CN112130118A - SNN-based ultra-wideband radar signal processing system and processing method - Google Patents
SNN-based ultra-wideband radar signal processing system and processing method Download PDFInfo
- Publication number
- CN112130118A CN112130118A CN202010841963.XA CN202010841963A CN112130118A CN 112130118 A CN112130118 A CN 112130118A CN 202010841963 A CN202010841963 A CN 202010841963A CN 112130118 A CN112130118 A CN 112130118A
- Authority
- CN
- China
- Prior art keywords
- snn
- uwb
- module
- neural network
- neural
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012545 processing Methods 0.000 title claims abstract description 27
- 238000003672 processing method Methods 0.000 title claims abstract description 12
- 238000004364 calculation method Methods 0.000 claims abstract description 36
- 238000000034 method Methods 0.000 claims abstract description 30
- 238000004458 analytical method Methods 0.000 claims abstract description 21
- 230000001537 neural effect Effects 0.000 claims abstract description 21
- 238000013528 artificial neural network Methods 0.000 claims description 45
- 238000001208 nuclear magnetic resonance pulse sequence Methods 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000003062 neural network model Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 5
- 230000003993 interaction Effects 0.000 claims description 4
- 230000008569 process Effects 0.000 abstract description 19
- 238000013473 artificial intelligence Methods 0.000 abstract description 6
- 238000001514 detection method Methods 0.000 abstract description 6
- 230000019771 cognition Effects 0.000 abstract description 3
- 230000008447 perception Effects 0.000 abstract description 3
- 230000008054 signal transmission Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 4
- 238000012706 support-vector machine Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005034 decoration Methods 0.000 description 2
- 238000002592 echocardiography Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000015654 memory Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000037081 physical activity Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012421 spiking Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/28—Details of pulse systems
- G01S7/285—Receivers
- G01S7/295—Means for transforming co-ordinates or for evaluating data, e.g. using computers
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/28—Details of pulse systems
- G01S7/285—Receivers
- G01S7/292—Extracting wanted echo-signals
Abstract
The invention discloses an SNN-based ultra-wideband radar signal processing system and a processing method, which relate to the technical field of brain-like artificial intelligence and comprise an IR-UWB sensor used for detecting and receiving a reflected pulse signal, wherein an IR-UWB module used for carrying out time-event coding on the reflected pulse signal is arranged in the IR-UWB sensor, the IR-UWB module inputs coded data into an SNN array module for carrying out neural mimicry calculation, the data obtained after the neural mimicry calculation is sent to a corresponding processor module for analysis, and the processor module feeds back result information to the SNN array module after the analysis is finished. Perception (radar pulse signal detection) and cognition (neural mimicry calculation) are integrated, the traditional radar signal processing and signal reconstruction process is avoided, and the result can be directly output through full-event-driven signal transmission and a model.
Description
Technical Field
The invention relates to the technical field of brain-like artificial intelligence, in particular to an SNN-based ultra-wideband radar signal processing system and a processing method.
Background
Impulse Radio-UI tra Wideband (IR-UWB, hereinafter referred to as ultra Wideband) is a short-range wireless communication technology. Unlike conventional wireless communication techniques, ultra-wideband does not use carrier modulation, but directly uses very narrow (nanosecond or sub-nanosecond) pulses to modulate, and these pulses are transmitted to the antenna directly or through a buffer, and generally the system does not need more complex circuits, such as local oscillator and up-conversion mixer.
The IR-UWB signal has extremely short pulse duration and extremely high time domain resolution, has lower complexity and low cost, strong anti-multipath interference, wide frequency spectrum range (up to several GHz or above), has the characteristics of penetrability, low power consumption and the like, and can be widely applied to the applications of Internet of things, wireless positioning, short-distance ranging, low-cost radars and the like.
The radar sensor based on IR-UWB generally integrates a plurality of components such as a transmitting end, an amplifier end, a receiving end and the like, and is consistent with other radar principles, and the UWB radar sensor can be reflected and received by the receiving end under the condition that the UWB signal is transmitted by the transmitting end and meets an obstacle. Compared with the traditional radar, the radar sensor based on the IR-UWB has the advantages of high time domain precision, small size, low power consumption and low cost, so that the radar sensor based on the IR-UWB is widely applied to non-contact sensors, such as human body monitoring, gesture recognition, human body activity analysis, target monitoring, man-machine interaction and the like, and meanwhile, because the IR-UWB signals have penetrability, the radar sensor based on the IR-UWB also can be applied to applications such as nondestructive internal detection and the like. The received signals acquired using an IR-UWB radar sensor are typically processed using a fast-time and slow-time sampling system, as shown in fig. 1. In the fast time dimension, storing L samples in a period of time after the sampling time begins in a digital memory to obtain L range gates; in the slow time dimension, the pulse radar transmits not only one pulse but a group of pulses, and the dimension of the repetition axes is the slow time axis.
