CN115002903B

CN115002903B - Personnel track positioning system and testing method for internal space of offshore equipment

Info

Publication number: CN115002903B
Application number: CN202210895496.8A
Authority: CN
Inventors: 冯玮; 史大虎; 李明; 徐升; 李伟; 薛乃耀; 杨星驰; 刘廉昊
Original assignee: China International Marine Containers Engineering Co ltd
Current assignee: China International Marine Containers Engineering Co ltd
Priority date: 2022-07-28
Filing date: 2022-07-28
Publication date: 2022-11-15
Anticipated expiration: 2042-07-28
Also published as: CN115002903A

Abstract

The application relates to the field of intelligent positioning, and particularly discloses a personnel trajectory positioning system and a testing method for an internal space of offshore equipment.

Description

Personnel track positioning system and testing method for internal space of offshore equipment

Technical Field

The invention relates to the field of intelligent positioning, in particular to a personnel trajectory positioning system and a testing method for an internal space of offshore equipment.

Background

The positioning system is of great significance to the normal operation of offshore equipment (such as offshore drilling platforms and offshore transportation ships), and objects (such as workers) entering the internal space of the offshore equipment can be positioned by the positioning system, so that the workers can be positioned, tracked, guaranteed to be safe and prevented from being disconnected.

However, compared to positioning systems applied in open environments, positioning systems applied to the inner space of marine equipment have special challenges due to complex and variable hull structures, signal transmission shielding of closed spaces, numerous cross construction personnel, large-scale power interference, numerous field sundries and the like.

Therefore, an optimized personnel trajectory positioning system for the internal space of offshore equipment and a test solution thereof are expected.

Disclosure of Invention

The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides a personnel track positioning system and a testing method for an internal space of offshore equipment, which fuse the testing methods of RSS and TDOA through a feature fusion method based on an artificial intelligence positioning technology so as to accurately position an object entering the internal space of the offshore equipment, and further introduce position information of each communication base station to represent a communication environment between a beacon and the positioning base station in the process of fusing the two methods for positioning, so that the positioning accuracy is higher.

According to one aspect of the present application, there is provided a personnel trajectory positioning system for an interior space of offshore equipment, comprising: a communication data acquisition module, configured to acquire, from a plurality of positioning base stations in communication with a positioning beacon, communication data between the plurality of positioning base stations and the positioning beacon, where the communication data includes a signal strength value of a positioning signal sent by the positioning beacon and received by each positioning base station, and an arrival time of the positioning signal sent by the positioning beacon and received by each positioning base station; the signal intensity coding module is used for arranging the signal intensity values in the communication data into signal intensity input vectors and then obtaining signal intensity characteristic vectors through a sequence coder comprising a one-dimensional convolutional layer; the arrival time coding module is used for arranging arrival time in each communication data into an arrival time input vector and then obtaining an arrival time characteristic vector through the sequence coder containing the one-dimensional convolutional layer; the positioning data acquisition module is used for acquiring three-dimensional coordinates of each positioning base station in the plurality of positioning base stations in the internal space of the offshore equipment; the positioning data coding module is used for converting three-dimensional coordinates of each positioning base station in the internal space of the maritime equipment into position characteristic values through a full connection layer, and then arranging the position characteristic values of each positioning base station into position characteristic vectors; a feature distribution correction module, configured to perform feature value distribution location correction on the signal intensity feature vector and the arrival time feature vector respectively based on the location feature vector to obtain a corrected signal intensity feature vector and a corrected arrival time feature vector; the first correlation coding module is used for performing vector multiplication on the transposed vector of the position characteristic vector and the corrected signal intensity characteristic vector to obtain a position-intensity correlation matrix and then obtaining a first positioning characteristic diagram through a first convolution neural network serving as a characteristic extractor; the second correlation coding module is used for performing vector multiplication on the transposed vector of the position characteristic vector and the corrected arrival time characteristic vector to obtain a position-time correlation matrix and then obtaining a second positioning characteristic diagram through a second convolutional neural network serving as a characteristic extractor; a positioning feature fusion module, configured to fuse the first positioning feature map and the second positioning feature map to obtain a decoding feature map; and the positioning data generation module is used for carrying out regression decoding on the decoding characteristic diagram through a decoder to obtain a decoding value, and the decoding value is positioning data of the positioning beacon.

According to another aspect of the present application, a method for testing a personnel trajectory positioning system for an interior space of offshore equipment, comprising: acquiring communication data between a plurality of positioning base stations and a positioning beacon from the plurality of positioning base stations in communication with the positioning beacon, wherein the communication data comprises a signal strength value of a positioning signal sent by the positioning beacon and received by each positioning base station and an arrival time of the positioning signal sent by the positioning beacon and received by each positioning base station; arranging the signal intensity values in the communication data into signal intensity input vectors, and then obtaining signal intensity characteristic vectors through a sequence encoder comprising one-dimensional convolutional layers; arranging arrival time in each communication data into an arrival time input vector, and then obtaining an arrival time characteristic vector through the sequence encoder comprising the one-dimensional convolutional layer; acquiring three-dimensional coordinates of each positioning base station in the plurality of positioning base stations in the internal space of the offshore equipment; converting the three-dimensional coordinates of each positioning base station in the internal space of the maritime equipment into position characteristic values through a full connection layer, and arranging the position characteristic values of each positioning base station into position characteristic vectors; respectively carrying out characteristic value distribution position correction on the signal intensity characteristic vector and the arrival time characteristic vector based on the position characteristic vector to obtain a corrected signal intensity characteristic vector and a corrected arrival time characteristic vector; performing vector multiplication on the transposed vector of the position feature vector and the corrected signal intensity feature vector to obtain a position-intensity correlation matrix, and then obtaining a first positioning feature map through a first convolution neural network serving as a feature extractor; the transposed vector of the position characteristic vector and the corrected arrival time characteristic vector are subjected to vector multiplication to obtain a position-time correlation matrix, and then a second positioning characteristic diagram is obtained through a second convolution neural network serving as a characteristic extractor; fusing the first positioning feature map and the second positioning feature map to obtain a decoding feature map; and performing regression decoding on the decoding characteristic diagram through a decoder to obtain a decoding value, wherein the decoding value is positioning data of the positioning beacon.

Compared with the prior art, the personnel track positioning system and the testing method for the internal space of the offshore equipment fuse the RSS testing method and the TDOA testing method through the feature fusion method based on the artificial intelligence positioning technology so as to accurately position the object entering the internal space of the offshore equipment, and in the process of fusing the RSS testing method and the TDOA testing method for positioning, the position information of each communication base station is further introduced to represent the communication environment between the beacon and the positioning base station, so that the positioning accuracy is higher.

Drawings

The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally represent like parts or steps.

FIG. 1 is a general flow diagram of a project implementation of a personnel trajectory positioning system for an interior space of a maritime vessel according to an embodiment of the present application.

FIG. 2 is a block diagram of a personnel trajectory positioning system for an interior space of offshore equipment according to an embodiment of the application.

FIG. 3 is a block diagram of a feature distribution correction module in a personnel trajectory positioning system for an interior space of offshore equipment according to an embodiment of the application.

