CN117269912A - Signal source number calculation method, training method of first model and signal processing method - Google Patents
Signal source number calculation method, training method of first model and signal processing method Download PDFInfo
- Publication number
- CN117269912A CN117269912A CN202210653922.7A CN202210653922A CN117269912A CN 117269912 A CN117269912 A CN 117269912A CN 202210653922 A CN202210653922 A CN 202210653922A CN 117269912 A CN117269912 A CN 117269912A
- Authority
- CN
- China
- Prior art keywords
- model
- incoming wave
- training
- signal
- echo signals
- 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.)
- Pending
Links
- 238000012549 training Methods 0.000 title claims abstract description 142
- 238000004364 calculation method Methods 0.000 title claims abstract description 89
- 238000000034 method Methods 0.000 title claims abstract description 87
- 238000003672 processing method Methods 0.000 title claims abstract description 15
- 238000012545 processing Methods 0.000 claims abstract description 46
- 238000005259 measurement Methods 0.000 claims description 35
- 238000004590 computer program Methods 0.000 claims description 24
- 230000006870 function Effects 0.000 claims description 17
- 238000004088 simulation Methods 0.000 claims description 16
- 238000001514 detection method Methods 0.000 claims description 15
- 238000004891 communication Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 description 18
- 239000011159 matrix material Substances 0.000 description 16
- 238000004422 calculation algorithm Methods 0.000 description 12
- 238000010586 diagram Methods 0.000 description 10
- 238000005457 optimization Methods 0.000 description 4
- 238000012706 support-vector machine Methods 0.000 description 4
- 230000004069 differentiation Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/417—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section involving the use of neural networks
-
- 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
- G01S3/00—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
- G01S3/02—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
- G01S3/14—Systems for determining direction or deviation from predetermined direction
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computer Networks & Wireless Communication (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The application relates to a signal source number calculation method, a first model training method and an echo signal processing method. The signal source number calculating method comprises the following steps: obtaining K incoming wave directions corresponding to echo signals, wherein K is more than or equal to 1; according to the K incoming wave directions, calculating to obtain first model features; and processing the first model characteristic according to a first model, and determining the number of signal sources corresponding to the echo signals. The method can ensure the accuracy of the number of the signal sources and expand the application range. The application also relates to a training method of the first model, a signal source number calculation device, a training device of the first model, an integrated circuit, a radio device and equipment.
Description
Technical Field
The present disclosure relates to the field of radar technologies, and in particular, to a method for calculating the number of signal sources, a first model training method, and an echo signal processing method.
Background
In recent years, as the application of radars in the fields of automatic driving, security protection, unmanned aerial vehicle and the like is more and more widespread, higher requirements are put forward on the angle measurement precision of the radars in the application process, and the angle measurement precision, namely the determination of the signal incoming wave direction, depends on the number of signal sources. For example, the number of known signal sources is needed in the DML (Maximum likelihood estimation ) algorithm, so that the incoming wave direction can be calculated, and the error estimation of the number of signal sources seriously affects the angular resolution of the DML algorithm.
The current algorithm for estimating the incoming wave direction generally carries out Hilbert transformation on the radar array received signals to obtain instantaneous phase components, so that a covariance matrix is constructed according to the instantaneous phase components, a characteristic value is obtained, and then the number of signal sources is obtained according to the characteristic value and a pre-trained model.
However, in the above method, the feature decomposition of the matrix is required, the operation amount is large, and under the condition that the number of snapshots of the signal is small, the covariance matrix estimation error is large, so that the deviation of the source data estimation is large.
Disclosure of Invention
In view of the above, it is necessary to provide a signal source number calculation method, a first model training method, and an echo signal processing method that can ensure calculation accuracy.
In a first aspect, the present application provides a signal source number calculation method, including:
obtaining K incoming wave directions corresponding to echo signals, wherein K is more than or equal to 1;
according to the K incoming wave directions, calculating to obtain first model features;
and processing the first model characteristic according to a first model, and determining the number of signal sources corresponding to the echo signals.
In one embodiment, the acquiring K incoming wave directions corresponding to the echo signals includes:
And estimating the incoming wave directions of the echo signals to obtain the K incoming wave directions.
In one embodiment, the estimating the incoming wave directions of the echo signals to obtain K incoming wave directions includes:
acquiring the estimated quantity of the incoming wave directions corresponding to the echo signals;
and estimating the incoming wave direction of the echo signals according to the estimated quantity, and obtaining the incoming wave direction corresponding to the estimated quantity.
In one embodiment, the obtaining the estimated number of the incoming wave directions corresponding to the echo signals includes:
acquiring a current application scene, and acquiring a corresponding first model according to the current application scene;
reading input parameters of the first model;
and determining the estimated quantity of the target incoming wave direction according to the input parameters.
In one embodiment, the first model feature comprises: single target power ratio and/or multiple target power ratio; the single target power ratio is calculated according to the energy of the signal source corresponding to each incoming wave direction and the energy of the echo signal to obtain the single target power ratio corresponding to each incoming wave direction; the multi-target power ratio is calculated according to at least two incoming wave directions to obtain comprehensive energy, and is calculated according to the comprehensive energy and the energy of the echo signal.
In one embodiment, the method for calculating the number of signal sources further includes:
in the case where K includes K1 and K1 is equal to 1, the K incoming wave directions correspond to a first incoming wave direction, and a first feature W (θ 0 ) The first characteristic W (θ 0 ) For the single target power ratio; and/or
In the case where K includes K2 and K2 is equal to 2, the K incoming wave directions include a second incoming wave direction and a third incoming wave direction, according to the second incoming wave directionThe wave direction and the third incoming wave direction are calculated to obtain a second characteristic W (theta 1 ) Third feature W (θ 2 ) Fourth feature W (θ 1 θ 2 ) The second characteristic W (θ 1 ) And the third feature W (θ 2 ) For the single target power ratio, the fourth characteristic W (θ 1 θ 2 ) For the multi-target power ratio.
In one embodiment, the processing the first model feature according to a first model to determine the number of signal sources corresponding to the echo signal includes:
and calculating the number of signal sources corresponding to the echo signals according to at least any two characteristics in the first model based on the first model.
In a second aspect, the present application further provides a training method of the first model, the training method including:
Acquiring training data, wherein the training data comprises training echo signals and the number of corresponding training signal sources;
calculating according to the training echo signals to obtain training characteristics;
respectively selecting at least two training features and the corresponding training signal source numbers to generate at least one training set;
and training according to at least two simulation features in the training set and the corresponding simulation signal source numbers to obtain at least one first model.
In one embodiment, after training at least two simulation features in the training set and the corresponding number of simulation signal sources according to the at least one first model to obtain a first model, the method includes:
acquiring actual measurement echo signals and the corresponding actual measurement signal source numbers under a set application scene;
processing according to the actually measured echo signals and at least one first model to obtain the corresponding number of model signal sources;
comparing the number of the model signal sources with the number of the corresponding actually measured signal sources to obtain the accuracy of each first model;
and selecting the first model with the accuracy meeting the requirement as the first model corresponding to the application scene.
In one embodiment, after training according to the training set to obtain at least one first model, the method includes:
acquiring actual measurement echo signals and the corresponding actual measurement signal source numbers under a set application scene;
and optimizing the first model according to the acquired actual measurement echo signals and the corresponding actual measurement signal source numbers.
In a third aspect, the present application further provides an echo signal processing method, including:
receiving an echo signal;
and processing the echo signals according to the signal source number calculation method to obtain the signal source number.
In a fourth aspect, the present application further provides a signal source number calculating device, where the signal source number calculating device is configured to perform the above signal source number calculating method.
In a fifth aspect, the present application further provides a training device of the first model, where the training device is configured to perform the above method for calculating the number of signal sources.
In a sixth aspect, the present application further provides an echo signal processing device, where the echo signal processing device is configured to perform the echo signal processing method described above.
In a seventh aspect, the present application further provides a processing device, including a memory and a processor, the memory storing a computer program, the processor implementing the method according to any one of the embodiments described above when executing the computer program.
In an eighth aspect, the present application also provides a computer readable storage medium having stored thereon a computer program comprising instructions for implementing the method described in any one of the embodiments above.
In a ninth aspect, the present application also provides a computer program product comprising a computer program comprising instructions for implementing the method as described in any one of the embodiments above.
In a tenth aspect, the present application further provides an operation control device comprising a memory and a processor, the memory storing a computer program which, when run on the processor, causes the operation control device or the processor to implement the method as described in any one of the embodiments above.
In an eleventh aspect, the present application also provides an integrated circuit comprising a digital circuit and an operation control device as described above for controlling the digital circuit to implement different functions of the integrated circuit
In a twelfth aspect, the present application also provides a radio device comprising:
a carrier;
an integrated circuit as described above disposed on the carrier;
the antenna is arranged on the supporting body or integrated with the integrated circuit into a whole;
The integrated circuit is connected with the antenna and is used for transmitting and receiving radio signals.
In one embodiment, the radio device is a radar or radar chip.
In a thirteenth aspect, the present application further provides a terminal, including:
a radio device as described above;
wherein the radio device is used for target detection and/or communication.
In one embodiment, the terminal is a vehicle, smart home device, or robot.
According to the signal source number calculation method, the first model training method and the echo signal processing method, firstly, based on an algorithm such as DML, the number of the incoming wave directions of echo signals is estimated to be K, so that calculation is simplified, calculated amount is reduced, compared with the calculation of instantaneous phase components in the prior art, the calculation amount for obtaining characteristic values is greatly reduced by constructing a covariance matrix according to the instantaneous phase components, the number of array elements and the number of the signal sources are not limited, and the application range is expanded. And secondly, calculating to obtain first model features according to the estimated incoming wave directions, and determining a first model through the first model features, so that the number of real signal sources corresponding to the first model features is determined through processing of the first model, accuracy is improved, and the calculated incoming wave directions are accurate on the premise that the number of the signal sources is accurate.
Drawings
FIG. 1 is a diagram of an application environment for a method of signal source number calculation in one embodiment;
FIG. 2 is a flow chart of a method for calculating the number of signal sources in one embodiment;
FIG. 3 is a flow diagram of a training process for a first model in one embodiment;
FIG. 4 is a diagram of one embodiment of selecting training features W (θ 1 )、W(θ 2 ) A corresponding relation graph of the number of the signal sources;
FIG. 5 is a diagram of one embodiment of selecting training features W (θ 0 ) And W (theta) 1 ,θ 2 ) A corresponding relation graph of the number of the signal sources;
FIG. 6 shows the use of measured characteristic W (θ 1 )、W(θ 2 ) Statistical profiles of time single/multiple targets;
FIG. 7 shows the use of measured characteristic W (θ 0 ) And W (theta) 1 ,θ 2 ) Statistical profiles of time single/multiple targets;
FIG. 8 is a flowchart of a method for calculating the number of signal sources according to another embodiment;
FIG. 9 is a flow chart of a method for calculating an incoming wave direction in one embodiment;
FIG. 10 is a flow chart of an echo signal processing method according to an embodiment;
FIG. 11 is a block diagram of a signal source number calculation device in one embodiment;
FIG. 12 is a block diagram of an incoming wave direction calculation device in one embodiment;
FIG. 13 is a block diagram of an echo signal processing device in one embodiment;
fig. 14 is an internal structural diagram of a computer device in one embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
The signal source number calculation method, the incoming wave direction calculation method and the echo signal processing method provided by the application can be applied to an application environment shown in fig. 1. Wherein the radar 102 sends a signal to the target 104, so that the target 104 reflects an echo signal to the radar 102, and the radar 102 processes the echo signal reflected by the target 104 to determine the distance, speed and angle of the target 104, thereby realizing the positioning of the target. Wherein the radar may also be referred to as a radar transceiver. The positioning method can be applied to the fields of automatic driving, security protection, unmanned aerial vehicle and the like. Specifically, after the radar 102 receives the echo signal, analog-to-digital conversion (AD) is performed on the echo signal, and then the obtained digital signal is sampled, so that one-dimensional fast fourier transform (1 DFFT) and two-dimensional fast fourier transform (2 DFFT) are sequentially performed on the sampled digital signal to accumulate the echo signal and obtain the doppler frequency of the target 104, and then radar constant false alarm detection (CFAR) is performed to determine whether the echo signal and noise exist or not, so that the target 104 with the same distance and speed is obtained, and then the target 104 with the same distance and speed is processed to obtain the number of targets 104, that is, the number of signal sources, so that the subsequent calculation of the direction of the incoming wave (DOA) is facilitated.
The signal source number calculation method provided by the application can optionally process the radar constant false alarm detection result, wherein the radar constant false alarm detection result may have the condition that a plurality of targets 104 are identified as one target, that is, the distances and the speeds of the plurality of targets and the radar are equal, so that the number of targets, that is, the number of signal sources, needs to be determined, and the accuracy of the subsequent wave direction calculation is ensured.
The method comprises the steps of taking a result of radar constant false alarm detection as an example, after obtaining a radar constant false alarm detection result, namely an echo signal, firstly estimating incoming wave directions to obtain K incoming wave directions, wherein K is more than or equal to 1, calculating according to the K incoming wave directions to obtain first model features, and finally inputting the first model features into a first model obtained through pre-training to obtain the number of target signal sources corresponding to the echo signal. The operation only needs to calculate the incoming wave direction under the preset condition, so that the calculation is simplified, the calculated amount is reduced, compared with the calculation of the instantaneous phase component in the prior art, the covariance matrix is constructed according to the instantaneous phase component, the calculated amount of the obtained characteristic value is greatly reduced, the number of array elements and information sources is not limited, and the application range is expanded. The radar 102 may be installed on any terminal, and the terminal may be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, vehicle-mounted systems, unmanned aerial vehicles, portable wearable devices, and the like.
In one embodiment, as shown in fig. 2, a method for calculating the number of signal sources is provided, and the method is applied to the radar transceiver in fig. 1 for illustration, and includes the following steps:
s202: k incoming wave directions corresponding to the echo signals are obtained.
In one embodiment, acquiring K incoming wave directions corresponding to the echo signals includes: and estimating the incoming wave directions of the echo signals to obtain the K incoming wave directions. Specifically, echo signals are acquired first, then incoming wave direction estimation is carried out on the echo signals, and the K incoming wave directions are obtained.
In particular, the echo signal, i.e. the signal received by the radar transceiver, may, in combination with the above, alternatively be a signal obtained after detection by radar constant false alarm, wherein the signal may distinguish between objects of different distance and/or speed from the radar system, whereas objects of the same distance and speed from the radar system may not be distinguished, thus in order to reduce misidentification of objects of the same distance and speedThe probability of the same object requires the calculation of the number of objects. Specifically, the echo signal is X (t) = [ X 1 (t),x 2 (t),...,x M (t)]Wherein M is the number of antenna virtual units, wherein:
X(t)=A(θ)S(t)+N(t) (1)
S (t) is a D×l dimension vector of the spatial signal, A (θ) is a spatial array flow pattern matrix of M×D dimension, D is the number of signal sources, and N (t) is noise.
cov[N(t)]=σ 2 I (2)
Where σ represents the power ratio of the noise and I is the diagonal matrix.
And then estimating the incoming wave directions of the echo signals according to a first algorithm to obtain the K incoming wave directions.
The first algorithm may be a DML algorithm, and in other embodiments, other algorithms may be used to estimate the directions of the incoming waves to obtain K directions of the incoming waves.
Specifically, the incoming wave direction is estimated according to the echo signal, the incoming wave direction is not necessarily the same as the incoming wave direction calculated by the subsequent DOA, only the estimation process is performed here, for example, when the estimated number of incoming wave directions is K, the processing is performed according to the echo signal to obtain K incoming wave directions, wherein the estimation of the estimated number of incoming wave directions can be performed according to the practical application scenario, for example, 1 or 2, without specific limitation, and in one embodiment, there may be multiple estimates of the incoming wave directions at the same time, that is, in one calculation, the estimated incoming wave direction is K at the same time 1 ……K q And q is more than or equal to 1, and the estimated quantity of each incoming wave direction is calculated to obtain the incoming wave direction with the corresponding estimated quantity so as to facilitate the calculation of the subsequent first model characteristics.
Taking the first algorithm as a DML algorithm as an example after determining the estimated number of the incoming wave directions, the radar transceiver can calculate the incoming wave directions by using the following formula:
wherein,for incoming wave direction estimation, P A(θ) For the orthographic projection matrix for A (θ), tr () represents the tracing matrix, ++>Representing an estimate of the covariance matrix.
Wherein, the orthogonal projection matrix is:
P A(θ) =A(θ)(A H (θ)A(θ)) -1 A H (θ) (4)
the covariance matrix is:
wherein the method comprises the steps ofIs X i Transposed conjugate matrix of (t).
Thus, the radar transceiver can simplify the solution of equation (3) after estimating the estimated number of incoming wave directions to obtain the incoming wave directions, wherein, for example, when the estimated number of incoming wave directions is 1, equation (3) is directly calculated to obtain an incoming wave direction, such as θ 0 When the estimated number of incoming wave directions is 2, directly solving the formula (3) to obtain two incoming wave directions, such as θ 1 θ 2 And so on.
Furthermore, the estimated number of incoming wave directions may be related to the selected first model, for example, to input parameters of the first model, the estimated number of incoming wave directions is determined from the input parameters, and the incoming wave directions corresponding to the estimated number are calculated. For example, when the estimated number of incoming wave directions of the input parameters is 1 and 2, it is necessary to find the incoming wave directions when the estimated number of incoming wave directions is 1 and 2, respectively. When the number of estimated incoming wave directions required for the calculation of the input parameters is 2, then only two incoming wave directions when the number of estimated incoming wave directions is 2 need to be calculated.
In other embodiments, if the obtained first model is only 1, the number of incoming wave directions may be directly fixed, for example, 1 or 2, so that the estimated number of incoming wave directions is directly obtained and calculated during calculation.
S204: and calculating to obtain a first model feature according to the K incoming wave directions.
Specifically, the first model feature is associated with an estimated direction of incoming waves, from which the first model feature is calculated, i.e. the direction of the estimated number of signal sources. The radar transceiver calculates first model features corresponding to different estimated numbers of incoming wave directions respectively. And under the condition that the number of the signal sources is the estimated number, respectively calculating a single target power ratio corresponding to each signal source corresponding to the echo signal and/or a multi-target power ratio corresponding to at least two signal sources as first model features, and further judging the actual number of the signal sources according to the first model features. Wherein the first model feature may be a feature for characterizing a power ratio of the signal source. Specifically, the first model feature comprises: single target power ratio and/or multiple target power ratio; the single target power ratio is calculated according to the energy of the signal source corresponding to each incoming wave direction and the energy of the echo signal to obtain the single target power ratio corresponding to each incoming wave direction; the multi-target power ratio is calculated according to at least two incoming wave directions to obtain comprehensive energy, and is calculated according to the comprehensive energy and the energy of the echo signal.
Specifically, in the case where K includes K1 and K1 is equal to 1, the K incoming wave directions correspond to first incoming wave directions from which first characteristics (W (θ 0 ) And a first feature (W (θ) 0 ) A single target power ratio; and/or
In the case that K includes K2 and K2 is equal to 2, the K incoming wave directions include a second incoming wave direction and a third incoming wave direction, and the second feature is calculated according to the second incoming wave direction and the third incoming wave directionSign W (θ) 1 ) Third feature W (θ 2 ) Fourth feature W (θ 1 θ 2 ). Second characteristic W (θ) 1 ) Third feature w (θ 2 ) For a single target power ratio, the fourth characteristic W (θ 1 θ 2 ) Is a multi-target power ratio.
Alternatively, where K may also include K3, and K3 is greater than 2, other features may be calculated as needed.
In one embodiment, two estimated numbers are taken as an example, one estimated number is 1 in the incoming wave direction, the other estimated number is 2 in the incoming wave direction, and in other embodiments, there may be other number of assumptions about the incoming wave direction, and no specific limitation is made herein.
Wherein when the number of the estimated incoming wave directions is 2, that is, the number of the estimated targets is two, that is, the number of the estimated signal sources is 2, the two-dimensional search is performed according to the above formula (4) to obtain the estimated incoming wave direction θ 1 ,θ 2 Then define object 1, i.e. incoming wave direction θ 1 The energy of (2) is:
P(θ 1 )=A(θ 1 )(A H (θ 1 )A(θ 1 )) -1 A H (θ 1 )X i (t) (7)
defining object 1, i.e. incoming wave direction θ 1 Is a single target power ratio of the second characteristic W (θ 1 ):
Defining object 2, i.e. incoming wave direction θ 2 Energy of (2):
P(θ 2 )=A(θ 2 )(A H (θ 2 )A(θ 2 )) -1 A H (θ 2 )X i (t) (9)
target 2, i.e. the direction of incoming waves is θ 2 Is a single target power ratio of the third characteristic W (θ 2 ):
Furthermore, the integrated energy of both target 1 and target 2 is defined:
P(θ 1 ,θ 2 )=A(θ 1 ,θ 2 )(A H (θ 1 ,θ 2 )A(θ 1 ,θ 2 )) -1 A H (θ 1 ,θ 2 )X i (t) (11)
defining a multiple target power ratio where target 1 and target 2 are present at the same time, i.e. a fourth characteristic W (θ 1 ,θ 2 ) The method comprises the following steps:
wherein when the number of the estimated incoming wave directions is 1, that is, the estimated target is one, that is, the estimated signal source is one, the incoming wave direction θ is obtained by performing the two-dimensional search according to the above formula (4) 0 Then define target 0, i.e. incoming wave direction θ 0 The energy of (2) is:
P(θ 0 )=A(θ 0 )(A H (θ 0 )A(θ 0 )) -1 A H (θ 0 )X i (t) (13)
target 0, a first characteristic W (θ 0 ):
Thus, in the case where the estimated number includes 1 and 2, the resulting feature includes W (θ 0 )、W(θ 1 )、W(θ 2 ) And W (theta) 1 ,θ 2 ) At least 2 of the features are selected as first model features.
In one embodiment, the radar transceiver does not need to calculate all the first model features at a time, and the radar transceiver can determine the corresponding first model features according to the determined input parameters of the first model and then calculate the corresponding first model features, so that the calculation amount of the first model features can be reduced. In other embodiments, the radar transceiver is required to calculate all of the first model features for input into the first model for calculation.
S206: and processing the first model characteristic according to a first model, and determining the number of signal sources corresponding to the echo signals.
In particular, the first model may be various types of neural network models, preferably an SVM (support vector machine) model. Wherein the first model may be a default model or a model selected according to the application scenario. The input of the first model is the first model characteristic, and the output is the number of target signal sources. And the radar transceiver inputs the calculated first model characteristics into a first model which is obtained through pre-training, so as to obtain the number of target signal sources corresponding to the echo signals.
In one embodiment, the same first model may be used in different scenarios, that is, only one first model is trained during training, or a plurality of first models are trained, but a default first model is selected in different scenarios, so that the model input parameters are fixed, the types of the first model features are fixed, therefore, the estimated number of the incoming wave directions can be determined directly according to the first model features of the preset type, then the incoming wave directions corresponding to each estimated number are calculated and obtained, finally the first model features are calculated according to each calculated incoming wave direction, and the corresponding target signal source number is obtained by processing the first model features through the first models.
In one embodiment, different scenes correspond to different first models, that is, a plurality of first models are trained according to the scenes during training, so that when the method is used, corresponding first model features are determined according to application scenes, corresponding first models are determined according to the first model features, the estimated number of incoming wave directions is determined according to the first models, the incoming wave directions corresponding to each estimated number are calculated, the first model features are calculated according to each calculated incoming wave direction, and the corresponding target signal source number is obtained by processing the first model features through the first models.
For easy understanding, for example, when the estimated number of incoming wave directions is 1, then the first model 1 is corresponded; when the estimated number of incoming wave directions is 1 and 2, then the first model 2 is corresponded. In other embodiments, other estimated numbers of the incoming wave directions also correspond to the unique first models, so that the calculated first model features are input into the corresponding first models to calculate to obtain the corresponding target signal source numbers.
In the above embodiment, the K incoming wave directions are estimated first, so that the calculation is simplified, and the calculated amount is reduced, compared with the calculation of the instantaneous phase components in the prior art, the calculation amount of obtaining the eigenvalues is greatly reduced by constructing the covariance matrix according to the instantaneous phase components, the application range is expanded without limitation on the number of array elements and information sources, the first model features are obtained by calculation according to the estimated incoming wave directions, and the real number of the signal sources corresponding to the first model features is determined by processing through the first model, so that the accuracy is improved, and the calculated incoming wave directions are also accurate on the premise that the number of the signal sources is accurate.
In one embodiment, estimating an incoming wave direction according to an echo signal to obtain at least one incoming wave direction includes: acquiring at least one estimated quantity of the incoming wave direction; and estimating the incoming wave directions of the echo signals according to the estimated quantity, and obtaining the incoming wave directions corresponding to the estimated quantity.
The estimated number refers to the estimated number of the incoming wave directions, and when the number of the target signal sources is calculated each time, the number of the target signal sources is estimated first, that is, the incoming wave directions are estimated, and the estimated number is the estimated number of the incoming wave directions. The estimate may be determined in relation to the input parameters of the first model, i.e. the number of directions of incoming waves needed for calculation of the first model features corresponding to the input parameters of the model, said first model features being W (θ 0 )、W(θ 1 )、W(θ 2 ) And W (theta) 1 ,θ 2 ) At least two of them are exemplified byIllustratively, when the first model feature is W (θ 1 )、W(θ 2 ) The corresponding estimated number is 2, i.e. the incoming wave direction is 2. When the first model feature is W (θ 0 )、W(θ 1 )、W(θ 2 ) And W (theta) 1 ,θ 2 ) The corresponding estimated numbers are 1 and 2, i.e. there are 1 and 2 incoming wave directions. When the first model feature is W (θ 0 ) And W (theta) 1 ,θ 2 ) The corresponding estimated numbers are 1 and 2, i.e. there are 1 and 2 incoming wave directions. When the first model feature is another feature, the direction of the incoming wave may be determined according to the feature, which is not limited in this embodiment.
Specifically, after the estimated number of the incoming wave directions is determined, the incoming wave directions of the echo signals are estimated according to the estimated number, so as to obtain incoming wave directions corresponding to the estimated number, that is, the above formula (3) is solved according to the estimated number of the incoming wave directions, so as to obtain corresponding incoming wave directions respectively, for example, when the estimated number is 1, the formula (3) is solved so as to obtain 1 incoming wave direction. And when the estimated number is 2, solving the formula (3) to obtain 2 incoming wave directions. In other embodiments, when the estimated number is other values, the equation (3) may be solved to obtain the directions of incoming waves corresponding to other numbers.
In the above embodiment, the estimated number of the incoming wave directions is obtained first, and then the echo signals are processed, so that the operation can be simplified, and the calculation amount can be reduced.
In one embodiment, obtaining the estimated number of incoming wave directions corresponding to the echo signals includes: acquiring a current application scene, and acquiring a corresponding first model according to the current application scene; reading input parameters of the first model; and determining the estimated quantity of the target incoming wave direction according to the input parameters.
Specifically, in this embodiment, the first model corresponds to an application scenario, that is, each application scenario corresponds to a first model, so during training, the first model may be first trained by training data, and then obtained by respectively optimizing measured data in the corresponding scenario, so as to ensure accuracy of the first model.
When the first model changes according to the application scene, one scheme is to download the corresponding first model according to the application scene in advance and install the corresponding first model, so as to determine the corresponding first model. And the other is that all the first models are downloaded and then selected in real time according to the application scene. Therefore, the step of acquiring the current application scene and acquiring the corresponding first model according to the current application scene in this embodiment may include the two schemes.
After the first model is determined, the input parameters of the first model, i.e. the type of the first model features, are determined, so that the estimated number of corresponding incoming wave directions can be determined according to the type of the first model features, as described above.
In the above embodiment, the corresponding first model is determined according to the application scenario, so that the corresponding first model feature can be determined according to the input parameter of the first model, the number of the incoming wave directions is estimated, and then the echo signals are processed, so that the operation can be simplified, and the calculation amount is reduced.
In one embodiment, the calculating according to the incoming wave direction obtains a first model feature, including: and calculating according to the incoming wave direction to obtain first model features corresponding to the input parameters of the first model.
Specifically, in this embodiment, in order to reduce the amount of calculation, the first model feature to be calculated may be determined according to the type of the model input parameter, for example, the first model feature includes W (θ 1 )、W(θ 2 ) In this case, only the two features need to be calculated, and other features do not need to be calculated, so that the calculation amount is reduced.
In the above embodiment, the calculation of the features of the first model is performed according to the input parameters of the first model, so that no additional features need to be calculated, and the calculation amount is reduced.
The foregoing mainly describes the use process of the first model, and the following describes the training process of the first model in detail, and referring specifically to fig. 3, fig. 3 is a flowchart of the training process of the first model in an embodiment, where the training method of the first model obtained by training in advance includes:
s302: training data is acquired, wherein the training data comprises training echo signals and the corresponding number of training signal sources.
Specifically, the training data includes simulation data and measured data, wherein the simulation data is obtained by respectively simulating combinations of different angles/different signal-to-noise ratios by setting angles of the incoming wave direction and the signal-to-noise ratio of the target, wherein the positions and speeds of the targets of the training data are the same, and thus only the different angles/different signal-to-noise ratios are defined herein. It is emphasized that the angle of the targets of the different sources must be different and the signal to noise ratio may be the same. The actual measurement data are data obtained in the actual use process of the radar receiving and transmitting system.
In practical application, it can be assumed that the incoming wave direction angle of a signal source, namely a target, is theta_int= [ -65:1:65 ] ° (the angle range is-65 DEG to 65 DEG, and the interval is 1 DEG); the signal-to-noise ratio snr_int= [ 10:1:22 ] dB (signal-to-noise ratio range 10-22 dB, interval 1 dB) of the target is respectively simulated by combining different angles and signal-to-noise ratios to obtain the original simulation data of the radar. For example, when the incoming wave directions are 1, the angle can be determined to be-65 ° to 65 °, the interval is 1 °, so that 121 incoming wave directions exist, each incoming wave direction corresponds to a signal to noise ratio range of 10-22 dB, the interval is 1dB, so that 13 different signal to noise ratios exist, when the incoming wave directions are 1, 121×13 combinations exist, and similarly, when the incoming wave directions are 2, the number of combinations can be calculated, and each combination corresponds to one simulation data. I.e. each simulation data defines a simulation radar echo signal by incoming wave direction and signal-to-noise ratio. The number of incoming wave directions is the number of simulated signal sources.
S304: and calculating according to the training echo signals to obtain training characteristics.