In view of the received signal form of IR-UWB, conventional IR-UWB radar sensors generally process binary images in the XY direction of three-dimensional echoes. The signal processing method and steps are shown in fig. 2, and after detecting and receiving the signal, the received radar echo signal is first preprocessed. The preprocessing includes clutter suppression, background noise removal, and the like, and in general, the preprocessing may use a direct subtraction-average method, a singular value decomposition method, a moving target detection method, and the like to perform clutter filtering. And then, extracting the signal characteristics by adopting different methods according to the application scene and the working requirement. The extraction method varies according to different application types, and in multi-target detection, Principal Component Analysis (PCA) is generally adopted; in gesture recognition, time-frequency feature extraction or a Range-Doppler (RD) algorithm is generally adopted. After the features are extracted, the features are analyzed by adopting a mode recognition or machine learning method to finally obtain a result. Conventional machine learning methods include Nearest Neighbor classifiers (K-Nearest neighbors), decision tree classifiers, Support Vector Machines (SVMs), and the like. With the improvement of computing power, an analysis method using a convolutional neural network is also widely applied to feature extraction of IR-UWB radar echoes.
With the rapid development of artificial intelligence, the method is widely applied to machine vision, and also extends from image recognition classification to microwave vision represented by radar application, and the method generally utilizes deep learning and an Artificial Neural Network (ANN) method to perform pattern recognition and classification on radar signals. The method is also used for IR-UWB radar sensors, including Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), long short term memory networks (LSTM), Support Vector Machines (SVM), etc.
In conclusion, artificial intelligence based on the ANN has great demand for hardware resources, and the difficulty of deploying artificial intelligence algorithms in low-power-consumption equipment is greatly increased. However, with the rise of the neural mimicry calculation, it is possible to realize low-power-consumption and high-efficiency artificial intelligence calculation (energy efficiency is improved by orders of magnitude) by simulating the human brain working mode. The present patent therefore proposes new improvements in the state of the art described above, in particular in the processing of radar signals.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides an SNN-based ultra-wideband radar signal processing system and a processing method, and by integrating perception (radar pulse signal detection) and cognition (neural mimicry calculation), the traditional radar signal processing and signal reconstruction process is avoided, and a result is directly output through a signal transmission and model driven by a full event.
In order to achieve the above object, the present invention provides the following technical solutions:
the SNN-based ultra-wideband radar signal processing system comprises an IR-UWB sensor used for detecting and receiving reflected pulse signals, wherein an IR-UWB module used for carrying out time-event coding on the reflected pulse signals is arranged in the IR-UWB sensor, the IR-UWB module inputs coded data into an SNN array module for carrying out neural mimicry calculation, the data obtained after the neural mimicry calculation is sent to a corresponding processor module for analysis, and the processor module feeds result information back to the SNN array module after the analysis is finished;
or the system also comprises a neural network state machine which can realize interaction with the SNN array module, wherein the neural network state machine adopts WTA rules, determines the number of state machines in the neural network state machine according to the quantity to be classified, and trains the neural network state machine through a pulse time domain sequence recognition training algorithm.
The processing method of the SNN-based ultra-wideband radar signal processing system comprises the following steps:
the IR-UWB sensor detects and receives a reflected pulse signal;
the IR-UWB module carries out time-event coding on the reflection pulse signal received in the step (1) to obtain coded data;
the SNN array module receives the data coded in the step (2), adds time domain expression to the coded data through the neural mimicry calculation to form a dynamic pulse sequence, and transmits the dynamic pulse sequence to the processor module for analysis;
and (4) after the processor module completes analysis, feeding back analysis result information to the SNN array module to realize real-time network update.
Preferably, the neural mimicry calculation in step (3) is specifically realized by a spiking neural network model.