Fig. 4 is a flowchart of a testing method for a personnel trajectory positioning system for an interior space of a maritime vessel according to an embodiment of the application.

Fig. 5 is a schematic structural diagram of a testing method of a personnel trajectory positioning system for an internal space of offshore equipment according to an embodiment of the application.

Detailed Description

Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only a few embodiments of the present application, and not all embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.

Overview of scenes

As mentioned above, the positioning system is important for the normal operation of offshore equipment (e.g., offshore drilling platform, marine transport vessel), and the positioning system can be used to position an object (e.g., a worker) entering the internal space of the offshore equipment, so as to position, track and protect the worker from loss of connection.

Based on the method, a feasibility design scheme is provided through on-site actual reconnaissance, a customized positioning system is developed, all indexes of the real ship are tested synchronously, and finally verification implementation and reliable operation of the positioning system are carried out on the real ship project. The deployment of the positioning system is mainly to establish an indoor and outdoor seamless positioning network. The construction of a positioning network is realized by deploying the Bluetooth beacon, and the construction of a communication backhaul network is realized by deploying the positioning communication base station. The overall flow of project implementation is shown in fig. 1.

And summarizing the characteristics of the site environment through site reconnaissance: the ship body in the limited closed space signal transmission shielding, dismounting and mounting construction is complex and changeable in structure, numerous in cross construction personnel, large-scale power interference and more in-site sundries. By analyzing the actual environment on the site and combining the drawing of the ship body construction, typical scenes are selected for key tests, such as a cabin, a ballast tank, an oil tank and a deck.

The test method after the platform is installed is as follows:

1. signal coverage strength test method. 4 typical environments (ballast tank, oil tank, engine room and deck) are selected, two positioning tags walk into the testing environment to carry out signal coverage strength testing, and special signal detection software is used for data collection. Overall results of signal testing for both tags: excellent (dark green) and good (light green), it can be determined whether the signal condition of the personnel positioning system is satisfied.

2. A method for testing personnel positioning dynamic effect and positioning accuracy. And (3) selecting an environment with obvious spacial sense, respectively holding the cards by 2 persons, walking on the deck according to a specified path, and observing whether the actual walking path is consistent with the system analysis path. And (4) actually measuring and judging the positioning dynamic effect (the precision is within 3-5 meters, the judgment is excellent, and the precision is within 5-8 meters, and the judgment is good).

3. Function realization condition test method. The operation and the personnel are cooperated to carry out the drill in the field through a plurality of key department purposes. The effectiveness of the direct experience function, the accuracy of the result and the timeliness of the data. Specific functional subjects: the method comprises the steps of real-time positioning, track tracking, one-key help seeking, early warning management, data analysis, video linkage, people's identity verification and fire control.

4. A test method for multi-service fusion capability. The fusion utilization of system data and functions is met, 3 typical fusion applications are selected, and the functional effectiveness is verified through actual operation. And verifying the testimony of a witness, controlling fire, checking while, and checking the validity of the data.

Accordingly, the inventor of the present application considers that both RSS (signal strength analysis) and TDOA (time difference of arrival location) can well locate an object entering the internal space of the maritime equipment in the location technology, but both methods have respective advantages and disadvantages and are interfered by the communication environment, for example, RSS calculates the distance to be measured through a received signal according to model parameters, so that it requires a rigorous model design and has a limited coverage range, and is not generally used alone, and TDOA requires time synchronization between base stations. Therefore, in the technical solution of the present application, the inventor expects to integrate RSS and TDOA methods to improve positioning accuracy, and considers that if location information of each communication base station is further introduced to represent a communication environment between a beacon and a positioning base station, including that a ship structure is complex and changeable, a closed space signal transmission is shielded, cross construction staff is numerous, large-scale dynamic interference, field sundries are numerous, and the like, characteristic information of the communication environment can be introduced into the positioning solution to perform positioning, so as to improve positioning accuracy.

Specifically, in the technical solution of the present application, first, communication data between a plurality of positioning base stations and a positioning beacon is obtained from the plurality of positioning base stations in communication with the positioning beacon, where the communication data includes a signal strength value of a positioning signal sent by the positioning beacon received by each positioning base station and an arrival time of the positioning signal sent by the positioning beacon received by each positioning base station. Considering that interference and relevance exist between communication data between the plurality of positioning base stations and the positioning beacons, in order to improve the accuracy of the measured data and more accurately position an object entering an internal space of maritime equipment, a sequence encoder comprising a one-dimensional convolutional layer is further used for respectively carrying out sequence encoding on a signal strength value in each communication data and an arrival time in each communication data so as to respectively extract local association implicit characteristic information of the signal strength value in each communication data and local association implicit characteristic information of the arrival time in each communication data, and therefore a signal strength characteristic vector and an arrival time characteristic vector are obtained.

It should be understood that, in order to obtain a positioning result more accurately, position information of each communication base station needs to be introduced to represent a communication environment between the positioning beacon and the positioning base station, for example, the communication environment includes that a ship structure is complex and varied, a closed space signal transmission shield, a lot of cross construction personnel, large power interference, a lot of field sundries, and the like, so as to express interference of communication signals. Specifically, first, three-dimensional coordinates of each of the plurality of positioning base stations in the internal space of the marine equipment are acquired. And then converting the three-dimensional coordinates of each positioning base station in the internal space of the offshore equipment into position characteristic values through a full connection layer, and arranging the position characteristic values of each positioning base station into position characteristic vectors.

In order to obtain the location feature information of each communication positioning base station and the implicit correlation feature of the signal strength value in each communication data and the correlation feature distribution information between the implicit correlation features of the arrival time in each communication data, a correlation matrix of the location feature vector and the signal strength feature vector and the arrival time feature vector respectively is further constructed. And respectively extracting features of the constructed position-intensity correlation matrix and position-time correlation matrix in a convolutional neural network model as a feature extractor to respectively extract a high-dimensional correlation feature of the position of the communication positioning base station and the signal strength value of the positioning signal, namely, a signal strength feature of the signal strength value of the positioning signal sent by the positioning beacon under the interference of a communication environment to obtain a first positioning feature map, and extract a high-dimensional correlation of the position of the communication positioning base station and the arrival time of the positioning signal sent by the positioning beacon, namely, a time feature of the arrival time of the positioning signal sent by the positioning beacon under the interference of the communication environment to obtain a second positioning feature map. Thus, when positioning is performed by the RSS positioning method and the TDOA positioning method, the positioning data of the positioning beacon can be obtained more accurately in consideration of the influence of interference of the environment.

Further, the first positioning feature map and the second positioning feature map are fused for decoding, so that a decoded value representing the positioning data of the positioning beacon can be obtained.

However, when fusing the first and second positioning feature maps, it is desirable to align their feature distributions as much as possible to obtain a better fusion effect. Also, since the first and second localization feature maps have a common dimension of a location feature vector, it is apparent that the above object can be achieved if the signal intensity feature vector and the temporal feature vector are pre-aligned with the location feature vector before feature extraction.

And, since the signal strength feature vector and the time feature vector are both deep coding features obtained by a sequence encoder, the signal strength feature vector is encoded

And the temporal feature vector

Using feature vectors based on said position

The hierarchical depth homography correction of (1), namely:

wherein

A feature vector representing the strength of the signal,

a feature vector representing the time of arrival,

a feature vector representing the location of the object is provided,

a feature vector representing the corrected signal strength,

representing the corrected arrival time feature vector,

represents a norm of a vector, and

representing the Frobenius norm of the matrix,

the difference in terms of position is indicated,

indicating a multiplication by a point in the position,

indicating a sum by position.

Here, the hierarchical depth homography alignment is to perform homography alignment of a scene depth stream based on vector differential expression according to hierarchical depth characteristics of vector fusion characterization, and superimpose a full scene homography incidence matrix as an offset, so as to implement the signal strength feature vector

And the temporal feature vector

Respectively with the position feature vector

Thereby reducing the signal strength feature vector

And the temporal feature vector

The characteristic distribution is staggered due to deep sequence coding, and the decoding accuracy is further improved.

Based on this, the present application proposes a personnel trajectory positioning system for an internal space of offshore equipment, comprising: a communication data acquisition module, configured to acquire, from a plurality of positioning base stations in communication with a positioning beacon, communication data between the plurality of positioning base stations and the positioning beacon, where the communication data includes a signal strength value of a positioning signal sent by the positioning beacon received by each positioning base station and an arrival time at which each positioning base station receives the positioning signal sent by the positioning beacon; the signal intensity coding module is used for arranging the signal intensity values in the communication data into signal intensity input vectors and then obtaining signal intensity characteristic vectors through a sequence coder comprising a one-dimensional convolutional layer; the arrival time coding module is used for arranging arrival time in each communication data into an arrival time input vector and then obtaining an arrival time characteristic vector through the sequence coder containing the one-dimensional convolutional layer; the position data acquisition module is used for acquiring three-dimensional coordinates of each positioning base station in the plurality of positioning base stations in the internal space of the maritime equipment; the positioning data coding module is used for converting three-dimensional coordinates of each positioning base station in the internal space of the maritime equipment into position characteristic values through a full connection layer, and then arranging the position characteristic values of each positioning base station into position characteristic vectors; a feature distribution correction module, configured to perform feature value distribution location correction on the signal intensity feature vector and the arrival time feature vector respectively based on the location feature vector to obtain a corrected signal intensity feature vector and a corrected arrival time feature vector; the first correlation coding module is used for performing vector multiplication on the transposed vector of the position characteristic vector and the corrected signal intensity characteristic vector to obtain a position-intensity correlation matrix and then obtaining a first positioning characteristic diagram through a first convolution neural network serving as a characteristic extractor; the second correlation coding module is used for performing vector multiplication on the transposed vector of the position characteristic vector and the corrected arrival time characteristic vector to obtain a position-time correlation matrix and then obtaining a second positioning characteristic diagram through a second convolutional neural network serving as a characteristic extractor; a positioning feature fusion module, configured to fuse the first positioning feature map and the second positioning feature map to obtain a decoding feature map; and the positioning data generation module is used for performing regression decoding on the decoding characteristic diagram through a decoder to obtain a decoding value, wherein the decoding value is the positioning data of the positioning beacon.

Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.

Exemplary System

FIG. 2 illustrates a block diagram of a personnel trajectory positioning system for an internal space of offshore equipment according to an embodiment of the application. As shown in fig. 2, a person trajectory positioning system 200 for an internal space of offshore equipment according to an embodiment of the present application includes: a communication data acquisition module 210, configured to acquire, from a plurality of positioning base stations in communication with a positioning beacon, communication data between the plurality of positioning base stations and the positioning beacon, where the communication data includes a signal strength value of a positioning signal sent by the positioning beacon received by each positioning base station and an arrival time of the positioning signal sent by the positioning beacon received by each positioning base station; a signal strength encoding module 220, configured to arrange the signal strength values in each piece of communication data into a signal strength input vector, and then obtain a signal strength feature vector through a sequence encoder including a one-dimensional convolutional layer; an arrival time encoding module 230, configured to arrange arrival times in the communication data into arrival time input vectors, and then obtain arrival time feature vectors through the sequence encoder including the one-dimensional convolutional layer; a position data acquisition module 240, configured to acquire a three-dimensional coordinate of each of the plurality of positioning base stations in an internal space of the marine equipment; the position data encoding module 250 is configured to convert three-dimensional coordinates of each positioning base station in the internal space of the marine equipment into position feature values through a full connection layer, and arrange the position feature values of each positioning base station into position feature vectors; a feature distribution correction module 260, configured to perform, based on the location feature vector, feature value distribution location correction on the signal intensity feature vector and the arrival time feature vector respectively to obtain a corrected signal intensity feature vector and a corrected arrival time feature vector; a first association encoding module 270, configured to perform vector multiplication on the transposed vector of the position feature vector and the corrected signal strength feature vector to obtain a position-strength association matrix, and then obtain a first positioning feature map through a first convolutional neural network serving as a feature extractor; a second correlation encoding module 280, configured to perform vector multiplication on the transposed vector of the position feature vector and the corrected arrival time feature vector to obtain a position-time correlation matrix, and then obtain a second positioning feature map through a second convolutional neural network serving as a feature extractor; a positioning feature fusion module 290, configured to fuse the first positioning feature map and the second positioning feature map to obtain a decoding feature map; and a positioning data generating module 300, configured to perform regression decoding on the decoded feature map through a decoder to obtain a decoded value, where the decoded value is positioning data of a positioning beacon.

Specifically, in this embodiment of the present application, the communication data collection module 210, the signal strength coding module 220, and the arrival time coding module 230 are configured to obtain, from a plurality of positioning base stations in communication with a positioning beacon, communication data between the plurality of positioning base stations and the positioning beacon, where the communication data includes a signal strength value of a positioning signal sent by the positioning beacon received by each positioning base station and an arrival time of a positioning signal sent by the positioning beacon received by each positioning base station, arrange the signal strength values in each communication data into a signal strength input vector, obtain a signal strength feature vector through a sequence encoder including a one-dimensional convolutional layer, and arrange the arrival time in each communication data into an arrival time input vector, and obtain an arrival time feature vector through the sequence encoder including a one-dimensional convolutional layer. As described above, in the positioning technology, it is considered that RSS (signal strength analysis) and TDOA (time difference of arrival positioning) can well locate an object entering the internal space of the maritime equipment, but both methods have their own advantages and disadvantages and are interfered by the communication environment, for example, RSS calculates the distance of a point to be measured from a received signal according to model parameters, so that it requires a rigorous model design and has a limited coverage range, and is not generally used alone, whereas TDOA requires time synchronization between base stations. Therefore, in the technical solution of the present application, it is desirable to integrate RSS and TDOA methods to improve the positioning accuracy, and in consideration of the fact that if the position information of each communication base station is further introduced to represent the communication environment between the beacon and the positioning base station, the communication environment includes a complex and variable hull structure, a closed space signal transmission shielding, a large number of cross constructors, large power interference, a large number of field sundries, and the like, the characteristic information of the communication environment can be introduced into the positioning scheme to perform positioning, thereby improving the positioning accuracy.

More specifically, in this embodiment of the application, the signal strength encoding module is further configured to: arranging the signal strength values in each of the communication data into the signal strength input vector; performing full-concatenation encoding on the signal strength input vector by using a full-concatenation layer of the sequence encoder according to the following formula to extract high-dimensional implicit features of feature values of each position in the signal strength input vector, wherein the formula is as follows:

in which

Is the input vector of the said one or more input vectors,

is the output vector of the output vector,

is a matrix of the weights that is,

is a vector of the offset to the offset,

represents a matrix multiplication; performing one-dimensional convolution encoding on the signal intensity input vector by using a one-dimensional convolution layer of the sequence encoder according to the following formula so as to extract high-dimensional implicit correlation characteristics among characteristic values of all positions in the signal intensity input vector, wherein the formula is as follows:

wherein the content of the first and second substances,ais a convolution kernelxWidth in the direction,FIs a convolution kernel parameter vector,GIs a matrix of local vectors operating with a convolution kernel,wis the size of the convolution kernel and,

representing the signal input intensity vector.

More specifically, in this embodiment of the application, the arrival time encoding module is further configured to: arranging arrival times in the respective communication data into the arrival time input vector; using a full-concatenation layer of the sequence encoder to perform full-concatenation encoding on the arrival time input vector by using the following formula to extract high-dimensional implicit features of feature values of each position in the arrival time input vector, wherein the formula is as follows:

in which

Is the input vector of the said one or more input vectors,

is the output vector of the digital video signal,

is a matrix of weights that is a function of,

is a vector of the offset to the offset,

represents a matrix multiplication; performing one-dimensional convolutional encoding on the arrival time input vector by using a one-dimensional convolutional layer of the sequence encoder according to the following formula so as to extract high-dimensional implicit correlation characteristics among characteristic values of all positions in the arrival time input vector, wherein the formula is as follows:

wherein, the first and the second end of the pipe are connected with each other,ais a convolution kernelxWidth in the direction,FIs a convolution kernel parameter vector,GIs a matrix of local vectors operating with a convolution kernel,wis the size of the convolution kernel and,

representing the arrival time input vector.

Specifically, in this embodiment of the application, the position data acquisition module 240 and the position data encoding module 250 are configured to acquire a three-dimensional coordinate of each of the multiple positioning base stations in an internal space of the marine equipment, convert the three-dimensional coordinate of each of the positioning base stations in the internal space of the marine equipment into a position characteristic value through a full connection layer, and arrange the position characteristic value of each of the positioning base stations into a position characteristic vector. It should be understood that, in order to obtain a positioning result more accurately, position information of each communication base station needs to be introduced to represent a communication environment between the positioning beacon and the positioning base station, for example, the communication environment includes that a ship structure is complex and varied, a closed space signal transmission shield, a lot of cross construction personnel, large power interference, a lot of field sundries, and the like, so as to express interference of communication signals. Specifically, first, three-dimensional coordinates of each of the plurality of positioning base stations in the internal space of the marine equipment are acquired. And then converting the three-dimensional coordinates of each positioning base station in the internal space of the offshore equipment into position characteristic values through a full connection layer, and arranging the position characteristic values of each positioning base station into position characteristic vectors. In this way, the position feature vector can be generated by using the global correlation feature information of the three-dimensional coordinates of each positioning base station in the internal space of the maritime equipment, so as to introduce global communication environment information for correlation, and the positioning accuracy can be improved.

Specifically, in this embodiment of the present application, the feature distribution correction module 260 is configured to perform feature value distribution location correction on the signal strength feature vector and the arrival time feature vector respectively based on the location feature vector to obtain a corrected signal strength feature vector and a corrected arrival time feature vector. It should be understood that when subsequently fusing the first positioning feature map and the second positioning feature map with associated feature information, it is desirable to align the feature distributions as much as possible to obtain a better fusion effect. Also, since the first and second localization feature maps have a common dimension of a location feature vector, it is apparent that the above object can be achieved if the signal intensity feature vector and the temporal feature vector are pre-aligned with the location feature vector before feature extraction. In addition, since the signal strength feature vector and the time feature vector are deep coding features obtained by a sequence encoder, in the technical solution of the present application, the signal strength feature vector is subjected to deep coding

And the arrival time feature vector

Using feature vectors based on said position

And (4) correcting the homography alignment of the layering depth. It should be understood that, here, the hierarchical depth homography alignment is performed by performing homography alignment of scene depth streams based on vector differential expression according to hierarchical depth characteristics characterized by vector fusion, and superimposing a full scene homography incidence matrix as a bias, so as to realize the signal strength feature vector

And the arrival time feature vector

Respectively with the position feature vector

Thereby reducing the signal strength feature vector

And the arrival time feature vector

The characteristic distribution is misplaced due to deep sequence coding, thereby improving the decoding accuracy.

More specifically, in this embodiment, the feature distribution correction module includes: firstly, performing hierarchical depth homography alignment correction on the signal strength eigenvector based on the position eigenvector to obtain the corrected signal strength eigenvector, wherein the hierarchical depth homography alignment correction is performed based on a difference eigenvector between the position eigenvector and the signal strength eigenvector and a Frobenius norm of a full-scene homography correlation matrix between the position eigenvector and the signal strength eigenvector, wherein the full-scene homography correlation matrix between the position eigenvector and the signal strength eigenvector is a product between a transposed vector of the position eigenvector and the signal strength eigenvector. Accordingly, in a specific example, based on the position feature vector, the signal strength feature vector is subjected to hierarchical depth homography alignment correction according to the following formula to obtain the corrected signal strength feature vector; wherein the formula is:

wherein

A feature vector representing the strength of the signal,

a feature vector representing the location is generated,

representing the corrected signal strength feature vector,

represents a norm of a vector, and

representing the Frobenius norm of the matrix,

the difference in terms of position is indicated,

indicating that the multiplication is performed by a point at a position,

indicating a sum by position.

Then, performing hierarchical depth homography alignment correction on the arrival time feature vector based on the position feature vector to obtain the corrected arrival time feature vector, wherein the hierarchical depth homography alignment correction is performed based on a difference feature vector between the position feature vector and the arrival time feature vector and a Frobenius norm of a full-scene homography correlation matrix between the position feature vector and the arrival time feature vector, and the full-scene homography correlation matrix between the position feature vector and the arrival time feature vector is a product between a transposed vector of the position feature vector and the arrival time feature vector. Accordingly, in a specific example, the arrival time feature vector is subjected to hierarchical depth homography alignment correction based on the position feature vector according to the following formula to obtain the corrected arrival time feature vector;

wherein the formula is:

wherein

A feature vector representing the time of arrival,

a feature vector representing the location is generated,

representing the corrected arrival time feature vector,

represents a norm of a vector, an

Represents the Frobenius norm of the matrix,

the difference in terms of position is indicated,

indicating a multiplication by a point in the position,

indicating a sum by position.

FIG. 3 illustrates a block diagram of a feature distribution correction module in a method for financial reimbursement invoice management based on blockchain techniques according to an embodiment of the present application. As shown in fig. 3, the feature distribution correction module 260 includes: a first correction unit 261, configured to perform hierarchical depth homography alignment correction on the signal strength eigenvector based on the position eigenvector to obtain the corrected signal strength eigenvector, where the hierarchical depth homography alignment correction is performed based on a difference eigenvector between the position eigenvector and the signal strength eigenvector and a Frobenius norm of a full-scene homography correlation matrix between the position eigenvector and the signal strength eigenvector, where the full-scene homography correlation matrix between the position eigenvector and the signal strength eigenvector is a product between a transposed vector of the position eigenvector and the signal strength eigenvector; a second correcting unit 262, configured to perform hierarchical depth homographic correction on the arrival time feature vector based on the position feature vector to obtain the corrected arrival time feature vector, where the hierarchical depth homographic correction is performed based on a difference feature vector between the position feature vector and the arrival time feature vector and a Frobenius norm of a full scene homographic association matrix between the position feature vector and the arrival time feature vector, where the full scene homographic association matrix between the position feature vector and the arrival time feature vector is a product between a transposed vector of the position feature vector and the arrival time feature vector.

Specifically, in this embodiment of the application, the first correlation encoding module 270 and the second correlation encoding module 280 are configured to perform vector multiplication on the transposed vector of the position feature vector and the corrected signal strength feature vector to obtain a position-strength correlation matrix, obtain a first positioning feature map through a first convolutional neural network serving as a feature extractor, perform vector multiplication on the transposed vector of the position feature vector and the corrected arrival time feature vector to obtain a position-time correlation matrix, and obtain a second positioning feature map through a second convolutional neural network serving as a feature extractor. It should be understood that, in order to obtain the correlation feature distribution information between the location feature information of each of the communication positioning base stations and the implicit correlation feature of the signal strength value in each of the communication data and the implicit correlation feature of the arrival time in each of the communication data, the correlation matrix of the location feature vector and the signal strength feature vector and the arrival time feature vector, respectively, is further constructed. And respectively extracting features of the constructed position-intensity correlation matrix and position-time correlation matrix in a convolutional neural network model as a feature extractor to respectively extract a high-dimensional correlation feature of the position of the communication positioning base station and the signal strength value of the positioning signal, namely, a signal strength feature of the signal strength value of the positioning signal sent by the positioning beacon under the interference of a communication environment to obtain a first positioning feature map, and extract a high-dimensional correlation of the position of the communication positioning base station and the arrival time of the positioning signal sent by the positioning beacon, namely, a time feature of the arrival time of the positioning signal sent by the positioning beacon under the interference of the communication environment to obtain a second positioning feature map. Thus, when positioning is performed by the RSS positioning method and the TDOA positioning method, the positioning data of the positioning beacon can be obtained more accurately in consideration of the influence of the environmental interference.

More specifically, in this embodiment of the present application, the first association encoding module includes: a first correlation matrix constructing unit, configured to perform vector multiplication on the transposed vector of the position feature vector and the corrected signal strength feature vector by using the following formula to obtain the position-strength correlation matrix; wherein the formula is:

wherein

For the feature vector of the position, the position is,

for the corrected signal strength feature vector,

representing the location-intensity correlation moment,

representing a vector multiplication. A first convolution encoding unit, configured to perform convolution processing, pooling processing, and activation processing on input data in forward pass of layers using layers of the first convolution neural network as a feature extractor to generate the first location feature map from a last layer of the first convolution neural network, wherein an input of the first layer of the first convolution neural network is the position-intensity correlation matrix.

More specifically, in this embodiment of the present application, the second correlation encoding module includes: a first correlation matrix constructing unit, configured to perform vector multiplication on the transposed vector of the position feature vector and the corrected arrival time feature vector by using the following formula to obtain the position-time correlation matrix; wherein the formula is:

wherein

For the feature vector of the position, the position is,

is a stand forThe corrected arrival time feature vector is described below,

representing the position-time correlation matrix and,

representing vector multiplication. A first convolution encoding unit, configured to perform convolution processing, pooling processing, and activation processing on input data in forward pass of layers using layers of the second convolutional neural network as a feature extractor to generate the second localization feature map from a last layer of the second convolutional neural network, wherein an input of the first layer of the second convolutional neural network is the position-time correlation matrix.

Specifically, in this embodiment of the present application, the positioning feature fusing module 290 and the positioning data generating module 300 are configured to fuse the first positioning feature map and the second positioning feature map to obtain a decoded feature map, and perform regression decoding on the decoded feature map through a decoder to obtain a decoded value, where the decoded value is positioning data of a positioning beacon. That is, in the technical solution of the present application, further, the first positioning feature map and the second positioning feature map are fused and decoded, so that a decoded value representing the positioning data of the positioning beacon can be obtained.

More specifically, in this embodiment of the present application, the localization feature fusion module includes: fusing the first positioning feature map and the second positioning feature map according to the following formula to obtain the decoding feature map; wherein the formula is:

wherein, the first and the second end of the pipe are connected with each other,

for the purpose of the decoding of the feature map,

in order to be the first positioning feature map,

for the second positioning feature map "

"means the addition of elements at the corresponding positions of the first and second localization profiles,

is a weighting parameter for controlling a balance between the first localization profile and the second localization profile in the decoding profile.

In particular, in the embodiment of the application, a plurality of positioning data of a plurality of positions of a positioning beacon in the internal space of the maritime equipment are obtained based on the personnel trajectory positioning system for the internal space of the maritime equipment, which is described in any one of the modules; and generating a test result based on a comparison between real position data of a plurality of positions of the positioning beacon within the maritime equipment interior space and the plurality of positioning data. Accordingly, in one specific example, first, real position data of a plurality of positions of the positioning beacon within the maritime equipment interior space is constructed as first trajectory data; then, constructing the plurality of positioning data into second track data; then, calculating a Euclidean distance between the first track data and the second track data; finally, the test result is generated based on a comparison between the euclidean distance and a predetermined threshold.

In summary, the personnel trajectory positioning system 200 for the maritime equipment internal space based on the embodiment of the present application is illustrated, which fuses the testing methods of both RSS and TDOA through the feature fusion method based on the artificial intelligence positioning technology to accurately position the object entering the maritime equipment internal space, and further introduces the location information of each communication base station to characterize the communication environment between the beacon and the positioning base station in the process of fusing the two methods for positioning, so as to make the positioning accuracy higher.

As described above, the person trajectory positioning system 200 for the maritime equipment internal space according to the embodiment of the present application may be implemented in various terminal devices, such as a server of a person trajectory positioning algorithm for the maritime equipment internal space, and the like. In one example, the personnel trajectory locating system 200 for an interior space of a maritime vessel according to an embodiment of the application may be integrated into a terminal device as one software module and/or hardware module. For example, the person trajectory positioning system 200 for the maritime equipment interior space may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the personnel trajectory locating system 200 for the interior space of the offshore facility can also be one of many hardware modules of the terminal device.

Alternatively, in another example, the person trajectory positioning system 200 for the maritime equipment interior space and the terminal device may also be separate devices, and the person trajectory positioning system 200 for the maritime equipment interior space may be connected to the terminal device through a wired and/or wireless network and transmit the interactive information according to an agreed data format.

Exemplary method

FIG. 4 illustrates a flow chart of a method of testing a personnel trajectory positioning system for an interior space of a marine installation. As shown in fig. 4, a method for testing a personnel trajectory positioning system for an internal space of offshore equipment according to an embodiment of the application includes the steps of: s110, obtaining, from a plurality of positioning base stations in communication with a positioning beacon, communication data between the plurality of positioning base stations and the positioning beacon, where the communication data includes a signal strength value of a positioning signal sent by the positioning beacon received by each positioning base station and an arrival time of the positioning signal sent by the positioning beacon received by each positioning base station; s120, arranging the signal intensity values in the communication data into signal intensity input vectors, and then obtaining signal intensity characteristic vectors through a sequence encoder comprising one-dimensional convolutional layers; s130, arranging arrival time in each communication data into an arrival time input vector, and then obtaining an arrival time characteristic vector through the sequence encoder containing the one-dimensional convolutional layer; s140, acquiring three-dimensional coordinates of each positioning base station in the plurality of positioning base stations in the internal space of the maritime equipment; s150, converting the three-dimensional coordinates of each positioning base station in the internal space of the maritime equipment into position characteristic values through a full connection layer, and arranging the position characteristic values of each positioning base station into position characteristic vectors; s160, respectively carrying out characteristic value distribution position correction on the signal intensity characteristic vector and the arrival time characteristic vector based on the position characteristic vector to obtain a corrected signal intensity characteristic vector and a corrected arrival time characteristic vector; s170, performing vector multiplication on the transposed vector of the position feature vector and the corrected signal intensity feature vector to obtain a position-intensity correlation matrix, and then obtaining a first positioning feature map through a first convolution neural network serving as a feature extractor; s180, performing vector multiplication on the transposed vector of the position feature vector and the corrected arrival time feature vector to obtain a position-time correlation matrix, and then obtaining a second positioning feature map through a second convolutional neural network serving as a feature extractor; s190, fusing the first positioning feature map and the second positioning feature map to obtain a decoding feature map; and S200, performing regression decoding on the decoding characteristic diagram through a decoder to obtain a decoding value, wherein the decoding value is positioning data of the positioning beacon.

Fig. 5 illustrates an architectural diagram of a testing method for a personnel trajectory positioning system for an interior space of offshore equipment according to an embodiment of the application. As shown in fig. 5, in the network architecture of the test method for the personnel trajectory positioning system for maritime equipment internal space, firstly, after arranging the signal strength values (e.g., P1 as illustrated in fig. 5) in each of the obtained communication data into a signal strength input vector (e.g., V1 as illustrated in fig. 5), a signal strength feature vector (e.g., VF1 as illustrated in fig. 5) is obtained by a sequence encoder (e.g., E as illustrated in fig. 5) comprising one-dimensional convolutional layers; then, arranging the arrival time (for example, P2 as illustrated in fig. 5) in each of the obtained communication data as an arrival time input vector (for example, V2 as illustrated in fig. 5) and then passing through the sequence encoder (for example, E as illustrated in fig. 5) including one-dimensional convolutional layers to obtain an arrival time feature vector (for example, VF2 as illustrated in fig. 5); then, after converting the obtained three-dimensional coordinates (e.g., Q as illustrated in fig. 5) of each of the positioning base stations in the internal space of the offshore equipment into position characteristic values (e.g., CV as illustrated in fig. 5) through a fully connected layer (e.g., FCL as illustrated in fig. 5), arranging the position characteristic values of each of the positioning base stations into a position characteristic vector (e.g., VF as illustrated in fig. 5); then, respectively performing eigenvalue distribution position correction on the signal intensity eigenvector and the arrival time eigenvector based on the position eigenvector to obtain a corrected signal intensity eigenvector (e.g., VC1 as illustrated in fig. 5) and a corrected arrival time eigenvector (e.g., VC2 as illustrated in fig. 5); then, vector-multiplying the transposed vector of the position feature vector and the corrected signal strength feature vector to obtain a position-strength correlation matrix (e.g., MF1 as illustrated in fig. 5), and then passing through a first convolutional neural network (e.g., CNN1 as illustrated in fig. 5) as a feature extractor to obtain a first positioning feature map (e.g., F1 as illustrated in fig. 5); then, the transposed vector of the position feature vector and the corrected arrival time feature vector are subjected to vector multiplication to obtain a position-time correlation matrix (for example, MF2 as illustrated in fig. 5), and then a second positioning feature map (for example, F2 as illustrated in fig. 5) is obtained through a second convolutional neural network (for example, CNN2 as illustrated in fig. 5) as a feature extractor; then, fusing the first positioning feature map and the second positioning feature map to obtain a decoding feature map (e.g., F as illustrated in fig. 5); and, finally, performing regression decoding on the decoded feature map by a decoder (e.g., D as illustrated in fig. 5) to obtain a decoded value (e.g., D as illustrated in fig. 5), which is the positioning data of the positioning beacon.

More specifically, in step S110, step S120, and step S130, communication data between the plurality of positioning base stations and the positioning beacon is obtained from a plurality of positioning base stations communicating with the positioning beacon, where the communication data includes a signal strength value of a positioning signal sent by the positioning beacon received by each positioning base station and an arrival time of the positioning signal sent by the positioning beacon received by each positioning base station, the signal strength value in each communication data is arranged as a signal strength input vector, and then a signal strength eigenvector is obtained by passing through a sequence encoder including a one-dimensional convolutional layer, and then the arrival time in each communication data is arranged as an arrival time input vector, and then a arrival time eigenvector is obtained by passing through the sequence encoder including a one-dimensional convolutional layer. It should be understood that considering that both RSS (signal strength analysis) and TDOA (time difference of arrival location) are good for locating objects entering the internal space of offshore equipment in the positioning technology, but both methods have their own advantages and disadvantages and are subject to interference of communication environment, for example, RSS calculates the distance to be measured from received signals according to model parameters, so that it requires rigorous model design and has limited coverage, and is not generally used alone, while TDOA requires time synchronization between base stations. Therefore, in the technical solution of the present application, it is desirable to fuse RSS and TDOA methods to improve positioning accuracy, and in consideration of the fact that if location information of each communication base station is further introduced to indicate a communication environment between a beacon and a positioning base station, the communication environment includes a complex and variable hull structure, a closed space signal transmission shielding, numerous cross constructors, large power interference, numerous field sundries, and the like, characteristic information of the communication environment can be introduced into the positioning scheme to perform positioning, so as to improve positioning accuracy.

More specifically, in step S140 and step S150, three-dimensional coordinates of each of the plurality of positioning base stations in the internal space of the marine equipment are obtained, and after the three-dimensional coordinates of each of the plurality of positioning base stations in the internal space of the marine equipment are converted into position characteristic values through a full connection layer, the position characteristic values of each of the plurality of positioning base stations are arranged as a position characteristic vector. It should be understood that, in order to obtain a positioning result more accurately, position information of each of the communication base stations needs to be introduced to represent a communication environment between the positioning beacon and the positioning base station, for example, a ship structure is complicated and varied, a closed space signal transmission shielding is adopted, a number of cross construction workers is increased, large power interference is caused, a number of field sundries is increased, and the like, so as to express interference of communication signals. Specifically, first, three-dimensional coordinates of each of the plurality of positioning base stations in the internal space of the marine equipment are acquired. And then, converting the three-dimensional coordinates of each positioning base station in the internal space of the maritime equipment into position characteristic values through a full connection layer, and arranging the position characteristic values of each positioning base station into position characteristic vectors. In this way, the position feature vector can be generated by using the global correlation feature information of the three-dimensional coordinates of each positioning base station in the internal space of the maritime equipment, so as to introduce global communication environment information for correlation, and the positioning accuracy can be improved.

More specifically, in step S160, feature value distribution position correction is performed on the signal strength feature vector and the arrival time feature vector based on the position feature vector to obtain a corrected signal strength feature vector and a corrected arrival time feature vector, respectively. It should be understood that when subsequently fusing the first positioning feature map and the second positioning feature map with associated feature information, it is desirable to align their feature distributions as much as possible to obtain a better fusion effect. Also, since the first and second positioning feature maps have a common dimension of a position feature vector, it is apparent that the above object can be achieved if the signal intensity feature vector and the time feature vector are pre-aligned with the position feature vector before feature extraction. In addition, since the signal strength feature vector and the time feature vector are deep coding features obtained by a sequence encoder, in the technical solution of the present application, the signal strength feature vector is subjected to deep coding

And the arrival time feature vector

Using feature vectors based on said position

And (4) performing homography correction on the layering depth. It should be understood that, here, the hierarchical depth homography alignment is performed by performing homography alignment of scene depth streams based on vector differential expression according to hierarchical depth characteristics characterized by vector fusion, and superimposing a full scene homography incidence matrix as a bias, so as to realize the signal strength feature vector

And said time of arrivalInter feature vector

Respectively with the position feature vector

Thereby reducing the signal strength feature vector

And the arrival time feature vector

More specifically, in step S170 and step S180, the transposed vector of the position feature vector and the corrected signal strength feature vector are vector-multiplied to obtain a position-strength correlation matrix, and then the position-strength correlation matrix is obtained through a first convolutional neural network as a feature extractor, and the transposed vector of the position feature vector and the corrected arrival time feature vector are vector-multiplied to obtain a position-time correlation matrix, and then the position-time correlation matrix is obtained through a second convolutional neural network as a feature extractor, so as to obtain a second positioning feature map. It should be understood that, in order to obtain the associated feature distribution information between the location feature information of each of the communication positioning base stations and the implicit associated feature of the signal strength value in each of the communication data and the implicit associated feature of the arrival time in each of the communication data, the association matrix of the location feature vector and the signal strength feature vector and the arrival time feature vector, respectively, is further constructed. And respectively extracting features of the constructed position-intensity correlation matrix and position-time correlation matrix in a convolutional neural network model as a feature extractor to respectively extract a high-dimensional correlation feature of the position of the communication positioning base station and the signal strength value of the positioning signal, namely, a signal strength feature of the signal strength value of the positioning signal sent by the positioning beacon under the interference of a communication environment to obtain a first positioning feature map, and extract a high-dimensional correlation of the position of the communication positioning base station and the arrival time of the positioning signal sent by the positioning beacon, namely, a time feature of the arrival time of the positioning signal sent by the positioning beacon under the interference of the communication environment to obtain a second positioning feature map. Thus, when positioning is performed by the RSS positioning method and the TDOA positioning method, the positioning data of the positioning beacon can be obtained more accurately in consideration of the influence of the environmental interference.

More specifically, in step S190 and step S200, the first positioning feature map and the second positioning feature map are fused to obtain a decoding feature map, and the decoding feature map is subjected to regression decoding by a decoder to obtain a decoded value, where the decoded value is the positioning data of the positioning beacon. That is, in the technical solution of the present application, further, the first positioning feature map and the second positioning feature map are fused and decoded, so that a decoded value representing the positioning data of the positioning beacon can be obtained.

In summary, the testing method for the personnel trajectory positioning system for the maritime equipment internal space based on the embodiment of the application is clarified, the testing method for the RSS and the TDOA is fused through a feature fusion method based on an artificial intelligence positioning technology so as to accurately position the object entering the maritime equipment internal space, and in the process of fusing the two methods for positioning, the position information of each communication base station is further introduced to represent the communication environment between the beacon and the positioning base station, so that the positioning accuracy is higher.

The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.

The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by one skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. The words "or" and "as used herein mean, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".

It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations are to be considered as equivalents of the present application.

The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A personnel trajectory positioning system for an interior space of offshore equipment, comprising:

a communication data acquisition module, configured to acquire, from a plurality of positioning base stations in communication with a positioning beacon, communication data between the plurality of positioning base stations and the positioning beacon, where the communication data includes a signal strength value of a positioning signal sent by the positioning beacon received by each positioning base station and an arrival time at which each positioning base station receives the positioning signal sent by the positioning beacon;

the signal intensity coding module is used for arranging the signal intensity values in the communication data into signal intensity input vectors and then obtaining signal intensity characteristic vectors through a sequence coder comprising a one-dimensional convolutional layer;

the arrival time coding module is used for arranging arrival time in each communication data into an arrival time input vector and then obtaining an arrival time characteristic vector through the sequence coder containing the one-dimensional convolutional layer;

the positioning data acquisition module is used for acquiring three-dimensional coordinates of each positioning base station in the plurality of positioning base stations in the internal space of the offshore equipment;

the positioning data coding module is used for converting three-dimensional coordinates of each positioning base station in the internal space of the maritime equipment into position characteristic values through a full connection layer, and then arranging the position characteristic values of each positioning base station into position characteristic vectors;

a feature distribution correction module, configured to perform feature value distribution location correction on the signal intensity feature vector and the arrival time feature vector respectively based on the location feature vector to obtain a corrected signal intensity feature vector and a corrected arrival time feature vector;

the first correlation coding module is used for performing vector multiplication on the transposed vector of the position characteristic vector and the corrected signal intensity characteristic vector to obtain a position-intensity correlation matrix and then obtaining a first positioning characteristic diagram through a first convolution neural network serving as a characteristic extractor;

the second correlation coding module is used for performing vector multiplication on the transposed vector of the position characteristic vector and the corrected arrival time characteristic vector to obtain a position-time correlation matrix and then obtaining a second positioning characteristic diagram through a second convolutional neural network serving as a characteristic extractor;

a positioning feature fusion module, configured to fuse the first positioning feature map and the second positioning feature map to obtain a decoding feature map; and

and the positioning data generation module is used for performing regression decoding on the decoding characteristic diagram through a decoder to obtain a decoding value, and the decoding value is the positioning data of the positioning beacon.

2. The personnel trajectory positioning system for an interior space of offshore equipment, as claimed in claim 1, wherein said signal strength encoding module is further configured to:

arranging the signal strength values in each of the communication data into the signal strength input vector;

performing full-concatenation encoding on the signal strength input vector using a full-concatenation layer of the sequence encoder to extract high-dimensional implicit features of feature values of respective positions in the signal strength input vector according to the following formula:

wherein X ₁ Is the signal strength input vector, Y ₁ Is the output vector, W is the weight matrix, B is the offset vector,

represents a matrix multiplication;

performing one-dimensional convolution encoding on the signal intensity input vector by using a one-dimensional convolution layer of the sequence encoder according to the following formula so as to extract high-dimensional implicit correlation characteristics among characteristic values of all positions in the signal intensity input vector, wherein the formula is as follows:

wherein a is a convolution kernel at x ₁ Width in direction, F is the convolution kernel parameter vector, G is the local vector matrix operating with the convolution kernel function, w is the size of the convolution kernel, X ₁ Representing the signal strength input vector.

3. The personnel trajectory positioning system for the internal space of offshore equipment, as recited in claim 2, wherein said time of arrival encoding module is further configured to:

arranging arrival time in each communication data into the arrival time input vector;

using a full-concatenation layer of the sequence encoder to perform full-concatenation encoding on the arrival time input vector by using a formula to extract high-dimensional implicit features of feature values of each position in the arrival time input vector, wherein the formula is as follows:

wherein, X ₂ Is the arrival time input vector, Y ₂ Is the output vector, W is the weight matrix, B is the offset vector,

represents a matrix multiplication;

performing one-dimensional convolutional encoding on the arrival time input vector by using a one-dimensional convolutional layer of the sequence encoder according to the following formula so as to extract high-dimensional implicit correlation characteristics among characteristic values of all positions in the arrival time input vector, wherein the formula is as follows:

wherein a is a convolution kernel at x ₂ Width in direction, F is the convolution kernel parameter vector, G is the local vector matrix operating with the convolution kernel function, w is the size of the convolution kernel, X ₂ Representing the arrival time input vector.

4. The personnel trajectory positioning system for an internal space of offshore equipment according to claim 3, characterized in that said feature distribution correction module comprises:

a first correction unit, configured to perform hierarchical depth homography alignment correction on the signal strength eigenvector based on the position eigenvector to obtain the corrected signal strength eigenvector, where the hierarchical depth homography alignment correction is performed based on a difference eigenvector between the position eigenvector and the signal strength eigenvector and a Frobenius norm of a full scene homography correlation matrix between the position eigenvector and the signal strength eigenvector, where the full scene homography correlation matrix between the position eigenvector and the signal strength eigenvector is a product between a transposed vector of the position eigenvector and the signal strength eigenvector;

a second correction unit, configured to perform hierarchical depth homography correction on the arrival time feature vector based on the position feature vector to obtain the corrected arrival time feature vector, where the hierarchical depth homography correction is performed based on a difference feature vector between the position feature vector and the arrival time feature vector and a Frobenius norm of a full scene homography correlation matrix between the position feature vector and the arrival time feature vector, where the full scene homography correlation matrix between the position feature vector and the arrival time feature vector is a product between a transposed vector of the position feature vector and the arrival time feature vector.

5. Personnel trajectory positioning system for an internal space of offshore equipment, according to claim 4, characterized in that,

the first correction unit is further configured to: carrying out hierarchical depth homography alignment correction on the signal intensity characteristic vector based on the position characteristic vector by using the following formula to obtain the corrected signal intensity characteristic vector;

wherein the formula is:

wherein V ₁ A feature vector, V, representing the signal strength ₃ Representing said position feature vector, V ₁ ' represents the corrected signal strengthSign the vector, | · | | non-conducting phosphor ₁ Represents a norm of the vector, and | · |. Non-woven counting _F Represents the Frobenius norm of the matrix,

indicating a difference in click position, an indication of a click-to-click position,

represents summing by position;

the second correction unit is further configured to: carrying out hierarchical depth homography alignment correction on the arrival time feature vector based on the position feature vector by using the following formula so as to obtain the corrected arrival time feature vector;

wherein the formula is:

wherein V ₂ Feature vector, V, representing the time of arrival ₃ Representing said position feature vector, V ₂ ' representing the corrected arrival time feature vector, | · | | | non-calculation ₁ Represents a norm of the vector, and | | · | cals |, calvert |, calcualty _F Representing the Frobenius norm of the matrix,

indicating a sum by location.

6. The personnel trajectory positioning system for maritime equipment interior space of claim 5, wherein said first correlation encoding module comprises:

a first correlation matrix constructing unit, configured to perform vector multiplication on the transposed vector of the position feature vector and the corrected signal strength feature vector by using the following formula to obtain the position-strength correlation matrix;

wherein the formula is:

wherein V ₃ For the position feature vector, V ₁ ' is the corrected signal strength feature vector, M ₁ Representing the location-intensity correlation moment,

represents a vector multiplication;

a first convolution encoding unit, configured to perform convolution processing, pooling processing, and activation processing on input data in forward pass of layers using layers of the first convolution neural network as a feature extractor to generate the first location feature map from a last layer of the first convolution neural network, wherein an input of the first layer of the first convolution neural network is the position-intensity correlation matrix.

7. The personnel trajectory positioning system for the internal space of offshore equipment, of claim 6, wherein said second associative coding module comprises:

a first correlation matrix constructing unit, configured to perform vector multiplication on the transposed vector of the position feature vector and the corrected arrival time feature vector by using the following formula to obtain the position-time correlation matrix;

wherein the formula is:

wherein V ₃ For the position feature vector, V ₂ ' is the corrected arrival time feature vector, M ₂ Representing the position-time correlation matrix and,

represents a vector multiplication;

a first convolution encoding unit, configured to perform convolution processing, pooling processing, and activation processing on input data in forward pass of layers using layers of the second convolutional neural network as a feature extractor to generate the second localization feature map from a last layer of the second convolutional neural network, wherein an input of the first layer of the second convolutional neural network is the position-time correlation matrix.

8. The personnel trajectory positioning system for an interior space of offshore equipment according to claim 7, wherein said positioning feature fusion module comprises:

fusing the first positioning feature map and the second positioning feature map according to the following formula to obtain the decoding feature map;

wherein the formula is:

F _s ＝αF ₁ +βF ₂

wherein, F _s For said decoded feature map, F ₁ Is the first positioning feature map, F ₂ For the second localization signature, "+" indicates the addition of elements at the corresponding positions of the first and second localization signatures, α and β are weighting parameters for controlling the balance between the first and second localization signatures in the decoding signature.

9. A method for testing a personnel trajectory positioning system for an interior space of offshore equipment, comprising:

acquiring a plurality of positioning data of a plurality of positions of a positioning beacon in the internal space of the maritime equipment based on the personnel trajectory positioning system for the internal space of the maritime equipment as claimed in any one of claims 1 to 8;

generating a test result based on a comparison between real location data of a plurality of locations of the positioning beacon within the maritime equipment interior space and the plurality of positioning data.

10. The method of claim 9, wherein generating a test result based on a comparison between the actual location data and the plurality of positioning data for the plurality of locations of the positioning beacon within the maritime equipment interior space comprises:

constructing real position data of a plurality of positions of the positioning beacon within the maritime equipment interior space as first trajectory data;

constructing the plurality of positioning data as second trajectory data;

calculating a Euclidean distance between the first track data and the second track data; and

generating the test result based on a comparison between the Euclidean distance and a predetermined threshold.