam width of each cooperative detection radar, so as to enable the ground detection regions of the beams of each cooperative detection radar to be consistent; for each cooperative detection radar, determining a weighting coefficient of an antenna array element of the cooperative detection radar by adopting a beamforming algorithm with controllable beamwidth according to the determined beamwidth;and transmitting a detection beam to a detection target based on the determined weighting coefficient of the antenna array element, and sending a reflected signal to a signal fusion processing centerthrough an underground optical fiber, so that the signal fusion processing center integrates the signals sent by the cooperative detection radars and performs cooperative processing. The complexity ofspace registration can be reduced, and multi-radar cooperative detection based on signal level data fusion can be realized.

The invention provides a high-voltage circuit breaker characteristic parameter prediction method based on multi-source signal fusion. The method comprises the following steps of judging whether a high-voltage circuit breaker is in a fault state or not, if yes, collecting fault data of the high-voltage circuit breaker at the Tth moment in the fault state, the fault data comprising a vibration signal, a sound signal, a temperature signal, a current signal and operation times; extracting feature vectors of the vibration signal, the sound signal, the temperature signal, the current signal and the number of operation times, dividing the feature vectors into a training set and a to-be-processed data set, taking the training set as the input of a preset deep learning model, and taking the predicted value of the fault data as the output of the deep learning model, performing learning training on the deep learning model; and inputting a to-be-processed data set into the trained deep learning model to obtain a predicted value of the to-be-processed data set at a T+n moment, n being any positive natural number.

The embodiment of the invention discloses a bimodal signal fusion system and method. The bimodal signal fusion system comprises: an image sensor unit, which is used for obtaining a bimodal image signal; an image processing unit which is connected with the image sensor unit, and is used for preprocessing synchronous signals and asynchronous signals in the bimodal image signals at the same time; andan image fusion unit which is connected with the image processing unit and is used for carrying out feature fusion on the synchronous signal and the asynchronous signal in the bimodal image signal. According to the scheme, the bimodal image signal processing efficiency of the image processing unit can be improved, the efficiency and real-time performance of the bimodal signal fusion system are further improved, heterogeneous feature fusion of the synchronous signal and the asynchronous signal is achieved, image signals with different types of image feature information are obtained, and the image fusion accuracy is improved. Therefore, the application range of the bimodal signal fusion system is wider.

The invention provides a signal processing method, and common signals of the quay crane are integrated so as to obtain more fusion characteristic parameters, so that information such as vibration, temperature and stress is more effectively utilized, and more accurate fault diagnosis and monitoring are carried out on the state of the quay crane. And a neural network is optimized by using an Adam algorithm, so convenience is provided for monitoring the quay crane state more accurately and quickly. The method comprises the steps: 1, collecting data; 2, preprocessing data, and determining input and output; 3, constructing a neural network model; 4, enabling the training set to train a neural network by using an Adam algorithm to obtain a quay crane state prediction model; 5, checking the quaycrane state prediction model by using the test set; 6, outputting a prediction result, and presenting the prediction result in the human-computer interaction interface. According to the invention, a plurality of signals are adopted for fusion prediction, the fault tolerance of a certain signal is increased, and the accuracy of quay crane operation state monitoring is improved.

The health state assessment method for the spaceflight measurement and control data transmission integrated equipment is high in feasibility and high in assessment conclusion credibility. According to the technical scheme, in AHP software, an equipment-level and component-level health management module, a parameter-level health management module, a subsystem-level health management module and a system-level health management grading module are adopted to form a layered health management evaluation system architecture; the equipment-level and component-level health management module collects electric signals and characterization parameters of equipment circuit modules or components and devices; various parameter signals are fused to jointly predict health states of components and equipment, and the parameter-level health management module carries out state prediction and management on parameter changes; the subsystem-level health management module performs health assessment and prediction; and the system-level health management grading module analyzes and predicts a task execution form, constructs and forms a multi-level analysis evaluation model, calculates a weight coefficient, performs layer-by-layer weight fusion and performs evidence synthesis to obtain an evaluation result of a system comprehensive state.

The invention relates to the technical field of intention recognition, and provides an upper limb wearable transfer robot motion recognition system based on multi-signal fusion. The upper limb wearable transfer robot motion recognition system comprises a sensor system and a data processing system. The sensor system comprises a surface dry electrode myoelectricity sensor, a six-axis inertial sensorand a silica gel air bag connected with an air pressure sensor. The data processing system is an upper computer integrating a human body physiological information calculation module, an electromyographic signal characteristic value calculation module, an upper limb joint angle calculation module, a four-stage processing module and an impedance regulator. The four-stage processing module comprisesa neural network module, a principal component analysis module, a moving average filtering module and an If-Then decision maker module. The output end of the sensor system is electrically connected with the input end of the data processing system, and the output end of the data processing system is electrically connected with the input end of a robot controller. According to the invention, the individual difference problem of physiological signals can be solved, and the accuracy of intention recognition and the effectiveness and safety of man-machine interaction are improved.

The invention relates to a complex multi-stagger signal sorting method based on EDW (Emitter Descriptor Word) fusion, which is mainly used to solve the problem that it is difficult to sort complex multi-stagger signals (stagger signals of which the stagger number N is larger, N is usually greater than or equal to 10). The method is implemented by the steps as follows: (1) for PDW (Pulse DescriptorWord) data input in real time within a sorting interval T, generating sorting interval EDWs through an improved PRI (Pulse Repetition Interval) transformation method, and representing all the generated sorting interval EDWs by a set Enow; (2) determining complex multi-stagger signal fusion conditions for any two EDWs in Enow (determining fusion conditions of frequency and pulse width first and then determining fusion conditions of PRI features), determining whether any two EDWs have multiple consecutive and identical sub PRIs, and fusing two EDWs if the two EDWs meet the conditions; and (3) fusing all EDWs meeting the complex multi-stagger signal fusion conditions in Enow and outputting a complex multi-stagger signal EDW formed by fusion. The method can improve the accuracy of complex multi-stagger signal sorting, and can be used in a radar detection system.