Advantageous effects
The SNN-based ultra-wideband radar signal processing system and the SNN-based ultra-wideband radar signal processing method have the following advantages:
1. due to the sparsity of the pulse sequence and the event-driven characteristic, the system avoids the collection, transmission and processing of a large amount of redundant data, and the IR-UWB radar has low energy consumption, so the system has extremely low power consumption.
And 2, the IR-UWB radar directly adopts pulses with very narrow time width (nanosecond or subnanosecond level) for modulation, and the time domain precision is extremely high, so that the IR-UWB radar has the characteristic of high real-time degree.
And 3, the SNN can naturally process pulse data, and for a pulse sequence generated by the IR-UWB radar, the pulse signal can be directly input into a network without processing steps of signal type conversion, feature extraction and the like.
In IR-UWB radar, the pulses are delivered to the antenna directly or via a buffer, typically without the need for more complex circuitry in the system, which is less complex and suitable for low cost deployment and application.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the structures shown in the drawings without creative efforts.
FIG. 1 is a diagram illustrating a received signal acquired by a conventional IR-UWB radar sensor;
FIG. 2 is a conventional IR-UWB radar sensor signal processing method;
FIG. 3 is a diagram of an architecture of an ultra-wideband radar signal processing system with a supervised learning impulse neural network for an SNN based ultra-wideband radar signal processing system according to the present invention;
FIG. 4 is an architecture diagram of an unsupervised learning impulse neural network ultra-wideband radar signal processing system of the SNN-based ultra-wideband radar signal processing system of the present invention;
FIG. 5 is a schematic diagram of the signal processing of the supervised learning impulse neural network ultra-wideband radar of the SNN-based ultra-wideband radar signal processing system of the present invention;
fig. 6 is a schematic diagram of processing an ultra-wideband radar signal of an unsupervised learning impulse neural network of the SNN-based ultra-wideband radar signal processing system according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The SNN-based ultra-wideband radar signal processing system comprises two forms of supervised learning and unsupervised learning, wherein the supervised learning form is shown in fig. 3 and comprises an IR-UWB sensor used for detecting and receiving reflected pulse signals, an IR-UWB module used for carrying out time-event coding on the reflected pulse signals is arranged in the IR-UWB sensor, the IR-UWB module inputs coded data into an SNN array module for carrying out neurostimid calculation, and sends the data obtained after the neurostimid calculation to a corresponding processor module for analysis, and the processor module feeds back result information to the SNN array module after the analysis is finished;
and in an unsupervised learning mode, on the basis of supervised learning, the system also comprises a neural network state machine capable of realizing interaction with the SNN array module, wherein the neural network state machine adopts WTA rules, determines the number of state machines in the neural network state machine according to the quantity to be classified, and trains the neural network state machine through a pulse time domain sequence recognition training algorithm.
The processing method of the SNN-based ultra-wideband radar signal processing system comprises the following steps:
the IR-UWB sensor detects and receives a reflected pulse signal;
the IR-UWB module carries out time-event coding on the reflection pulse signal received in the step (1) to obtain coded data;
the SNN array module receives the data coded in the step (2), adds time domain expression to the coded data through the neural mimicry calculation to form a dynamic pulse sequence, and transmits the dynamic pulse sequence to the processor module for analysis; the neural mimicry calculation is specifically realized through a pulse neural network model;
and (4) after the processor module completes analysis, feeding back analysis result information to the SNN array module to realize real-time network update.
Specifically, as shown in fig. 5, an algorithm flow of supervised learning is that, firstly, according to an echo of a radar, an IR-UWB module performs time-event encoding, and encoded data can be directly input to perform a neuromorphic calculation.
The process of the neural mimicry calculation is carried out by using a pulse neural network model, and the calculation is mainly divided into the following two processes:
the first is the calculation of the existing output, the pulse passes through each node of each layer of neural network and generates the output, and finally the neural network can generate a final output which is irrelevant to the target application and is the result of self calculation;
secondly, according to the type of the output pulse sequence which is desired to be obtained, new weight is obtained through analysis and result output and neural mimicry calculation, the process is simulation Propagation, and the result of next self-calculation after learning is gradually similar to the target result;
by repeating the two processes, supervised identification and classification of the targets can be realized according to the application scene. In addition, the impulse neural network can be customized according to application scenes, and the larger the network is, the richer the pattern types which can be identified are, and more differentiated classifications can be accommodated. The model ensures that the recognition object can be accurately recognized at any position and any direction of the IR-UWB input matrix.
Specifically, taking gesture recognition under supervised learning as an example, firstly, according to the echo of the radar, the IR-UWB module performs time-event encoding, and the encoded data can be directly input to perform neuromorphic calculation.
The SNN neural network for gesture recognition is simple in structure and only comprises an input layer, a hidden layer and an output layer.
The process of the neural mimicry calculation is carried out by using a pulse neural network model, and the calculation is mainly divided into the following two processes:
the first is the calculation of the existing output, the pulse passes through each node of each layer of neural network and generates the output, finally, the neural network generates a final output, and the number of neurons in the output layer is determined by the action to be identified.
The second is that according to the type of the output pulse sequence that we want to obtain, for example, we can design different time sequence impulse response sequences according to different gestures, through analysis and result output, calculate the loss function, and calculate new weight through the loss function and response input and output, this process is a slope Gradient Learning rule, and the result of the next self-calculation after Learning will be gradually similar to the target result.
Two processes are repeated, supervised identification and classification of targets can be achieved according to application scenes, and in the repeated process, the learning rate is gradually reduced, and the model is guaranteed not to stay in the local optimal solution.
An algorithm flow of unsupervised learning is shown in fig. 6, firstly, according to the echo of the radar, the IR-UWB module performs time-event coding, and the coded data can be directly input to perform neuromorphic calculation; the process of the neural mimicry calculation is carried out by using a pulse neural network model, and the calculation is mainly divided into the following two processes:
the first is the calculation of the existing output, the pulse passes through each node of each layer of neural network and generates the output, finally, the neural network can generate a final output, the output is irrelevant to the target application and is the result of self calculation, and the output can lead the neural network to generate a fixed state;
secondly, after a plurality of neural network state machines are generated according to the required classification, signals input into the whole neural network state machine system at intervals are generated, the process is called as the initialization of the neural network, after one signal is input each time, a control neuron restores the state of the whole neural network, and meanwhile, a special weight attenuation structure is used for ensuring that the same neural network state machine cannot be continuously excited in two continuous samples (the input signals are required to be respectively in different classes); when all the state machines are excited at least once and the weights of all the state machines of the neural network are updated once, the control neurons are removed, various random pulse samples are input into the state machines of the neural network, and the weight attenuation structure is removed through some mode phase change, so that the related state machines of the neural network can be continuously excited (if continuous same type signal input is met) and the weights are continuously updated. The termination condition of the process is that no signal is adjusted and classified to another neural network state machine, and the method enables all similar inputs to excite a fixed neural network state;
by repeating the two processes, the unsupervised identification and classification of the target can be realized according to the application scene. Different from supervised learning, an auto-supervised learning structure of a neural network state machine is added, the neural network state machine adopts a WTA (wind-Take-All) rule, the number of state machines in the neural network state machine is determined according to the number of required classifications, the neural network state machine is trained through a pulse time domain sequence recognition training algorithm, and for example, if 5 gestures are required to be recognized, 5 neural network state machines are generated.
In summary, the invention integrates perception (radar pulse signal detection) and cognition (neural mimicry calculation), avoids the traditional radar signal processing and signal reconstruction process, and directly outputs results through signal transmission and model driven by full events, under the system framework:
1. the-10 db relative bandwidth of the signal generator is greater than 20% of the central frequency, or the-10 db bandwidth of the signal exceeds 500MHz, then the signal is UWB (Ultra wide Band, UWB) technology is a wireless carrier communication technology, it does not adopt sinusoidal carrier, but utilizes nanosecond-level non-sinusoidal wave narrow pulse to transmit data, therefore its occupied frequency spectrum range is very wide. Wherein the definition of the relative bandwidth η is:
in the formula fhIs the upper bound of the signal spectrum, flThe lower bound of the signal spectrum.
2. After the echo signal is detected, an Event sequence (Event lndex) is directly generated, and the echo signal directly enters a neural mimicry computing unit without transformation and is processed intelligently by using a pulse neural network;
3. performing SNN using a neuromorphic architecture, including but not limited to general purpose computers and servers, GPUs, FPGAs, application specific integrated circuit chips, or novel brain-like computing devices and architectures including memristors;
4. the SNN model training method is typically applied to gesture recognition, heartbeat and respiration recognition, human motion recognition and the like.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (3)
1. SNN-based ultra-wideband radar signal processing system is characterized in that: the sensor comprises an IR-UWB sensor used for detecting and receiving a reflected pulse signal, wherein an IR-UWB module used for carrying out time-event coding on the reflected pulse signal is arranged in the IR-UWB sensor, the IR-UWB module inputs coded data into an SNN array module for carrying out neural mimicry calculation, the data obtained after the neural mimicry calculation is sent to a corresponding processor module for analysis, and the processor module feeds back result information to the SNN array module after the analysis is finished;
or the system also comprises a neural network state machine which can realize interaction with the SNN array module, wherein the neural network state machine adopts WTA rules, determines the number of state machines in the neural network state machine according to the quantity to be classified, and trains the neural network state machine through a pulse time domain sequence recognition training algorithm.
2. The processing method of the SNN-based ultra-wideband radar signal processing system according to claim 1, characterized in that: the method comprises the following steps:
the IR-UWB sensor detects and receives a reflected pulse signal;
the IR-UWB module carries out time-event coding on the reflection pulse signal received in the step (1) to obtain coded data;
the SNN array module receives the data coded in the step (2), adds time domain expression to the coded data through the neural mimicry calculation to form a dynamic pulse sequence, and transmits the dynamic pulse sequence to the processor module for analysis;
and (4) after the processor module completes analysis, feeding back analysis result information to the SNN array module to realize real-time network update.
3. The processing method of the SNN-based ultra-wideband radar signal processing system according to claim 2, characterized in that: the calculation of the neural mimicry in the step (3) is specifically realized through a pulse neural network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010841963.XA CN112130118B (en) | 2020-08-19 | 2020-08-19 | Ultra-wideband radar signal processing system and method based on SNN |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010841963.XA CN112130118B (en) | 2020-08-19 | 2020-08-19 | Ultra-wideband radar signal processing system and method based on SNN |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112130118A true CN112130118A (en) | 2020-12-25 |
| CN112130118B CN112130118B (en) | 2023-11-17 |
Family
ID=73851721
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010841963.XA Active CN112130118B (en) | 2020-08-19 | 2020-08-19 | Ultra-wideband radar signal processing system and method based on SNN |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112130118B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112699956A (en) * | 2021-01-08 | 2021-04-23 | 西安交通大学 | Neural morphology visual target classification method based on improved impulse neural network |
| CN113056750A (en) * | 2021-01-25 | 2021-06-29 | 成都时识科技有限公司 | Equipment anomaly detection method, computer readable storage medium, chip and device |
| CN113050060A (en) * | 2021-03-25 | 2021-06-29 | 南京敏智达科技有限公司 | Life detection method and detection device based on sparse reconstruction and neural network |
| CN113255905A (en) * | 2021-07-16 | 2021-08-13 | 成都时识科技有限公司 | Signal processing method of neurons in impulse neural network and network training method |
| CN114399033A (en) * | 2022-03-25 | 2022-04-26 | 浙江大学 | Brain-like computing system and method based on neuron instruction coding |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2358611A1 (en) * | 2000-10-10 | 2002-04-10 | Heidelberger Druckmaschinen Ag | System and method for interfacing with a production scanner |
| CN103036590A (en) * | 2013-01-01 | 2013-04-10 | 复旦大学 | High-swing impulse radio ultra-wideband transmitter |
| US20140019392A1 (en) * | 2012-06-01 | 2014-01-16 | Brain Corporation | Intelligent modular robotic apparatus and methods |
| WO2016070188A1 (en) * | 2014-11-02 | 2016-05-06 | Kim Revyn | Smart audio headphone system |
| CN106344005A (en) * | 2016-10-28 | 2017-01-25 | 张珈绮 | Mobile ECG (electrocardiogram) monitoring system and monitoring method |
| CN108919218A (en) * | 2018-06-07 | 2018-11-30 | 北京邮电大学 | A kind of contactless number of people in car and the method and device of position judgement |
| CN109407067A (en) * | 2018-10-13 | 2019-03-01 | 中国人民解放军海军航空大学 | Radar moving targets detection and classification integral method based on time-frequency figure convolutional neural networks |
| KR20190040822A (en) * | 2017-10-11 | 2019-04-19 | 엘지전자 주식회사 | Mobile terminal |
| CN109906101A (en) * | 2016-08-25 | 2019-06-18 | 天堂医疗公司 | System and method for handling nerve signal |
| CN110210563A (en) * | 2019-06-04 | 2019-09-06 | 北京大学 | The study of pattern pulse data space time information and recognition methods based on Spike cube SNN |
| WO2019182265A1 (en) * | 2018-03-21 | 2019-09-26 | 엘지전자 주식회사 | Artificial intelligence device and method for operating same |
| CN110288510A (en) * | 2019-06-11 | 2019-09-27 | 清华大学 | A kind of nearly sensor vision perception processing chip and Internet of Things sensing device |
| CN110426560A (en) * | 2019-06-28 | 2019-11-08 | 北京大学 | A kind of pulse array signals space-time up-sampler generation method |
| US20190383904A1 (en) * | 2018-06-13 | 2019-12-19 | Metawave Corporation | Autoencoder assisted radar for target identification |
| US20200184556A1 (en) * | 2018-05-06 | 2020-06-11 | Strong Force TX Portfolio 2018, LLC | Adaptive intelligence and shared infrastructure lending transaction enablement platform responsive to crowd sourced information |
| CN111272764A (en) * | 2020-01-22 | 2020-06-12 | 哈尔滨工业大学 | Large intelligent temporary platform non-contact image recognition mobile management and control system and method |
| CN111334982A (en) * | 2018-12-19 | 2020-06-26 | Lg电子株式会社 | Laundry treatment apparatus and method of operating the same |
-
2020
- 2020-08-19 CN CN202010841963.XA patent/CN112130118B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2358611A1 (en) * | 2000-10-10 | 2002-04-10 | Heidelberger Druckmaschinen Ag | System and method for interfacing with a production scanner |
| US20140019392A1 (en) * | 2012-06-01 | 2014-01-16 | Brain Corporation | Intelligent modular robotic apparatus and methods |
| CN103036590A (en) * | 2013-01-01 | 2013-04-10 | 复旦大学 | High-swing impulse radio ultra-wideband transmitter |
| WO2016070188A1 (en) * | 2014-11-02 | 2016-05-06 | Kim Revyn | Smart audio headphone system |
| CN109906101A (en) * | 2016-08-25 | 2019-06-18 | 天堂医疗公司 | System and method for handling nerve signal |
| CN106344005A (en) * | 2016-10-28 | 2017-01-25 | 张珈绮 | Mobile ECG (electrocardiogram) monitoring system and monitoring method |
| KR20190040822A (en) * | 2017-10-11 | 2019-04-19 | 엘지전자 주식회사 | Mobile terminal |
| WO2019182265A1 (en) * | 2018-03-21 | 2019-09-26 | 엘지전자 주식회사 | Artificial intelligence device and method for operating same |
| US20200184556A1 (en) * | 2018-05-06 | 2020-06-11 | Strong Force TX Portfolio 2018, LLC | Adaptive intelligence and shared infrastructure lending transaction enablement platform responsive to crowd sourced information |
| CN108919218A (en) * | 2018-06-07 | 2018-11-30 | 北京邮电大学 | A kind of contactless number of people in car and the method and device of position judgement |
| US20190383904A1 (en) * | 2018-06-13 | 2019-12-19 | Metawave Corporation | Autoencoder assisted radar for target identification |
| CN109407067A (en) * | 2018-10-13 | 2019-03-01 | 中国人民解放军海军航空大学 | Radar moving targets detection and classification integral method based on time-frequency figure convolutional neural networks |
| CN111334982A (en) * | 2018-12-19 | 2020-06-26 | Lg电子株式会社 | Laundry treatment apparatus and method of operating the same |
| CN110210563A (en) * | 2019-06-04 | 2019-09-06 | 北京大学 | The study of pattern pulse data space time information and recognition methods based on Spike cube SNN |
| CN110288510A (en) * | 2019-06-11 | 2019-09-27 | 清华大学 | A kind of nearly sensor vision perception processing chip and Internet of Things sensing device |
| CN110426560A (en) * | 2019-06-28 | 2019-11-08 | 北京大学 | A kind of pulse array signals space-time up-sampler generation method |
| CN111272764A (en) * | 2020-01-22 | 2020-06-12 | 哈尔滨工业大学 | Large intelligent temporary platform non-contact image recognition mobile management and control system and method |
Non-Patent Citations (4)
| Title |
|---|
| A TAVANAEI等: "Deep learning in spiking neural networks", 《NEURAL NETWORKS》, pages 47 - 63 * |
| M. R. YUCE;J-M. REDOUTÉ;KASUN M. THOTAHEWA;HO CHEE KEONG: "Development of low-power UWB body sensors", 《INFORMATION TECHNOLOGIES (ISCIT)》, pages 1 - 9 * |
| 崔丽辉;赵安兴;宁方正;: "基于EMD和BP神经网络的雷达体征信号检测算法", 计算机系统应用, no. 08 * |
| 霍凯;姜卫东;黎湘;毛钧杰;: "一种新的OFDM相位编码频率步进雷达信号及其特性", 电子与信息学报, no. 03 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112699956A (en) * | 2021-01-08 | 2021-04-23 | 西安交通大学 | Neural morphology visual target classification method based on improved impulse neural network |
| CN112699956B (en) * | 2021-01-08 | 2023-09-22 | 西安交通大学 | Neuromorphic visual target classification method based on improved impulse neural network |
| CN113056750A (en) * | 2021-01-25 | 2021-06-29 | 成都时识科技有限公司 | Equipment anomaly detection method, computer readable storage medium, chip and device |
| CN113056750B (en) * | 2021-01-25 | 2023-05-12 | 成都时识科技有限公司 | Equipment abnormality detection method, computer readable storage medium, chip and device |
| CN113050060A (en) * | 2021-03-25 | 2021-06-29 | 南京敏智达科技有限公司 | Life detection method and detection device based on sparse reconstruction and neural network |
| CN113050060B (en) * | 2021-03-25 | 2024-04-26 | 南京敏智达科技有限公司 | Life detection method and device based on sparse reconstruction and neural network |
| CN113255905A (en) * | 2021-07-16 | 2021-08-13 | 成都时识科技有限公司 | Signal processing method of neurons in impulse neural network and network training method |
| CN113255905B (en) * | 2021-07-16 | 2021-11-02 | 成都时识科技有限公司 | Signal processing method of neurons in impulse neural network and network training method |
| CN114399033A (en) * | 2022-03-25 | 2022-04-26 | 浙江大学 | Brain-like computing system and method based on neuron instruction coding |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112130118B (en) | 2023-11-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112130118B (en) | Ultra-wideband radar signal processing system and method based on SNN | |
| Sun et al. | Real-time radar-based gesture detection and recognition built in an edge-computing platform | |
| US11269058B2 (en) | Autoencoder assisted radar for target identification | |
| Al Hadhrami et al. | Transfer learning with convolutional neural networks for moving target classification with micro-Doppler radar spectrograms | |
| Omidvar et al. | Neural networks and pattern recognition | |
| CN110245581B (en) | Human behavior recognition method based on deep learning and distance-Doppler sequence | |
| Safa et al. | Improving the accuracy of spiking neural networks for radar gesture recognition through preprocessing | |
| EP3882654A1 (en) | Object classification using low-level radar data | |
| Arsalan et al. | Resource efficient gesture sensing based on fmcw radar using spiking neural networks | |
| Tang et al. | Human activity recognition based on mixed CNN with radar multi-spectrogram | |
| Cheng et al. | A real-time human activity recognition method for through-the-wall radar | |
| Zhang et al. | Temporal-range-doppler features interpretation and recognition of hand gestures using MmW FMCW radar sensors | |
| Akyon et al. | Classification of intra-pulse modulation of radar signals by feature fusion based convolutional neural networks | |
| Kazemi et al. | Deep learning for direct automatic target recognition from SAR data | |
| Safa et al. | On the use of spiking neural networks for ultralow-power radar gesture recognition | |
| Gan et al. | Gesture recognition system using 24 GHz FMCW radar sensor realized on real-time edge computing platform | |
| Jin et al. | Robust dynamic hand gesture recognition based on millimeter wave radar using atten-TsNN | |
| CN115343704A (en) | Gesture recognition method of FMCW millimeter wave radar based on multi-task learning | |
| Liang et al. | A posture recognition based fall detection system using a 24GHz CMOS FMCW radar SoC | |
| Arsalan et al. | Radar-based gesture recognition system using spiking neural network | |
| CN115469303A (en) | Cognitive biological radar method and device for detecting human body posture and vital signs | |
| Ruan et al. | Automatic recognition of radar signal types based on CNN-LSTM | |
| Park et al. | Increasing accuracy of hand gesture recognition using convolutional neural network | |
| Yang et al. | A lightweight multi-scale neural network for indoor human activity recognition based on macro and micro-doppler features | |
| Shaaban et al. | Evaluation of spiking neural networks for time domain-based radar hand gesture recognition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |