CN115436881A - Positioning method, system, computer equipment and readable storage medium - Google Patents
Positioning method, system, computer equipment and readable storage medium Download PDFInfo
- Publication number
- CN115436881A CN115436881A CN202211269720.9A CN202211269720A CN115436881A CN 115436881 A CN115436881 A CN 115436881A CN 202211269720 A CN202211269720 A CN 202211269720A CN 115436881 A CN115436881 A CN 115436881A
- Authority
- CN
- China
- Prior art keywords
- chaotic
- ultrasonic
- signal
- signals
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S5/00—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations
- G01S5/18—Position-fixing by co-ordinating two or more direction or position line determinations; Position-fixing by co-ordinating two or more distance determinations using ultrasonic, sonic, or infrasonic waves
- G01S5/22—Position of source determined by co-ordinating a plurality of position lines defined by path-difference measurements
-
- 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
Abstract
The application belongs to the field of ultrasonic ranging and discloses a positioning method, a positioning system, computer equipment and a readable storage medium, wherein the method comprises the following steps: sending a plurality of first chaotic ultrasonic signals of different frequency bands, wherein the first chaotic ultrasonic signals are generated and transmitted by combining a plurality of ultrasonic wave generation modules on a component with a chaotic system; receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air; splicing the plurality of second chaotic ultrasonic signals to obtain spliced signals; calculating by adopting a neural network according to the splicing signal to obtain the position information of the ultrasonic wave generating module corresponding to the first chaotic ultrasonic signal; obtaining attitude information according to the position information, and calculating according to the position information and the attitude information to obtain a component attitude adjusting instruction and a displacement instruction; according to the component posture adjustment instruction and the displacement instruction, the posture adjustment and the displacement of the component are controlled, and the effect of improving the component alignment efficiency and accuracy in industrial manufacturing is achieved.
Description
Technical Field
The present application relates to the field of ultrasonic ranging technologies, and in particular, to a positioning method, a positioning system, a computer device, and a readable storage medium.
Background
In industrial production, the assembly and welding of two industrial product components requires that the assembly mechanisms and welds of the components be sufficiently close and aligned. However, large or very large industrial components are characterized by high mass and difficult handling, which makes them difficult and risky to access and align. If collision happens in the approaching process of the components, the components can be damaged, production accidents can be caused, and even casualties can be caused. If the component alignment has large deviation, the quality of industrial products can not reach the standard, and economic loss is caused. The component is accurately positioned, the problems of approach collision, alignment deviation and the like of the component can be effectively solved, and the risk of accidents is reduced.
The existing method for accurately positioning the component comprises the following steps:
(1) Manual alignment:
in the assembling and welding process of the components, the components are aligned in a mode of human eye observation and experience judgment. The method depends on subjective perception and experience accumulation of people, the precision of component alignment can be influenced by physical and mental states of people, and each alignment can be difficult to ensure to be successful. In addition, manual alignment is inefficient and time consuming.
(2) The traditional alignment method based on ultrasonic positioning comprises the following steps:
the alignment method based on the traditional ultrasonic positioning utilizes ultrasonic waves to position the components to be assembled and welded, can accurately measure and align the components, and has higher alignment precision and efficiency compared with a manual alignment method. However, the transmitted waveform signal has problems of multipath effect and internal period, which can cause deviation of time estimation and further affect positioning accuracy.
In addition, the accuracy of conventional ultrasonic location methods depends on the effectiveness of human design algorithms (e.g., TDOA, MUSIC, etc.). The algorithms are often deviated from the actual conditions in application, and a large amount of additional experiments are needed for compensating and correcting the algorithm models. The difficulty and cost of implementation are high.
(3) The alignment method based on electromagnetic wave positioning comprises the following steps:
the propagation speed of electromagnetic wave is about 3X 10 8 m/s, if the aligned system produces a slight deviation in the measurement of time, the positioning will produce a large error.
Disclosure of Invention
Therefore, the embodiment of the application provides a positioning method, a positioning system and a readable storage medium, which can solve the technical problems of low alignment efficiency and low accuracy of the existing components, and the specific technical scheme content is as follows:
in a first aspect, an embodiment of the present application provides a positioning method, including:
sending a plurality of first chaotic ultrasonic signals with different frequency bands, wherein the first chaotic ultrasonic signals are generated and emitted by combining a plurality of ultrasonic generation modules arranged on a member with a chaotic system;
receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air;
splicing the plurality of second chaotic ultrasonic signals to obtain spliced signals;
calculating by adopting a neural network according to the splicing signal to obtain the position information of an ultrasonic transmitting module corresponding to the first chaotic ultrasonic signal; the neural network comprises a backbone network, a coarse positioning network and a fine positioning network, and an output signal of the backbone network is input into the coarse positioning network and the fine positioning network; the output signal of the coarse positioning network is input into the fine positioning network;
obtaining attitude information according to the position information, and calculating to obtain a component attitude adjusting instruction and a displacement instruction according to the position information and the attitude information;
and controlling the posture adjustment and displacement of the component according to the posture adjustment instruction of the component and the displacement instruction.
Preferably, the backbone network consists of five bottleneck structures, a global average pooling layer and two attention modules;
the convolution layer of the first bottleneck structure is a two-dimensional convolution, and the convolution layers of the other bottleneck structures are one-dimensional convolutions;
after the output signal of the first bottleneck structure is input into the global average pooling layer, the output signal of the first bottleneck structure is sequentially input into the other bottleneck structures;
and the output signal of the last bottleneck structure is sequentially input into the attention module.