In particular, the specific definition of the training features is similar to the first model features, which include the power ratio features of the echo signals of each signal source. Taking the example above as an example, the training features of each training data combination include W (θ 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) Features, i.e. if there is only one estimated number of incoming wave directions, the training features are W (θ 0 ) If there are two estimated numbers of incoming wave directions, the training feature is W (θ 1 )、 W(θ 2 ) And W (theta) 1 ,θ 2 ) If there are one and two estimated number of incoming wave directions, the training feature is W (θ 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 )。
S306: and respectively selecting at least two training features and the corresponding training signal source number to generate at least one training set.
S308: and training according to the training set to obtain at least one first model.
In particular, the training set is selected in this case in relation to the first model to be trained, i.e. the different model input data corresponds to different first models, so that the differences in the training characteristics selected during training lead to different first models.
Therefore, at least two training features and the number of training signal sources of training data corresponding to the training features can be selected as required during training to generate at least one training set, for example, the training features W (theta 1 ) And W (theta) 2 ) Model training, and in particular discrimination of the trained model, may be found in fig. 4 and table 1, where training features W (θ 1 ) And W (theta) 2 ) With a distinct degree of differentiation, where the straight line is the use of W (θ 1 )、W(θ 2 ) The feature classifies the single/multiple objects as a boundary. Table 1 shows the characteristic W (θ 1 )、W(θ 2 ) Statistics of the accuracy of (2).
Table 1 training characteristics are W (θ) 1 )、W(θ 2 ) Discrimination probability at the time
In other embodiments, the training features W (θ 0 ) And W (theta) 1 ,θ 2 ) Model training, and in particular discrimination of the trained model, may be found in fig. 5 and table 2, where training features W (θ 0 ) And W (theta) 1 ,θ 2 ) With obvious differentiation, where the straight line is the use of the SVM with W (θ 0 ) And W (theta) 1 ,θ 2 ) The approach feature classifies the single/multiple objects at the boundary. Table 2 shows the characteristic W (θ 0 ) And W (theta) 1 ,θ 2 ) Statistics of the accuracy.
Table 2 training characteristics are W (θ) 0 ) And W (theta) 1 ,θ 2 ) Discrimination probability at the time
In other embodiments, the training features W (θ 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) Model training, and in particular discrimination of the trained model, may be found in Table 3, where training characteristics W (θ 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) With obvious differentiation, where the straight line is the use of the SVM with W (θ 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) The approach feature classifies the single/multiple objects at the boundary. Table 3 shows the training characteristics W (θ) 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) Statistics of the accuracy.
Table 3 training characteristics are W (θ) 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) Discrimination probability at the time
Table 4 discriminant probability using AIC criteria
Simulation results show that: four training features W (θ) 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) The highest accuracy of the information source number estimation can reach 99.653%, and training characteristics W (theta 1 )、W(θ 2 ) The accuracy was the lowest, as little as 93.879%, but training features W (θ 0 ) And W (theta) 1 ,θ 2 ) When the information source number is estimated, the accuracy is only 0.137% lower than that when four characteristics are adopted, and when the information source number is estimated by the traditional AIC algorithm, the accuracy is only 70.8596% which is far lower than that of the algorithm proposed by the embodiment.
In the above embodiment, a training method of the first model is provided, and training data is obtained through simulation, so that the sufficiency of the data is ensured, and then training is performed, thereby ensuring the accuracy of the trained model.
In one embodiment, after training according to the training set to obtain at least one first model, the method includes: acquiring actual measurement radar array receiving signals and the corresponding actual measurement signal source numbers under a set application scene; processing according to the actually measured radar array receiving signals and at least one first model to obtain the corresponding model signal source number; comparing the number of the model signal sources with the number of the corresponding actually measured signal sources to obtain the accuracy of each first model; and selecting the first model with the accuracy meeting the requirement as the first model corresponding to the application scene.
Specifically, in this embodiment, in order to further ensure the validity of the first model, the model is verified by actually measured echo signals, where actually measured echo signals under a set application scenario, for example, actually measured radar array receiving signals, are first obtained, actually measured features are calculated according to the actually measured radar receiving signals, where the actually measured features are similar to the training features described above, only the data values are different, then the actually measured features are input into the corresponding first model to process to obtain the corresponding number of model signal sources, and the accuracy of the first model is obtained by comparing the number of model signal sources with the corresponding number of actually measured signal sources, so that the first model with the highest accuracy is selected as the first model corresponding to the application scenario.
In particular, reference can be made to fig. 6 and 7, wherein fig. 6 is a graph of the measured characteristic W (θ 1 )、W(θ 2 ) Statistical profiles of time single/multiple targets; FIG. 7 shows the use of measured characteristic W (θ 0 ) And W (theta) 1 ,θ 2 ) Statistical profiles of time single/multiple targets; table 5 shows the values when W (θ 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) And when the actual measurement characteristics are obtained according to the actual measurement echo signals, classifying and comparing the accuracy statistics results of the traditional AIC method according to the actual measurement characteristics.
Table 5 training characteristics are W (θ) 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) Discrimination probability at the time
Experimental results show that the number of the information sources can be effectively estimated by the four extracted features in an actual environment, wherein the highest accuracy can reach 93.724% under the condition of adopting the four features, and compared with the traditional AIC method, the method has a better improvement result. Thus, a first model trained by four features can be selected as the first model corresponding to the application scenario.
In the above embodiment, the accuracy of the first model is evaluated by actually measuring the received signal of the radar array, so that the signal source data calculation model with the highest accuracy is selected for subsequent calculation, the accuracy can be improved, the robustness of the information source number estimation can be improved, and the method can work in various environments.
In one embodiment, after training at least two training features in the training set and the corresponding number of training signal sources to obtain a first model, the method includes: acquiring actual measurement echo signals and the corresponding actual measurement signal source numbers under a set application scene; and optimizing the first model according to the acquired actual measurement echo signals and the corresponding actual measurement signal source numbers.
Specifically, in this embodiment, since the first model is trained according to training data, in order to improve accuracy of the first model, in this embodiment, the first model is optimized by using actual measurement echo signals under a set application scenario, for example, actual measurement radar array receiving signals and corresponding actual measurement signal source numbers, where an optimization process is similar to a training process, that is, corresponding features are calculated according to actual measurement radar array receiving signals first, and then optimization training is performed according to the features and corresponding actual measurement signal source numbers, so as to ensure adaptability of the first model to the scenario.
In addition, in one embodiment, the actual measurement radar array receiving signals and the corresponding actual measurement signal source numbers can be periodically acquired, so that each first model is periodically optimized, and the first model with the highest accuracy rate is selected as the first model under the scene to be used after the optimization.
In one embodiment, the first model may be selected according to the requirement of the accuracy and the requirement of the calculation amount in the corresponding application scenario, and only the first model with the highest accuracy need not be selected, for example, the first model with the next highest accuracy but with smaller calculation amount, for example, only the first model with 2 features need to be calculated for actual processing.
In the above embodiment, the accuracy of the first model is further improved by optimizing the first model through the actually measured radar array receiving signals and the corresponding actually measured signal source numbers.
In one embodiment, referring to fig. 8, fig. 8 is a flowchart of a signal source number calculating method in another embodiment, in this embodiment, training data is first obtained, and then 4 training features are calculated for the training data: w (θ) 1 )、W(θ 2 )、W(θ 0 ) And W (theta) 1 ,θ 2 ) Then respectively selecting at least two characteristicsTraining the first models to obtain a plurality of first models. And optimizing the received signals of the actual measurement radar array and the corresponding actual measurement signal source numbers according to the application scene, selecting a first model with the accuracy and/or calculated amount meeting the requirements after optimizing, and downloading and installing.
When the method is actually applied, echo signals are firstly obtained, then the estimated quantity of the incoming wave directions is determined according to the input parameters of the signal source data calculation model, the incoming wave directions corresponding to the estimated quantity are obtained through calculation according to the estimated quantity, the calculation of first model features is carried out according to the input parameters and the calculated incoming wave directions, and finally the first model features are input into the first model to obtain the number of signal sources.
In the above embodiment, the K incoming wave directions are estimated first, so that the calculation is simplified, and the calculated amount is reduced, compared with the calculation of the instantaneous phase components in the prior art, the calculation amount of obtaining the characteristic values is greatly reduced by constructing the covariance matrix according to the instantaneous phase components, the application range is expanded without limitation on the number of array elements and information sources, the first model features are obtained by calculating according to the estimated incoming wave directions, and the corresponding first models are determined by the first model features, so that the real number of signal sources corresponding to the first model features is determined by processing through the first models, the accuracy is improved, and the calculated incoming wave directions are accurate on the premise that the number of the signal sources is accurate.
In one embodiment, as shown in fig. 9, there is provided an incoming wave direction calculation method, which is described by taking the radar transceiver in fig. 1 as an example, and includes the following steps:
s902: an echo signal is acquired.
Specifically, the echo signal, that is, the signal after processing the echo signal received by the radar transceiver, is combined with the signal after radar constant false alarm detection, wherein the signal sources with different distances and/or speeds can be distinguished, and the signal sources with the same distances and speeds cannot be distinguished, so that in order to avoid mistaking the signal sources with the same distances and speeds as one signal source, the number of the signal sources needs to be calculated.
S904: and processing the echo signals according to the signal source number calculation method in any one of the embodiments to obtain the signal source number.
Specifically, the calculation of the number of signal sources may be specifically referred to above, and will not be described herein.
S906: and processing the echo signals according to the number of the signal sources to obtain the incoming wave direction.
Specifically, the calculation of the incoming wave direction according to the data of the signal source may be specifically performed according to the above formula (3), which is not described herein. In one embodiment, after the number of signal sources is determined, the corresponding estimated number of incoming wave directions can be directly obtained, for example, the number of signal sources is 1, the incoming wave directions calculated when the estimated number is 1 are obtained, and when the number of signal sources is 2, the incoming wave directions calculated when the estimated number is 2 are obtained, so that calculation according to the formula (3) is not needed, and the processing efficiency is improved.
In one embodiment, the processing the echo signal according to the method for calculating the number of signal sources in any one of the above embodiments to obtain the number of signal sources includes: when the number of the echo signals is greater than or equal to 2, the echo signals are processed sequentially according to the signal source number calculation method in any one embodiment to obtain the signal source number.
Specifically, in this embodiment, when the number of echo signals is greater than or equal to 2, that is, when the number of targets obtained after radar constant false alarm detection is at least 2, the echo signals are sequentially processed at this time to obtain the number of signal sources, that is, serially processed, so as to respectively calculate the number of corresponding signal sources. In other embodiments, the echo signals may also be processed in parallel to obtain the number of signal sources corresponding to each echo signal.
In the above embodiment, when the number of echo signals is greater than or equal to 2, serial processing or parallel processing is performed to calculate the number of corresponding signal sources respectively.
In one embodiment, as shown in fig. 10, there is provided an echo signal processing method, which is described by taking an example that the method is applied to the radar transceiver in fig. 1, and includes the following steps:
S1002: an echo signal is received.
Specifically, the echo signal may be an echo signal from radar 102 to target 104, such that target 104 is to radar 102.
S1004: the echo signals are processed.
Specifically, processing the echo signal herein includes, but is not limited to, first performing analog-to-digital conversion (AD) on the echo signal, then sampling the resulting digital signal, and thus sequentially performing one-dimensional fast fourier transform (1 DFFT) and two-dimensional fast fourier transform (2 DFFT) on the sampled digital signal to accumulate the radar echo signal and obtain the doppler frequency of the target 104, and then performing radar constant false alarm detection (CFAR) to discriminate the radar echo signal from noise to determine whether the target signal is present, thus obtaining the target 104 with the same distance and speed.
S1006: the method for calculating the incoming wave direction according to any one of the embodiments processes the echo signal to obtain the incoming wave direction.
Specifically, the calculation of the incoming wave direction in this step may be referred to above, and will not be described herein.
S1008: processing continues according to the incoming wave direction.
Specifically, after the incoming wave direction is calculated, the processing may be continued as needed, which is not particularly limited herein. The continuous processing comprises the steps of determining the incoming wave directions corresponding to the signal sources according to the number of the obtained signal sources, and further utilizing the incoming wave directions to detect targets.
In the above embodiment, the K incoming wave directions are estimated first, so that the calculation is simplified, and the calculated amount is reduced, compared with the calculation of the instantaneous phase components in the prior art, the calculation amount of obtaining the eigenvalues is greatly reduced by constructing the covariance matrix according to the instantaneous phase components, the application range is expanded without limitation on the number of array elements and information sources, the first model features are obtained by calculation according to the estimated incoming wave directions, and the real number of the signal sources corresponding to the first model features is determined by processing through the first model, so that the accuracy is improved, and the calculated incoming wave directions are also accurate on the premise that the number of the signal sources is accurate.
It should be understood that, although the steps in the flowcharts of fig. 2, 8, 9, and 10 are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least a portion of the steps of fig. 2, 8, 9, and 10 may include a plurality of steps or stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of execution of the steps or stages is not necessarily sequential, but may be performed in turn or alternately with at least a portion of the steps or stages of other steps or other steps.
In one embodiment, as shown in fig. 11, there is provided a signal source number calculating apparatus including: an incoming wave direction acquisition module 1202, a first model feature calculation module 1204, and a target signal source number calculation module 1206, wherein:
an incoming wave direction acquisition module 1202, configured to acquire K incoming wave directions corresponding to echo signals, where K is greater than or equal to 1;
the first model feature calculation module 1204 is configured to calculate a first model feature according to the K incoming wave directions;
and the target signal source number calculation module 1206 is configured to process the first model feature according to a first model, and determine the signal source number corresponding to the echo signal.
In one embodiment, the incoming wave direction obtaining module 1202 is configured to perform incoming wave direction estimation on the echo signal, so as to obtain the K incoming wave directions.
In one embodiment, the incoming wave direction acquisition module 1202 includes:
an estimated quantity obtaining unit, configured to obtain an estimated quantity of an incoming wave direction corresponding to the echo signal;
and the incoming wave direction calculation unit is used for estimating the incoming wave direction of the echo signals according to the estimated quantity and obtaining the incoming wave direction corresponding to the estimated quantity.
In one embodiment, the estimated quantity acquisition unit includes:
the model acquisition subunit is used for acquiring a current application scene and acquiring a corresponding first model according to the current application scene;
a parameter reading subunit, configured to read an input parameter of the first model;
and the estimated quantity determination subunit is used for determining the estimated quantity of the target incoming wave direction according to the input parameters.
In one embodiment, the first model feature comprises: single target power ratio and/or multiple target power ratio; the single target power ratio is calculated according to the energy of the signal source corresponding to each incoming wave direction and the energy of the echo signal to obtain the single target power ratio corresponding to each incoming wave direction; the multi-target power ratio is calculated according to at least two incoming wave directions to obtain comprehensive energy, and is calculated according to the comprehensive energy and the energy of the echo signal.
In one embodiment, in the case where K includes K1 and K1 is equal to 1, the K incoming wave directions correspond to a first incoming wave direction, and according to the first incoming wave direction, a first feature (W (θ 0 ) First feature W (θ) 0 ) For the single target power ratio; and/or
In the case where K includes K2 and K2 is equal to 2, the K incoming wave directions include a second incoming wave direction and a third incoming wave direction, and the second feature W (θ 1 ) Third feature W (θ 2 ) Fourth feature W (θ 1 θ 2 ) Second characteristic W (θ 1 ) And the third feature W (θ 2 ) For the single target power ratio, the fourth featureW(θ 1 θ 2 ) For the multi-target power ratio.
In one embodiment, the target signal source number calculation module 1206 includes: and calculating the number of signal sources corresponding to the echo signals based on at least any two characteristics in the first model characteristics.
In one embodiment, there is provided a training apparatus for a first model, the apparatus comprising:
the training data acquisition module is used for acquiring training data, wherein the training data comprises training echo signals and the corresponding number of training signal sources;
the training characteristic calculation module is used for calculating training characteristics according to the training echo signals;
the training set generation module is used for respectively selecting at least two training features and the corresponding training signal source number to generate at least one training set;
And the training module is used for training according to the training set to obtain at least one first model.
In one embodiment, the training device of the first model further includes:
the first measured data acquisition module is used for acquiring measured echo signals and the corresponding number of measured signal sources in a set application scene;
the model processing module is used for processing according to the actually measured echo signals and at least one first model to obtain the corresponding number of model signal sources;
the accuracy rate calculation module is used for comparing the number of the model signal sources with the number of the corresponding actually measured signal sources to obtain the accuracy rate of each first model;
the model selection module is used for selecting a first model with the accuracy meeting the requirement as a first model corresponding to the application scene.
In one embodiment, the training device of the first model further includes:
the first measured data acquisition module is used for acquiring measured echo signals and the corresponding number of measured signal sources in a set application scene;
and the optimization module is used for optimizing the first model according to the acquired actual measurement echo signals and the corresponding actual measurement signal source numbers.
In one embodiment, as shown in fig. 12, there is provided an incoming wave direction calculation apparatus including: a second signal acquisition module 1302, a signal source number calculation module 1304, and a first incoming wave direction calculation module 1306, wherein:
A second signal acquisition module 1302, configured to acquire an echo signal;
a signal source number calculating module 1304, configured to calculate the number of signal sources according to the signal source number calculating device in any one of the foregoing embodiments;
the first incoming wave direction calculation module 1306 is configured to process the echo signals according to the number of signal sources to obtain an incoming wave direction.
In one embodiment, the signal source number calculating module 1304 is configured to process the echo signals according to the signal source number calculating method in any one of the embodiments to obtain the signal source number when the number of echo signals is greater than or equal to 2.
In one embodiment, as shown in fig. 13, there is provided an echo signal processing device including: an echo signal receiving module 1402, a first signal processing module 1404, a second incoming wave direction calculating module 1406, and a second signal processing module 1408, wherein:
an echo signal receiving module 1402, configured to receive an echo signal;
a first signal processing module 1404, configured to process the echo signal to an echo signal;
a second incoming wave direction calculation module 1406, configured to process the echo signal according to the incoming wave direction calculation device in any one of the foregoing embodiments to obtain an incoming wave direction;
The second signal processing module 1408 is configured to continue processing according to the incoming wave direction.
The specific limitations of the signal source number calculation device, the training device of the first model, the incoming wave direction calculation device, and the echo signal processing device may be referred to above as limitations of the signal source number calculation method, the training method of the first model, the incoming wave direction calculation method, and the echo signal processing method, which are not described herein. The above-mentioned signal source number calculation device, training device of the first model, incoming wave direction calculation device, and echo signal processing device may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a terminal or an embedded structure in a terminal, and an internal structure diagram thereof may be as shown in fig. 14. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a signal source number calculation method, an incoming wave direction calculation method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the structure shown in fig. 14 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the computer device to which the present application applies, and that a particular computer device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, there is also provided a processing device including a memory and a processor, the memory storing a computer program including instructions for implementing the method embodiments described above.
In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program comprising instructions for implementing the method embodiments described above.
In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program comprising instructions for implementing the method embodiments described above.
In one embodiment, the present application also provides an integrated circuit comprising a digital circuit and an operation control device electrically connected to the digital circuit, the operation control device for controlling the digital circuit to implement different functions of the integrated circuit. Optionally, the integrated circuit further comprises a digital functional module, and the digital functional module is respectively in communication connection with the digital circuit and the operation control device; the digital function module is used for detecting whether the digital circuit is abnormal or not, and the operation control equipment is used for controlling the digital function module to work. The operation control device is configured to execute the steps in the signal source number calculation method, the incoming wave direction calculation method, and the echo signal processing method in any of the foregoing embodiments.
Specifically, in an integrated circuit, the integrated circuit comprises a digital circuit, a digital function module and an operation control device, wherein various digital circuits are basic components of the integrated circuit, different digital circuits can realize different functions of the integrated circuit, the digital function module is used for detecting whether each digital circuit works normally, the operation control device can perform unified configuration management on the digital function module, a digital controller in the operation control device can send control signals for performing function detection to the digital function module through a digital control interface, configuration information and state information are stored in the configuration module, the configuration information can be obtained from the outside, a state machine is used for controlling the working flow of the integrated circuit, the state machine can read the configuration information stored in the configuration module, and corresponding control signals are generated for controlling the digital controller to be output to the digital function module so as to realize the control of the digital function module to detect each digital circuit.
The integrated circuit adopts the unified digital controller to be connected with the digital function module of the system on chip through the digital control interface, and then realizes the unified configuration management of the running state of the digital function module of the system on chip through the configuration module and the state machine, thereby improving the running control efficiency of the system on chip in the integrated circuit.
Alternatively, in one embodiment, the integrated circuit may be a millimeter wave radar chip. The kind of digital functional modules in the integrated circuit can be determined according to the actual requirements. For example, in a millimeter wave radar chip, the digital function module may be a power ratio detector or the like, which may be used to detect whether the voltage value of the antenna power ratio amplifier is abnormal, and the operation control device may operate the power ratio detector.
In one embodiment, the present application also provides a radio device comprising: a carrier; the integrated circuit of the above embodiment is disposed on the carrier; the antenna is arranged on the carrier; the integrated circuit is connected with the antenna through a first transmission line and is used for receiving and transmitting radio signals. The carrier may be a printed circuit board PCB, and the first transmission line may be a PCB trace. Wherein the radio device is a radar chip or radar, preferably a millimeter wave radar.
In one embodiment, the present application further provides a terminal, including: the radio device of the above embodiment; wherein the radio is used for object detection and/or communication. In one embodiment, the terminal is a vehicle, smart home device, or robot.
Specifically, on the basis of the above-described embodiments, in one embodiment of the present application, the radio device may be disposed outside the terminal, in another embodiment of the present application, the radio device may also be disposed inside the terminal, and in other embodiments of the present application, the radio device may also be disposed partially inside the terminal, and partially outside the terminal. The present application is not limited thereto, and is specific to the case.
It should be noted that the radio device may perform functions such as object detection and communication by transmitting and receiving signals.
In an alternative embodiment, the terminal may be a component and a product applied in fields such as smart home, transportation, smart home, consumer electronics, monitoring, industrial automation, in-cabin detection, and health care; for example, the terminal may be an intelligent transportation device (such as an automobile, a bicycle, a motorcycle, a ship, a subway, a train, etc.), a security device (such as a camera), an intelligent wearable device (such as a bracelet, glasses, etc.), an intelligent home device (such as a television, an air conditioner, an intelligent lamp, etc.), various communication devices (such as a mobile phone, a tablet computer, etc.), etc., and may also be various instruments for detecting vital sign parameters and various devices carrying the instruments, such as a barrier gate, an intelligent traffic indicator, an intelligent sign, a traffic camera, various industrial manipulators (or robots), etc. The radio device may be a radio device described in any embodiment of the present application, and the structure and working principle of the radio device are described in detail in the above embodiments, which are not described in detail herein.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, or the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples merely represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.
Claims (23)
1. The signal source number calculating method is characterized by comprising the following steps:
obtaining K incoming wave directions corresponding to echo signals, wherein K is more than or equal to 1;
according to the K incoming wave directions, calculating to obtain first model features;
and processing the first model characteristic according to a first model, and determining the number of signal sources corresponding to the echo signals.
2. The method for calculating the number of signal sources according to claim 1, wherein the acquiring K incoming wave directions corresponding to the echo signals includes:
and estimating the incoming wave directions of the echo signals to obtain the K incoming wave directions.
3. The method according to claim 2, wherein the estimating the incoming wave directions of the echo signals to obtain K incoming wave directions includes:
Acquiring the estimated quantity of the incoming wave directions corresponding to the echo signals;
and estimating the incoming wave direction of the echo signals according to the estimated quantity, and obtaining the incoming wave direction corresponding to the estimated quantity.
4. The method for calculating the number of signal sources according to claim 3, wherein the obtaining the estimated number of incoming wave directions corresponding to the echo signals includes:
acquiring a current application scene, and acquiring a corresponding first model according to the current application scene;
reading input parameters of the first model;
and determining the estimated quantity of the target incoming wave direction according to the input parameters.
5. The signal source number calculation method according to claim 2, wherein the first model feature includes: single target power ratio and/or multiple target power ratio; the single target power ratio is calculated according to the energy of the signal source corresponding to each incoming wave direction and the energy of the echo signal to obtain the single target power ratio corresponding to each incoming wave direction; the multi-target power ratio is calculated according to at least two incoming wave directions to obtain comprehensive energy, and is calculated according to the comprehensive energy and the energy of the echo signal.
6. The method for calculating the number of signal sources according to claim 5, wherein,
in the case where K includes K1 and K1 is equal to 1, the K incoming wave directions correspond to a first incoming wave direction, and a first feature W (θ 0 ) The first characteristic W (θ 0 ) For the single target power ratio; and/or
In the case where K comprises K2, and K2 is equal to 2,the K incoming wave directions comprise a second incoming wave direction and a third incoming wave direction, and a second characteristic W (theta) is calculated according to the second incoming wave direction and the third incoming wave direction 1 ) Third feature W (θ 2 ) Fourth feature W (θ 1 θ 2 ) The second characteristic W (θ 1 ) And the third feature W (θ 2 ) For the single target power ratio, the fourth characteristic W (θ 1 θ 2 ) For the multi-target power ratio.
7. The method of claim 6, wherein the processing the first model feature according to a first model to determine the number of signal sources corresponding to the echo signal comprises:
and calculating the number of signal sources corresponding to the echo signals according to at least any two characteristics in the first model based on the first model.
8. A training method for a first model, the training method comprising:
acquiring training data, wherein the training data comprises training echo signals and the number of corresponding training signal sources;
calculating according to the training echo signals to obtain training characteristics;
respectively selecting at least two training features and the corresponding training signal source numbers to generate at least one training set;
and training according to at least two simulation features in the training set and the corresponding simulation signal source numbers to obtain at least one first model.
9. The method of training a first model according to claim 8, wherein training at least two of the simulation features in the training set and the corresponding number of the simulation signal sources according to the at least one first model to obtain a first model comprises:
acquiring actual measurement echo signals and the corresponding actual measurement signal source numbers under a set application scene;
processing according to the actually measured echo signals and at least one first model to obtain the corresponding number of model signal sources;
comparing the number of the model signal sources with the number of the corresponding actually measured signal sources to obtain the accuracy of each first model;
And selecting the first model with the accuracy meeting the requirement as the first model corresponding to the application scene.
10. The method of calculating the number of signal sources according to claim 8, wherein after training according to the training set to obtain at least one first model, respectively, the method comprises:
acquiring actual measurement echo signals and the corresponding actual measurement signal source numbers under a set application scene;
and optimizing the first model according to the acquired actual measurement echo signals and the corresponding actual measurement signal source numbers.
11. The echo signal processing method is characterized by comprising the following steps:
receiving an echo signal;
the method for calculating the number of signal sources according to any one of claims 1 to 7, wherein the echo signal is processed to obtain the number of signal sources.
12. Signal source number calculation means for performing the signal source number calculation method according to any one of claims 1 to 7.
13. Training means of a first model, characterized in that the training means are adapted to perform the signal source number calculation method according to any of the claims 8 to 10.
14. An echo signal processing device for performing the echo signal processing method according to claim 11.
15. Processing means comprising a memory and a processor, said memory storing a computer program, characterized in that the processor implements the method of any one of claims 1 to 7,8 to 10, or 11 when executing said computer program.
16. A computer readable storage medium having stored thereon a computer program, characterized in that the computer program comprises instructions for implementing the method of any of claims 1 to 7,8 to 10, or 11.
17. Computer program product comprising a computer program, characterized in that the computer program comprises instructions for implementing the method of any one of claims 1 to 7,8 to 10, or 11.
18. An operation control device comprising a memory and a processor, the memory storing a computer program, characterized in that the computer program, when run on the processor, causes the operation control device or the processor to implement the method of any one of claims 1 to 7,8 to 10, or 11.
19. An integrated circuit comprising a digital circuit and an operation control device according to claim 18, the operation control device being arranged to control the digital circuit to implement different functions of the integrated circuit.
20. A radio device, comprising:
a carrier;
the integrated circuit of claim 19 disposed on a carrier;
the antenna is arranged on the supporting body or integrated with the integrated circuit into a whole;
the integrated circuit is connected with the antenna and is used for transmitting and receiving radio signals.
21. The radio device of claim 20, wherein the radio device is a radar or a radar chip.
22. A terminal, comprising:
the radio device of claim 20 or 21;
wherein the radio device is used for target detection and/or communication.
23. The terminal of claim 22, wherein the terminal is a vehicle, a smart home device, or a robot.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210653922.7A CN117269912A (en) | 2022-06-10 | 2022-06-10 | Signal source number calculation method, training method of first model and signal processing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210653922.7A CN117269912A (en) | 2022-06-10 | 2022-06-10 | Signal source number calculation method, training method of first model and signal processing method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117269912A true CN117269912A (en) | 2023-12-22 |
Family
ID=89199565
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210653922.7A Pending CN117269912A (en) | 2022-06-10 | 2022-06-10 | Signal source number calculation method, training method of first model and signal processing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117269912A (en) |
-
2022
- 2022-06-10 CN CN202210653922.7A patent/CN117269912A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Terán et al. | IoT-based system for indoor location using bluetooth low energy | |
Ma et al. | Fusion of RSS and phase shift using the Kalman filter for RFID tracking | |
CN109917390A (en) | Vehicle checking method and system based on radar | |
CN112859033B (en) | Target detection method and device and related equipment | |
CN111935820A (en) | Positioning implementation method based on wireless network and related equipment | |
CN106716170A (en) | Method for locating a device and a mobile locator device | |
CN113325363B (en) | Method, device and related equipment for determining direction of arrival | |
CN112689773B (en) | Radar signal processing method and radar signal processing device | |
CN117269912A (en) | Signal source number calculation method, training method of first model and signal processing method | |
CN109298391B (en) | Fixed-place information source positioning system, method and application | |
CN112578354B (en) | Method for determining azimuth angle of target object, computer device and storage medium | |
CN116033345A (en) | Indoor abnormal signal high-precision positioning method, system and device | |
US9838994B1 (en) | Production system and method for location-aware environment | |
Cheng et al. | Localization in inconsistent wifi environments | |
Kosaka et al. | RSSI-based indoor localization using two-step XGBoost | |
Ding et al. | RFTraffic: a study of passive traffic awareness using emitted RF noise from the vehicles | |
Grzechca et al. | Indoor localization of objects based on RSSI and MEMS sensors | |
Tian et al. | SAPIL: single access point based indoor localisation using Wi‐Fi L‐shaped antenna array | |
CN112379328A (en) | Robust TDOA (time difference of arrival) positioning method with low power consumption and robust median filter | |
WO2019107237A1 (en) | Location estimating apparatus, location estimating method and program storing recording medium, and location estimating system | |
CN111726743A (en) | Wifi positioning method and system based on online learning | |
Al Amiri et al. | Signals of Opportunity for Car Density Estimation with Limited Training Data | |
CN113840373B (en) | Positioning method, positioning device, electronic equipment and storage medium | |
CN113543136B (en) | Eavesdropping source 3D positioning method and related equipment | |
US20240114542A1 (en) | Methods and systems for ultra-wideband localization |
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 |