6 results about "Emitter identification" patented technology

The application relates to a kind of open set transmitting source identification method, it is related to wireless communication and signal processing technical field, solve the problem such as existing method under open set scene, it is difficult to accurately exclude unknown device, classification accuracy and robustness are insufficient etc..The method only retains Q component by cutting to fixed length IQ signal, and extracts the absolute value and spectrum amplitude of the Q component, and is fused into three-channel pseudo-image as neural network input.Classification branch is optimized for cross entropy for known training equipment, to realize closed set classification;Contrast branch simultaneously utilizes known and unknown training samples to construct positive samples, known negative samples and open set negative sample mask, reduces the similarity of unknown samples through dynamic suppression boundary mechanism and calculates contrast loss.Furthermore, according to the dynamic adjustment of classification loss weight, the convergence of known category and the dispersion of unknown samples are realized.

The present application relates to the field of radio signal identification, and particularly relates to a cross-domain few-shot specific emitter identification method based on self-supervised learning and transfer learning, comprising: adopting a self-supervised joint embedding method, pre-training a first model and a second model by using an upstream task pre-training dataset, regularizing a first autocorrelation matrix output by the first model by using von Neumann entropy, calculating the cosine similarity between the feature projections output by the first model and the second model, and updating the parameters of the first model based on the regularized autocorrelation matrix and the cosine similarity; migrating the pre-trained first encoder to a feature extractor, inputting a downstream task fine-tuning dataset into the connected feature extractor and classification head to fine-tune the feature extractor and the classification head; inputting a wireless signal into the fine-tuned feature extractor and classification head to obtain the predicted transmitter category of the signal. The present application achieves good identification accuracy under the conditions of few samples and low cost.

The present disclosure relates to a method for detecting a skimming device in a point of sale, POS, the method comprising: a plurality of successive scanning cycles each comprising scanning for emitters using a wireless communication protocol to receive emitter identifiers; generating and updating a list of detected emitters comprising for each currently detected emitter an emitter identifier and a first detection time; for each of the currently detected emitters in the detected emitter list, comparing an active time threshold with an active time of the currently detected emitter from the first detection time of the currently detected emitter, and / or comparing the identifier of the currently detected emitter with identifiers in a whitelist of authorized identifiers; and when a currently detected emitter has an active time exceeding the active time threshold and an identifier not belonging to the whitelist, activating an alert signal.

A method for time synchronization of sensors in a distributed system, wherein an up-to-date value of an internal time unit of a sensor is stored when capturing sensor data, at least the value of the internal time unit stored when a characteristic time feature of an identification signal is received, an associated transmitter identification of the identification signal and a further value of the internal time unit is transferred together with the sensor data to the central unit, and where the central unit captures voltage and phase values from at least one phase conductor of a power grid as the current pointer over more than one measuring period, and the current pointers of the respective sensor unit are aggregated to form aggregated current pointers that are transferred to the central unit, and correct-phase powers are calculated from the voltage values and the aggregated current pointers in the central unit.

Method and system for time synchronization of sensor units. Method for time synchronization of sensor units in a distributed system, wherein sensor data is acquired by the sensor units and transmitted to a central unit, wherein, upon acquisition of the sensor data, a further current value of an internal time unit of the sensor unit is stored, wherein the sensor data is at least assigned the value of the internal time unit stored upon receipt of a characteristic signal, the transmitter identification of the characteristic signal and the further value of the internal time unit stored upon acquisition of the sensor data and transmitted to the central unit together with the sensor data, wherein the central unit acquires voltage and phase values of at least one phase conductor as current vectors over more than one measurement cycle and aggregates the current vectors to an aggregated current vector and transmits the aggregated current vector to the central unit and calculates the power from the voltage values and the aggregated current vector.

This invention proposes a class-incremental radar radiation source identification method based on temporal adaptive clustering and orthogonal perturbation pseudo-feature generation. The steps are as follows: The acquired radar radiation source signal is constructed in three channels, including the time-domain waveform, the real part and imaginary part of the Fast Fourier Transform, and then standardized; fused features are extracted using multi-scale convolution and a multi-head self-attention network; the feature extractor is frozen, and the old class features are subjected to temporal adaptive clustering according to category, dynamically determining the number of clusters based on temporal complexity; pseudo-feature sets are generated based on the cluster centers through intra-class Gaussian perturbation and class boundary orthogonal perturbation; the old class pseudo-features are merged with the new class real data, and multi-stage incremental training is used to achieve rapid adaptation of the new class and maintain the performance of the old class; features are extracted and classified and output during the inference stage using the same steps. This invention maintains high recognition accuracy even without old class real data and is suitable for continuous identification of individual radars in dynamic electromagnetic environments.