Preferably, the bottleneck structure comprises 3 convolutional layers, 3 ReLU activation functions, and 1 downsampling layer;
the convolution layer and the ReLU activation function are arranged in a matched mode, and output signals of the convolution layer are input into the ReLU activation function and then output;
and accumulating the output signals of the ReLU activation function and inputting the accumulated output signals into the downsampling layer.
Preferably, the coarse positioning network comprises 3 convolutional layers, 3 ReLU activation functions and 1 adaptive average pooling layer;
the convolution layer and the ReLU activation function are arranged in a matched mode, and output signals of the convolution layer are input into the ReLU activation function and then output;
and after the output signals of the ReLU activation function are accumulated, the output signals are input into the self-adaptive average pooling layer.
Preferably, the fine positioning network comprises 5 convolutional layers, 5 ReLU activation functions and 1 adaptive average pooling layer;
the convolution layer and the ReLU activation function are arranged in a matched mode, and output signals of the convolution layer are input into the ReLU activation function and then output;
and after the output signals of the ReLU activation function are accumulated, the output signals are input into the self-adaptive average pooling layer.
Preferably, the receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air is:
setting a receiving window with the length longer than that of the first chaotic ultrasonic signal so as to receive a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air.
Preferably, the splicing the plurality of second chaotic ultrasonic signals to obtain a spliced signal includes:
zero padding is carried out on the prestored first chaotic ultrasonic signal, so that the length of the first chaotic ultrasonic signal is the same as that of the second chaotic ultrasonic signal;
and splicing the first chaotic ultrasonic signal subjected to zero padding and a second chaotic ultrasonic signal of the same frequency band to obtain a spliced signal.
Preferably, the obtaining of the position information of the ultrasonic emission module corresponding to the first chaotic ultrasonic signal by adopting neural network calculation according to the splicing signal includes:
constructing a sample set by placing an ultrasonic transmitting module at a preset space coordinate and receiving a transmitted waveform signal of the ultrasonic transmitting module;
training the neural network in a supervised learning mode through the sample set, and calculating the position information by using a plurality of spliced signals in learning;
and inputting the splicing signal into the trained neural network to obtain the position information of the ultrasonic emission module corresponding to the first chaotic ultrasonic signal.
In a second aspect, embodiments of the present application provide a positioning system, the system comprising:
an ultrasonic transmitting module: sending a plurality of first chaotic ultrasonic signals of different frequency bands, wherein the first chaotic ultrasonic signals are generated and transmitted by combining an ultrasonic transmitting module on a component with a chaotic system;
an ultrasonic acquisition module: receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air;
a data preprocessing module: splicing the received second chaotic ultrasonic signals to obtain spliced signals;
the deep neural network module: calculating by adopting a neural network according to the splicing signal to obtain the position information of an ultrasonic transmitting module corresponding to the first chaotic ultrasonic signal; the neural network comprises a backbone network, a coarse positioning network and a fine positioning network, and an output signal of the backbone network is input into the coarse positioning network and the fine positioning network; the output signal of the coarse positioning network is input into the fine positioning network;
the position information post-processing module: obtaining attitude information according to the position information, and calculating according to the position information and the attitude information to obtain a component attitude adjustment instruction and a displacement instruction;
a component alignment module: and receiving the attitude adjustment instruction and the displacement instruction, and controlling the attitude adjustment and displacement of the component.
Preferably, the ultrasonic transmitting module includes at least four ultrasonic transmitting devices, and different ultrasonic transmitting devices transmit a plurality of first chaotic ultrasonic signals of different frequency bands.
Preferably, the ultrasound acquisition module comprises at least four ultrasound receiving devices, and each ultrasound receiving device is provided with a plurality of signal acquisition channels.
In a third aspect, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements the steps of the positioning method according to any one of the preceding claims when executing the computer program.
In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and the computer program, when executed by a processor, implements the steps of the positioning method according to any one of the preceding claims.
In summary, compared with the prior art, the beneficial effects brought by the technical scheme provided by the embodiment of the present application at least include:
the first chaotic ultrasonic signal is used as an ultrasonic transmitting module to be combined with a waveform transmitting signal of a chaotic system, so that the matching error of signal phases caused by multipath effects and signal internal periods is effectively avoided, the accuracy of component alignment is further improved, a neural network comprising a backbone network, a coarse positioning network and a fine positioning network is further used as a signal processing means at the rear end of calculation, the distance from the ultrasonic transmitting module at a specific position of a component to an ultrasonic receiving module in an assembly and welding field is measured, and the spatial coordinates of the component are positioned. The method has the advantages that the method has higher precision and adaptability compared with the traditional measuring and positioning method, and the error is lower indoors compared with the electromagnetic wave distance measuring method due to the adoption of the ultrasonic distance measuring and positioning method.
Drawings
Fig. 1 is a schematic flowchart of a positioning method according to an embodiment of the present application.
Fig. 2 is a flowchart illustrating a positioning method according to another embodiment of the present application.
Fig. 3 is a second flowchart of a positioning method according to another embodiment of the present application.
Fig. 4 is a main structural block diagram of a neural network of a positioning method according to an exemplary embodiment of the present application.
Fig. 5 is a diagram of a backbone network structure of a neural network according to an exemplary embodiment of the present application.
Fig. 6 is a schematic diagram of a bottleneck structure of a backbone network of a neural network according to an exemplary embodiment of the present application.
Fig. 7 is a diagram of a coarse positioning network structure of a neural network provided in an exemplary embodiment of the present application.
Fig. 8 is a diagram of a fine positioning network structure of a neural network according to an exemplary embodiment of the present application.
Fig. 9 is a schematic structural diagram of a positioning system according to an embodiment of the present application.
Detailed Description
The present embodiment is only for explaining the present application, and it is not limited to the present application, and those skilled in the art can make modifications of the present embodiment without inventive contribution as needed after reading the present specification, but all of them are protected by patent law within the scope of the claims of the present application.
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In addition, the term "and/or" in the present application is only one kind of association relationship describing the associated object, and means that three kinds of relationships may exist, for example, a and/or B may mean: a exists alone, A and B exist simultaneously, and B exists alone. In addition, the character "/" in this application generally indicates that the preceding and following related objects are in an "or" relationship, unless otherwise specified.
The terms "first," "second," and the like in this application are used for distinguishing between similar items and items that have substantially the same function or similar functionality, and it should be understood that "first," "second," and "nth" do not have any logical or temporal dependency or limitation on the number or order of execution.
The term "at least one" in this application means one or more, and the meaning of "a plurality" means three or more, e.g. a plurality of first locations means three or more first locations.
The embodiments of the present application will be described in further detail with reference to the drawings attached hereto.
Referring to fig. 1, in one embodiment of the present application, there is provided a positioning method, the main steps of which are described as follows:
s1: sending a plurality of first chaotic ultrasonic signals of different frequency bands, wherein the first chaotic ultrasonic signals are generated and emitted by combining a plurality of ultrasonic generation modules on a component with a chaotic system;
s2: receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air;
s3: splicing the plurality of second chaotic ultrasonic signals to obtain spliced signals;
s4: calculating by adopting a neural network according to the splicing signal to obtain the position information of an ultrasonic transmitting module corresponding to the first chaotic ultrasonic signal; the neural network comprises a backbone network, a coarse positioning network and a fine positioning network, and an output signal of the backbone network is input into the coarse positioning network and the fine positioning network; the output signal of the coarse positioning network is input into the fine positioning network;
s5: obtaining attitude information according to the position information, and calculating to obtain a component attitude adjusting instruction and a displacement instruction according to the position information and the attitude information;
s6: and controlling the posture adjustment and displacement of the component according to the posture adjustment instruction of the component and the displacement instruction.
Specifically, the method is suitable for a system for positioning the component in the assembling and welding processes of the component, the system comprises an ultrasonic transmitting module, an ultrasonic acquisition module and a data processing module, the ultrasonic transmitting module adopts chaotic ultrasonic signals as signals, and the matching error of signal phases caused by multipath effects and internal periods of the signals is effectively avoided.
Specifically, in this embodiment, the ultrasonic emission module sends a plurality of first chaotic ultrasonic signals of different frequency bands, and a plurality of sets of first chaotic ultrasonic signals of different frequency bands are used, so as to achieve the effects of preventing interference and improving the measurement accuracy.
The chaotic ultrasonic signal is a continuous and ergodic (all samples in the space can appear at least once) non-periodic signal, and the non-periodicity can effectively avoid the matching error of the signal phase and is beneficial to improving the positioning precision.
There are many ways of generating chaotic ultrasonic signals, and the present application takes an analytic solution chaotic system as an example for introduction. The analytic solution chaotic system is a chaotic system proposed by Corron N.J. et al in 2010. The kinetic equation of the first chaotic ultrasonic signal can be represented by a differential equation:
wherein the content of the first and second substances,
when the temperature of the water is higher than the set temperature,
wherein w represents the angular frequency of the mixed chaotic ultrasonic signal;
the negative attenuation coefficient is a left-open and right-close interval from 0 to ln 2; s = sgn (x) is a switching function, s is a symbol sequence, and takes the value 1 or-1;
(ii) a x is a continuous first chaotic ultrasonic signal.
Formula (1) is a differential equation, g is a negative attenuation coefficient, the value is 0 to ln2, and the system is chaotic only in the range. x is a function of time t, the amplitude of the x changes with t, the x is a continuous chaotic ultrasonic signal which is a one-dimensional signal and can be transmitted only according to the intensity of x (t) when the x is transmitted.
s = sgn (x) is a switching function, s is defined as a switching condition, i.e. when the first derivative of x is 0, s (t) is assigned sgn (u (t)), s is a sequence of signs, the value ± 1, s is not changed depending on the case where the signal derivative is 0, and s (t) remains unchanged for the rest of the time until the next switching condition is met.
Deriving an analytical solution from equations (1) and (2)
In the case of the above-described situation,
is the sign value at t = n,
the chaotic ultrasonic signal sampling value when t = n satisfies the following iterative relationship:
according to the analytic solution iterative formula, one optimal value of x (0) is-0.3776, namely the chaotic waveform can be generated and is emitted by an ultrasonic wave generation module on the component.
Referring to fig. 2, further, S2 is S2':
and S2': and setting a receiving window with the length longer than that of the first chaotic ultrasonic signal so as to receive a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air.
The ultrasonic acquisition module can not synchronously receive the data sent by the ultrasonic transmitting module, and the receiving window of the ultrasonic acquisition module is set to be larger than the data volume sent by the ultrasonic transmitting module, so that the condition that the ultrasonic receiving module is missed or less received can be effectively reduced. For example, a signal of 1.5 seconds is transmitted and received in a window of two seconds.
Referring to fig. 3, further, S3 includes:
s31: zero padding is carried out on the prestored first chaotic ultrasonic signal, so that the length of the first chaotic ultrasonic signal is the same as that of the second chaotic ultrasonic signal;
s32: and splicing the first chaotic ultrasonic signal subjected to zero padding and a second chaotic ultrasonic signal of the same frequency band to obtain a spliced signal.
Through the arrangement of the embodiment, the first chaotic ultrasonic signal and the second chaotic ultrasonic signal are processed in the same computing device, and after the computing device generates the first chaotic ultrasonic signal, the first chaotic ultrasonic signal is stored in advance so as to be spliced with the second chaotic ultrasonic signal to form a spliced signal.
For example, the computer respectively transmits signals of the same frequency band received by the ultrasound acquisition modules of the four acquisition channels (the same frequency band represents the signals transmitted by the ultrasound transmitting device on the same ultrasound transmitting module), and splices the first chaotic ultrasonic signal subjected to zero padding and the four channels of received second chaotic ultrasonic signals with the same frequency. For example, the data length of the transmission and reception waveform signal is [1, 20000], and the shape of the spliced data is [5, 20000].
Referring to fig. 3, further, S4 includes:
s41: constructing a sample set by placing an ultrasonic transmitting module at a preset space coordinate and receiving a transmitted waveform signal of the ultrasonic transmitting module;
s42: training the neural network in a supervised learning mode through the sample set, and calculating the position information by utilizing a plurality of spliced signals in learning;
s43: and inputting the splicing signal into the trained neural network to obtain the position information of the ultrasonic emission module corresponding to the first chaotic ultrasonic signal.
Knowing that the propagation speed of the ultrasonic wave in the air medium at room temperature is about 340m/s, the distance from the ultrasonic transmitting module to the ultrasonic acquisition module can be calculated by using the time required by the signal from the ultrasonic transmitting module to the ultrasonic acquisition module. When the spatial position of the ultrasound acquisition module is fixed (i.e. the spatial coordinates of the ultrasound acquisition module are known), the spatial coordinates of the ultrasound transmission module can be calculated by using the distances from the ultrasound transmission module to the plurality of ultrasound acquisition modules. It is noted that positioning the ultrasound transmission module within a two-dimensional plane requires at least three ultrasound acquisition modules. Location ultrasonic emission module needs four at least ultrasonic acquisition modules in three-dimensional space, if ultrasonic acquisition device's quantity is less than four, then the information of receiving is not used for neural network study inadequately, and neural network is difficult to learn effective information promptly, and ultrasonic acquisition device sets up quantity too much, can increase the handling capacity of data, and also can increase the cost, consequently in this application, ultrasonic acquisition device is preferred four.
With the increase of the number of the ultrasonic acquisition modules, the system is more robust, but because inevitable noise exists in a real system, the system can calculate a plurality of spatial coordinates for the ultrasonic transmitting device of the same ultrasonic transmitting module, so that the precision of the traditional ultrasonic positioning system is difficult to guarantee, a large amount of additional experiments are needed to compensate and correct the system, and the implementation difficulty and the cost are higher.
Referring to fig. 4, the neural network of the present application has a structure including a backbone network, a coarse positioning network, and a fine positioning network. After the preprocessed signals are input into the backbone network, the output signals of the backbone network are input into the coarse positioning network and the fine positioning network, and the output signals of the coarse positioning network are input into the fine positioning network.
When the execution precision requirement is only in the centimeter-level range, the fine positioning network of the neural network does not work, and the position information is directly output through the coarse positioning network of the neural network.
Referring to fig. 5, the backbone network structure is composed of five bottleneck structures, one global average pooling layer and two attention modules. The attention module includes a channel attention module and a spatial attention module. The two attention modules can enable the neural network to pay more attention to key information, and reduce attention to local invalid information, so that the purposes of simplifying a model and accelerating calculation are achieved.
The convolution layer of the first bottleneck structure is a two-dimensional convolution, and the convolution layers of the other bottleneck structures are one-dimensional convolutions. The global average pooling layer maps the output of the first bottleneck structure to [1,10000], and then inputs the output into the remaining bottleneck structures in sequence.
And setting the number C of convolution kernels of the bottleneck structure, the size K of the convolution kernels and a value P supplemented by the convolution process. In the first bottleneck structure (i.e. bottleneck structure-1), C32K5 represents 32 convolution kernels of 5 × 5, and in the remaining bottleneck structures, K is fixedly set to 1x9, and p is set to 4, which is intended to prevent data from missing boundary data and supplementing the boundary data due to convolution characteristics in the convolution process.
The final bottleneck structure output result is multiplied by the channel attention module and then multiplied by the space attention module to be output, and the final output size of the backbone network is [512,625].
Referring to fig. 6, the bottleneck structure includes 3 convolutional layers, 3 ReLU activation functions, and 1 downsampling layer. The convolution layer is matched with the ReLU activation function, and an output signal of the convolution layer is input into the ReLU activation function and then output; the output signal of the ReLU activation function is accumulated and then input into a down-sampling layer.
Referring to fig. 7, the coarse positioning network structure includes 3 convolutional layers, 3 ReLU activation functions, and 1 adaptive average pooling layer. The convolution layer is matched with the ReLU activation function, and an output signal of the convolution layer is input into the ReLU activation function and then output; and accumulating the output signals of the ReLU activation function and inputting the accumulated output signals into the adaptive average pooling layer.
The coarse positioning network receives the output of the backbone network and outputs a vector shaped as [1,3 ]. Three elements in the vector are respectively coarse positioning space coordinates of the key nodes, are expressed by (x, y, z) coordinates, and have the precision of centimeter level.
Referring to fig. 8, the fine positioning network structure contains 5 convolutional layers, 5 ReLU activation functions, and 1 adaptive average pooling layer. The convolution layer is matched with the ReLU activation function, and an output signal of the convolution layer is input into the ReLU activation function and then output; and accumulating the output signals of the ReLU activation function and inputting the accumulated output signals into the adaptive average pooling layer.
And the fine positioning network receives a matrix formed by splicing the outputs of the backbone network and the coarse positioning network. The matrix is shaped [515,625], with the first 512 rows being the output matrix of the backbone network and the last 3 rows being the repeated spatial coordinates of 3 coarse positions. The output of the fine positioning network is a 1 [3, 11] matrix. The position where the element is largest represents the output result of the fine positioning. 11 elements of the output vector correspond to integers from-5 to 5 respectively, and represent correction of the positioning output result of the coarse positioning neural network under millimeter accuracy.
The characteristic information is extracted through the backbone network of the neural network, the coarse positioning network of the neural network outputs coordinates, and the coarse positioning network can output the position information of the key nodes through effective training due to the fact that the precision is centimeter level.
The fine positioning network of the neural network receives the position information output by the coarse positioning network and the characteristic information extracted by the main network, and realizes effective correction of the positioning error of the coarse positioning network by classifying the offset under centimeter precision.
In particular, assume that the output of the coarse positioning network is
The output of the fine positioning network is
Then the final output result of the neural network
Comprises the following steps:
wherein
Are respectively as
The coordinates on different coordinate axes are respectively obtained by the following formulas:
wherein
All in three-dimensional coordinate form:
wherein, the first and the second end of the pipe are connected with each other,
、
and
respectively for fine positioning network transmissionOut of [3, 11]The position index of the maximum value in the first, second and third rows of the matrix. The method adopts a neural network (an artificial intelligence technology) as a signal processing means at the back end of calculation, measures the distance from an ultrasonic generating system at a specific position of a component to an ultrasonic receiving system in an assembly and welding field, and positions the spatial coordinates of the component. Due to the strong learning capability of the deep neural network, compared with the traditional measurement positioning method, the method has higher precision and adaptability. In addition, the method only needs to collect limited component positioning sample data for fine tuning training, and can be directly used for positioning of the component, so that the defect that an algorithm model needs to be compensated and corrected before use is overcome, and the implementation difficulty and the cost are low.
The neural network adopted by the application is a mathematical model with learning ability. During training, the neural network can obtain an output result according to the input sample calculation, and adjust internal parameters according to the output result and the error of the sample label. Therefore, the neural network can learn how to get the desired result from the data in the training process without manually designing the calculation flow thereof.
In the application, a sample set required by neural network training can be constructed in a mode of placing an ultrasonic transmitting module at a preset space coordinate and receiving a transmitted waveform signal of the ultrasonic transmitting module. The sample is a preprocessed received waveform signal, and the sample label is a preset space coordinate. The positioning system can automatically learn how to accurately position the spatial coordinates of the ultrasonic emission source by using the received waveform signals acquired by a plurality of (at least four) ultrasonic receiving devices in a supervision learning mode by benefiting from the powerful learning capability of a neural network, particularly a deep neural network, so that the system is prevented from being compensated and corrected by a large number of additional experiments of the traditional positioning system, and the implementation difficulty and cost are reduced.
In practical application, collected ultrasonic signal samples (corresponding to different ultrasonic transmitting devices) of different frequency bands are respectively input into a trained neural network, and then the spatial coordinates of different ultrasonic transmitting devices can be obtained.
It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by functions and internal logic of the process, and should not limit the implementation process of the embodiments of the present invention in any way.
Referring to fig. 9, in one embodiment of the present application, a positioning system is provided that is used in conjunction with the positioning method of the above-described embodiment. The positioning system includes:
an ultrasonic transmitting module: sending a plurality of first chaotic ultrasonic signals with different frequency bands, wherein the first chaotic ultrasonic signals are generated and transmitted by an ultrasonic transmitting module on a member in combination with a chaotic system;
an ultrasonic acquisition module: receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air;
a data preprocessing module: splicing the received second chaotic ultrasonic signals to obtain spliced signals;
the deep neural network module: calculating by adopting a neural network according to the splicing signal to obtain the position information of an ultrasonic transmitting module corresponding to the first chaotic ultrasonic signal; the neural network comprises a backbone network, a coarse positioning network and a fine positioning network, and an output signal of the backbone network is input into the coarse positioning network and the fine positioning network; the output signal of the coarse positioning network is input into the fine positioning network;
the position information post-processing module: obtaining attitude information according to the position information, and calculating to obtain a component attitude adjusting instruction and a displacement instruction according to the position information and the attitude information;
a component alignment module: and receiving the attitude adjustment instruction and the displacement instruction, and controlling the attitude adjustment and displacement of the component.
Further, the ultrasonic transmitting module comprises at least four ultrasonic transmitting devices, and different ultrasonic transmitting devices transmit a plurality of first chaotic ultrasonic signals with different frequency bands.
The ultrasonic acquisition module comprises at least four ultrasonic receiving devices, and each ultrasonic receiving device is provided with a plurality of signal acquisition channels.
The various modules of the positioning system described above may be implemented in whole or in part by software, hardware, and combinations thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment of the embodiments of the present application, a computer device is provided, which may be a server. The computer device includes a processor, a memory, and a network interface connected by a system bus. Including but not limited to a CPU (central processing unit), a GPU (graphics processing unit), among others, and the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device may be implemented by any type or combination of volatile or non-volatile storage devices, including but not limited to: magnetic disk, optical disk, EEPROM (Electrically-Erasable Programmable Read Only Memory), EPROM (Erasable Programmable Read Only Memory), SRAM (Static Random Access Memory), ROM (Read-Only Memory), magnetic Memory, flash Memory, PROM (Programmable Read-Only Memory). The memory of the computer device provides an environment for the running of the operating system and computer programs stored within it. The network interface of the computer device is used for communicating with an external terminal through a network connection. Which when executed by a processor performs the steps of the positioning method described in the above embodiments.
In an embodiment of the present application, a computer-readable storage medium is provided, which stores a computer program, which when executed by a processor implements the positioning method steps described in the above embodiment. The computer-readable storage medium includes a ROM (Read-only memory), a RAM (Random-access memory), a CD-ROM (compact disc-Read-only memory), a magnetic disk, a floppy disk, and the like.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-mentioned division of the functional units and modules is illustrated, and in practical applications, the above-mentioned function distribution may be performed by different functional units and modules according to needs, that is, the internal structure of the system described in this application is divided into different functional units or modules to perform all or part of the above-mentioned functions.
Claims (13)
1. A method of positioning, the method comprising:
sending a plurality of first chaotic ultrasonic signals with different frequency bands, wherein the first chaotic ultrasonic signals are generated and transmitted by an ultrasonic transmitting module on a member in combination with a chaotic system;
receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air;
splicing the plurality of second chaotic ultrasonic signals to obtain spliced signals;
calculating by adopting a neural network according to the splicing signal to obtain the position information of an ultrasonic transmitting module corresponding to the first chaotic ultrasonic signal; the neural network comprises a backbone network, a coarse positioning network and a fine positioning network, and an output signal of the backbone network is input into the coarse positioning network and the fine positioning network; the output signal of the coarse positioning network is input into the fine positioning network;
obtaining attitude information according to the position information, and calculating to obtain a component attitude adjusting instruction and a displacement instruction according to the position information and the attitude information;
and controlling the posture adjustment and displacement of the component according to the posture adjustment instruction of the component and the displacement instruction.
2. The positioning method according to claim 1, wherein the backbone network is composed of five bottleneck structures, one global average pooling layer and two attention modules;
the convolution layers of the first bottleneck structure are two-dimensional convolutions, and the convolution layers of the other bottleneck structures are one-dimensional convolutions;
after the output signal of the first bottleneck structure is input into the global average pooling layer, the output signal of the first bottleneck structure is sequentially input into the other bottleneck structures;
and the output signal of the last bottleneck structure is sequentially input into the attention module.
3. The method of claim 2, wherein the bottleneck structure comprises 3 convolutional layers, 3 ReLU activation functions, and 1 downsampling layer;
the convolution layer and the ReLU activation function are arranged in a matched mode, and output signals of the convolution layer are input into the ReLU activation function and then output;
and after the output signals of the ReLU activation functions are accumulated, inputting the accumulated output signals into the downsampling layer.
4. The positioning method according to claim 1, wherein the coarse positioning network comprises 3 convolutional layers, 3 ReLU activation functions, and 1 adaptive average pooling layer;
the convolution layer and the ReLU activation function are arranged in a matched mode, and output signals of the convolution layer are output after being input into the ReLU activation function;
and after the output signals of the ReLU activation function are accumulated, the output signals are input into the self-adaptive average pooling layer.
5. The positioning method according to claim 1, wherein the fine positioning network comprises 5 convolutional layers, 5 ReLU activation functions, and 1 adaptive average pooling layer;
the convolution layer and the ReLU activation function are arranged in a matched mode, and output signals of the convolution layer are input into the ReLU activation function and then output;
and after the output signals of the ReLU activation function are accumulated, the output signals are input into the self-adaptive average pooling layer.
6. The positioning method according to claim 1, wherein the receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air is:
and setting a receiving window with the length longer than that of the first chaotic ultrasonic signal so as to receive a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air.
7. The positioning method according to claim 6, wherein the obtaining a spliced signal by splicing the plurality of second chaotic ultrasonic signals comprises:
zero padding is carried out on the prestored first chaotic ultrasonic signal, so that the length of the first chaotic ultrasonic signal is the same as that of the second chaotic ultrasonic signal;
and splicing the first chaotic ultrasonic signal subjected to zero padding and a second chaotic ultrasonic signal of the same frequency band to obtain a spliced signal.
8. The positioning method according to claim 7, wherein the obtaining of the position information of the ultrasonic emission module corresponding to the first chaotic ultrasonic signal by adopting neural network calculation according to the splicing signal comprises:
constructing a sample set by placing an ultrasonic transmitting module at a preset space coordinate and receiving a transmitted waveform signal of the ultrasonic transmitting module;
training the neural network in a supervised learning mode through the sample set;
and inputting the splicing signal into the trained neural network to obtain the position information of the ultrasonic transmitting module corresponding to the first chaotic ultrasonic signal.
9. A positioning system, characterized in that the system comprises:
an ultrasonic transmitting module: sending a plurality of first chaotic ultrasonic signals with different frequency bands, wherein the first chaotic ultrasonic signals are generated and transmitted by an ultrasonic transmitting module on a member in combination with a chaotic system;
an ultrasonic acquisition module: receiving a plurality of corresponding second chaotic ultrasonic signals formed by a plurality of first chaotic ultrasonic signals propagated in the air;
a data preprocessing module: splicing the received second chaotic ultrasonic signals to obtain spliced signals;
the deep neural network module: calculating by adopting a neural network according to the splicing signal to obtain the position information of an ultrasonic transmitting module corresponding to the first chaotic ultrasonic signal; the neural network comprises a backbone network, a coarse positioning network and a fine positioning network, and an output signal of the backbone network is input into the coarse positioning network and the fine positioning network; the output signal of the coarse positioning network is input into the fine positioning network;
the position information post-processing module: obtaining attitude information according to the position information, and calculating to obtain a component attitude adjusting instruction and a displacement instruction according to the position information and the attitude information;
a component alignment module: and receiving the posture adjustment instruction and the displacement instruction, and controlling the posture adjustment and the displacement of the component.
10. The positioning system of claim 9, wherein the ultrasound transmitting module comprises at least four ultrasound transmitting devices, and different ones of the ultrasound transmitting devices transmit a plurality of first chaotic ultrasound signals of different frequency bands.
11. The positioning system of claim 10, wherein the ultrasound acquisition module comprises at least four ultrasound receiving devices fixed within the field, each of the ultrasound receiving devices being provided with a plurality of signal acquisition channels.
12. A computer arrangement comprising a memory, a processor and a computer program stored in the memory and run on the processor, the processor implementing the steps of the positioning method of any one of claims 1 to 8 when executing the computer program.
13. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when being executed by a processor, carries out the steps of the positioning method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211269720.9A CN115436881B (en) | 2022-10-18 | 2022-10-18 | Positioning method, positioning system, computer equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211269720.9A CN115436881B (en) | 2022-10-18 | 2022-10-18 | Positioning method, positioning system, computer equipment and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115436881A true CN115436881A (en) | 2022-12-06 |
| CN115436881B CN115436881B (en) | 2023-07-07 |
Family
ID=84250472
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211269720.9A Active CN115436881B (en) | 2022-10-18 | 2022-10-18 | Positioning method, positioning system, computer equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115436881B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115713844A (en) * | 2022-10-18 | 2023-02-24 | 兰州交通大学 | Alarm method and system |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6072433A (en) * | 1996-07-31 | 2000-06-06 | California Institute Of Technology | Autonomous formation flying sensor |
| US6461316B1 (en) * | 1997-11-21 | 2002-10-08 | Richard H. Lee | Chaos therapy method and device |
| CN1888932A (en) * | 2006-07-17 | 2007-01-03 | 天津大学 | Chaos pulse sequence ultrasonic distance-measuring method and apparatus |
| WO2008118354A1 (en) * | 2007-03-26 | 2008-10-02 | Luna Innovations Incorporated | Method and apparatus for classifying gaseous and non-gaseous objects |
| CA2995409A1 (en) * | 2017-02-15 | 2018-08-15 | Microline Technology Corporation | Method and apparatus for metamaterial enhanced chaotic cavity transducer |
| WO2019051040A2 (en) * | 2017-09-06 | 2019-03-14 | Locix, Inc. | Systems and methods for coarse and fine time of flight estimates for precise radio frequency localization in the presence of multiple communication paths |
| WO2019212276A1 (en) * | 2018-05-04 | 2019-11-07 | Lg Electronics Inc. | Plurality of autonomous mobile robots and controlling method for the same |
| WO2019226686A2 (en) * | 2018-05-23 | 2019-11-28 | Movidius Ltd. | Deep learning system |
| CN113222875A (en) * | 2021-06-01 | 2021-08-06 | 浙江大学 | Image harmonious synthesis method based on color constancy |
| CN114137079A (en) * | 2021-11-30 | 2022-03-04 | 东莞理工学院 | Ultrasonic guided wave nondestructive testing method based on combination of deep learning and Duffing system |
| CN114494701A (en) * | 2022-02-14 | 2022-05-13 | 浙江大学 | Semantic segmentation method and device based on graph structure neural network |
| WO2022126392A1 (en) * | 2020-12-15 | 2022-06-23 | 中山大学孙逸仙纪念医院 | Model training method and apparatus, and electronic device, medium and bone mass measurement system |
| CN114998444A (en) * | 2022-05-20 | 2022-09-02 | 燕山大学 | Robot high-precision pose measurement system based on two-channel network |
-
2022
- 2022-10-18 CN CN202211269720.9A patent/CN115436881B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6072433A (en) * | 1996-07-31 | 2000-06-06 | California Institute Of Technology | Autonomous formation flying sensor |
| US6461316B1 (en) * | 1997-11-21 | 2002-10-08 | Richard H. Lee | Chaos therapy method and device |
| CN1888932A (en) * | 2006-07-17 | 2007-01-03 | 天津大学 | Chaos pulse sequence ultrasonic distance-measuring method and apparatus |
| WO2008118354A1 (en) * | 2007-03-26 | 2008-10-02 | Luna Innovations Incorporated | Method and apparatus for classifying gaseous and non-gaseous objects |
| CA2995409A1 (en) * | 2017-02-15 | 2018-08-15 | Microline Technology Corporation | Method and apparatus for metamaterial enhanced chaotic cavity transducer |
| WO2019051040A2 (en) * | 2017-09-06 | 2019-03-14 | Locix, Inc. | Systems and methods for coarse and fine time of flight estimates for precise radio frequency localization in the presence of multiple communication paths |
| WO2019212276A1 (en) * | 2018-05-04 | 2019-11-07 | Lg Electronics Inc. | Plurality of autonomous mobile robots and controlling method for the same |
| WO2019226686A2 (en) * | 2018-05-23 | 2019-11-28 | Movidius Ltd. | Deep learning system |
| WO2022126392A1 (en) * | 2020-12-15 | 2022-06-23 | 中山大学孙逸仙纪念医院 | Model training method and apparatus, and electronic device, medium and bone mass measurement system |
| CN113222875A (en) * | 2021-06-01 | 2021-08-06 | 浙江大学 | Image harmonious synthesis method based on color constancy |
| CN114137079A (en) * | 2021-11-30 | 2022-03-04 | 东莞理工学院 | Ultrasonic guided wave nondestructive testing method based on combination of deep learning and Duffing system |
| CN114494701A (en) * | 2022-02-14 | 2022-05-13 | 浙江大学 | Semantic segmentation method and device based on graph structure neural network |
| CN114998444A (en) * | 2022-05-20 | 2022-09-02 | 燕山大学 | Robot high-precision pose measurement system based on two-channel network |
Non-Patent Citations (7)
| Title |
|---|
| ELIF DERYA ÜBEYLI: "Probabilistic neural networks employing Lyapunov exponents for analysis of Doppler ultrasound signals", 《COMPUTERS IN BIOLOGY AND MEDICINE》, pages 82 - 89 * |
| YU WENG 等: "Nas-unet: Neural architecture search for medical image segmentation", 《IEEE ACCESS》, pages 44247 - 44257 * |
| ZHEN YANG 等: "An Overview of PCNN Model’s Development and Its Application in Image Processing", 《ARCHIVES OF COMPUTATIONAL METHODS IN ENGINEERING》, pages 491 - 505 * |
| 梅晓丹 等: "粗神经网络的禁止搜索训练算法研究", 《电子学报》, pages 1908 - 1911 * |
| 段亚玲 等: "基于混沌理论的医学超声影像提取研究", 《畜牧兽医杂志》, pages 82 - 84 * |
| 雍恩米 等: "一类受周期扰动航天器的混沌姿态运动", 《宇航学报》, pages 535 - 540 * |
| 马义德 等: "基于混沌脉冲耦合神经网络系统的信号跟踪", 《兰州大学学报(自然科学版)》, pages 133 - 136 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115713844A (en) * | 2022-10-18 | 2023-02-24 | 兰州交通大学 | Alarm method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115436881B (en) | 2023-07-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106355151B (en) | A kind of three-dimensional S AR images steganalysis method based on depth confidence network | |
| CN110849369A (en) | Robot tracking method, device, equipment and computer readable storage medium | |
| CN108872932B (en) | Beyond-visual-range target direct positioning result deviation rectifying method based on neural network | |
| CN115436881A (en) | Positioning method, system, computer equipment and readable storage medium | |
| CN111830493B (en) | System and method for forecasting strength of middle-high frequency sound target in underwater target receiving and transmitting division | |
| WO2017222385A1 (en) | An automotive testing method, system and computer program product | |
| CN107202976A (en) | The distributed microphone array sound source localization system of low complex degree | |
| CN106353722A (en) | RSSI (received signal strength indicator) distance measuring method based on cost-reference particle filter | |
| CN106323272B (en) | A kind of method and electronic equipment obtaining track initiation track | |
| CN107705359A (en) | The method and apparatus for obtaining wireless channel model | |
| US5488589A (en) | Neural network based three dimensional ocean modeler | |
| CN106842112A (en) | Sound localization method based on parametrization Bayes's dictionary learning under strong reverberant ambiance | |
| CN114223270B (en) | Training method and device for antenna signal processing model, antenna and storage medium | |
| CN113993205B (en) | UWB positioning system and method based on digital twinning | |
| CN112036011B (en) | Very low frequency wave communication transmission analysis method and system for underwater vehicle | |
| JP2020063958A (en) | Position estimation device and method | |
| CN107133385B (en) | Multi-mode synthetic aperture radar echo signal simulation method based on mode variation | |
| CN117241215A (en) | Wireless sensor network distributed node cooperative positioning method based on graph neural network | |
| CN115877377B (en) | Radar networking vector flow field synthesis method, system, equipment and storage medium | |
| CN108513355B (en) | Network positioning method, device and equipment based on single base station | |
| CN113835064B (en) | Weighted multi-dimensional scale TDOA (time difference of arrival) positioning method for cooperative correction source observation information | |
| CN113514796B (en) | Passive positioning method, system and medium | |
| CN115713844B (en) | Alarm method and system | |
| Wolfle et al. | A recursive model for the field strength prediction with neural networks | |
| Wang et al. | Optimization of Radio Frequency Identification Reference Tag Location Algorithm Based on Back Propagation Neural Network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |