CN117591984B - Bridge pushing path monitoring method and system based on deep learning - Google Patents
Bridge pushing path monitoring method and system based on deep learning Download PDFInfo
- Publication number
- CN117591984B CN117591984B CN202410065927.7A CN202410065927A CN117591984B CN 117591984 B CN117591984 B CN 117591984B CN 202410065927 A CN202410065927 A CN 202410065927A CN 117591984 B CN117591984 B CN 117591984B
- Authority
- CN
- China
- Prior art keywords
- bridge
- pose
- pushing
- attribute
- value
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 75
- 238000012544 monitoring process Methods 0.000 title claims abstract description 41
- 238000013135 deep learning Methods 0.000 title claims abstract description 26
- 230000008569 process Effects 0.000 claims abstract description 39
- 230000002159 abnormal effect Effects 0.000 claims abstract description 23
- 230000004044 response Effects 0.000 claims abstract description 9
- 238000005259 measurement Methods 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 10
- 238000011478 gradient descent method Methods 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 abstract description 15
- 238000010586 diagram Methods 0.000 description 6
- 238000009499 grossing Methods 0.000 description 5
- 230000001105 regulatory effect Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 239000003550 marker Substances 0.000 description 2
- 230000010355 oscillation Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 210000002435 tendon Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The application relates to the technical field of bridge construction, in particular to a method and a system for monitoring a bridge pushing path based on deep learning, wherein the method comprises the following steps: in the pushing process of the bridge, a pushing force sequence and a pushing speed sequence at the current moment are collected; inputting the initial pose, the pushing force sequence and the pushing speed sequence of the bridge into a trained pose prediction network to output a predicted pose of the bridge; inputting bridge predicted poses into a trained attribute prediction network to obtain predicted attribute values of each set mark point; comparing the predicted attribute value with the theoretical attribute value to obtain a safety discrimination result; and in response to the safety judgment result being unsafe, adjusting the pushing force and the pushing speed, and monitoring the pushing path of the bridge. Through the technical scheme of the application, when the pushing path is abnormal, the control parameters in the pushing process can be timely adjusted, so that the accurate monitoring of the pushing path is realized.
Description
Technical Field
The application relates to the technical field of bridge construction, in particular to a method and a system for monitoring a bridge pushing path based on deep learning.
Background
The bridge is one of important infrastructure in urban traffic, and in the bridge construction process, pushing method construction is a common and important construction method, wherein the pushing method construction refers to a bridge construction method for pushing the precast beam forward in place section by section along the bridge axis direction at the bridge head. Specifically, the pushing construction is to set a construction site behind the bridge abutment, cast the beam body in sections, integrate the cast sections with the completed Liang Tilian by using longitudinal prestressed tendons, push the beam body forward out of the construction site by using a pushing device, and repeat the procedures to complete the construction of all the beam body.
In the bridge pushing process, the thrust data can reflect the problems of potential structural deformation, displacement or bearing capacity reduction and the like of the bridge, further reflect the monitoring of the pushing path in the bridge pushing process, and can take necessary maintenance or reinforcement measures early through the monitoring of the thrust data so as to ensure the smooth completion of the bridge pushing construction.
At present, a patent application document with publication number of CN117113263A discloses a real-time monitoring method for a bridge pushing structure, which comprises the following steps: collecting thrust time sequence data in a bridge pushing structure; obtaining an oscillation discrete degree value and an elastic rebound resilience value of the thrust data corresponding to each time point according to the thrust time sequence data; obtaining a classification evaluation index of thrust data corresponding to each time point according to the oscillation discrete degree value and the elastic rebound resilience value, and dividing data in the thrust time sequence data to obtain abnormal data points and non-abnormal data points; carrying out smoothing factor adjustment on abnormal data points to obtain smoothing factors after corresponding adjustment of abnormal data points, and carrying out smoothing factor adjustment on non-abnormal data points to obtain smoothing factors after corresponding adjustment of non-abnormal data points; and according to the regulated smoothing factors and the conventional index weighted moving average formulas corresponding to the abnormal data points and the non-abnormal data points, obtaining a regulated index weighted moving average formula, regulating the thrust time sequence data according to the regulated index weighted moving average formula, and obtaining a group of accurate thrust time sequence data.
However, although the method can acquire accurate thrust time sequence data to realize the monitoring of the pushing path, the influence of control parameters except the thrust time sequence data on the pushing path in the pushing process is ignored, and the control parameters in the pushing process cannot be timely adjusted when the pushing path is abnormal, so that the accurate monitoring of the pushing path cannot be realized.
Disclosure of Invention
In order to solve the technical problems, the application provides a bridge pushing path monitoring method and system based on deep learning, so that when the pushing path is abnormal, control parameters in the pushing process are timely adjusted, and accurate monitoring of the pushing path is realized.
In a first aspect of the present application, a method for monitoring a bridge pushing path based on deep learning is provided, where the monitoring method includes: in the bridge pushing process, a pushing force sequence and a pushing speed sequence at the current moment are collected, wherein the pushing force sequence comprises pushing force at each moment from the beginning moment to the current moment of the bridge pushing, and the pushing speed sequence comprises pushing speed at each moment from the beginning moment to the current moment of the bridge pushing; acquiring an initial bridge pose, inputting the initial bridge pose, the pushing force sequence and the pushing speed sequence into a trained pose prediction network to output a bridge prediction pose, wherein the bridge prediction pose comprises three-dimensional coordinates of a plurality of set marking points on a bridge at the next adjacent moment, and the initial bridge pose comprises three-dimensional coordinates of a plurality of set marking points on the bridge at the initial moment of the bridge pushing process; inputting the bridge predicted pose into a trained attribute prediction network to obtain a predicted attribute value of each set mark point, wherein the predicted attribute value of one set mark point comprises a predicted stress value and a predicted strain value; comparing the predicted attribute value with a theoretical attribute value to obtain a safety discrimination result, wherein the safety discrimination result comprises safety and unsafe, and the theoretical attribute value comprises a maximum allowable stress value and a maximum allowable strain value of the bridge; responding to the safety judging result as unsafe, adjusting the pushing force and the pushing speed to realize monitoring of a bridge pushing path; the pose prediction network comprises a first time sequence module, a second time sequence module and a regression module; the attribute prediction network is a fully-connected neural network.
In one embodiment, in the pose prediction network, the first timing module is configured to perform timing feature extraction on the thrust sequence to obtain a thrust timing feature; the second time sequence module is used for extracting time sequence characteristics of the pushing speed sequence and acquiring time sequence characteristics of the pushing speed; and inputting the jacking force time sequence characteristic, the jacking speed time sequence characteristic and the initial pose of the bridge into the regression module after splicing so as to output the predicted pose of the bridge.
In one embodiment, the training method of the pose prediction network includes: in a historical bridge pushing process, collecting the initial pose of the bridge in the historical bridge pushing process, and a pushing force sequence and a pushing speed sequence at any moment, as a set of pose samples, and collecting the three-dimensional coordinates of all set marking points on the bridge at the next adjacent moment at any moment to obtain pose labels of the pose samples; inputting the pose sample into the pose prediction network to obtain an output result; calculating a mean square error loss function value between the output result and the pose label as pose loss; updating the pose prediction network by using a gradient descent method so as to reduce the pose loss; and continuously collecting pose samples, and iteratively updating the pose prediction network until the pose loss is smaller than a set loss value or the iteration number is larger than the set number, so as to obtain the trained pose prediction network.
In one embodiment, the attribute prediction network includes an input layer, a hidden layer, and an output layer; the input layer includes 3×n neurons for receiving the bridge prediction poses; the hidden layer is used for carrying out dimension transformation on the bridge prediction pose; the output layer comprises 2×n neurons, and is used for mapping the output result of the hidden layer into a predicted attribute value of each set mark point, where N is the number of all set mark points on the bridge.
In one embodiment, the training method of the attribute prediction network includes: in the history bridge pushing process, collecting three-dimensional coordinates of each set mark point on a bridge at any moment to serve as a group of attribute samples, and collecting actual measurement attribute values of each set mark point at any moment to serve as attribute labels of the attribute samples through a preset sensor, wherein the actual measurement attribute values comprise actual measurement stress values and actual measurement strain values; inputting the attribute sample into the attribute prediction network to obtain a prediction result corresponding to the attribute sample; calculating a mean square error loss function value between the prediction result and the attribute tag as attribute loss; updating the attribute prediction network by using a gradient descent method to reduce the attribute loss; and continuously collecting attribute samples, and iteratively updating the attribute prediction network until the attribute loss is smaller than a set loss value or the iteration number is larger than the set number, so as to obtain the trained attribute prediction network.
In one embodiment, comparing the predicted attribute value with the theoretical attribute value to obtain the security discrimination result includes: for any one set mark point, in response to the predicted stress value of the set mark point being greater than the maximum allowable stress value or the predicted strain value of the set mark point being greater than the maximum allowable strain value, marking the set mark point as an outlier; responding to the number of the abnormal points being equal to 0, wherein the safety discrimination result is safety; and responding to the number of the abnormal points to be at least one, wherein the safety judgment result is unsafe.
In one embodiment, adjusting the jacking force and the jacking speed comprises: initializing a bridge expected pose, wherein the bridge expected pose is a three-dimensional coordinate of a plurality of set marking points on a bridge at the next adjacent moment after the jacking force and the pushing speed are adjusted; inputting the bridge expected pose into a trained attribute prediction network to output a predicted attribute value of each set mark point under the bridge expected pose; calculating expected pose loss based on the theoretical attribute value and the predicted attribute value of each set marking point under the expected pose of the bridge, wherein the expected pose loss is used for reflecting the rationality of the expected pose of the bridge; in response to the expected pose loss being greater than the maximum allowable loss, back-propagating according to the expected pose loss to update the bridge expected pose, until the expected pose loss corresponding to the updated bridge expected pose is not greater than the maximum allowable loss, taking the updated bridge expected pose as a target bridge pose; and adjusting the pushing force and the pushing speed according to the target bridge pose and the real-time bridge pose at the current moment so that the three-dimensional coordinates of each set marking point on the bridge at the next adjacent moment meet the target bridge pose.
In one embodiment, the desired pose loss satisfies the relationship:
wherein,is->Predicted stress values of each set mark point, +.>Is->The predicted strain value of each set mark point,and->Maximum allowable stress value and maximum allowable strain value, respectively, < >>For all the number of set mark points, +.>For the bridge the desired pose +.>For the real-time pose of the bridge at the current moment +.>In order to preset the adjustment coefficient, the adjustment coefficient is set,representation->Corresponding to the set mark pointsMaximum value of>For the desired pose loss, +.>Is thatActivating function->The activation function satisfies the relation:
。
in one embodiment, the bridge real-time pose at the current moment comprises three-dimensional coordinates of a plurality of set marking points on the bridge at the current moment, and the bridge real-time pose at the current moment can be obtained through a position sensor deployed at each set marking point.
The second aspect of the application also provides a bridge pushing path monitoring system based on deep learning, which comprises a processor and a memory, wherein the memory stores computer program instructions, and the bridge pushing path monitoring method based on deep learning according to the first aspect of the application is realized when the computer program instructions are executed by the processor.
The technical scheme of the application has the following beneficial technical effects:
according to the technical scheme provided by the application, in the bridge pushing process, the initial pose, the pushing force sequence and the pushing speed sequence of the bridge are input into the trained pose prediction network to output the predicted pose of the bridge, and the predicted pose of the bridge can reflect the pose of the bridge at the next adjacent moment; further, the bridge predicted pose is input into a trained attribute prediction network to obtain a predicted attribute value of each set mark point under the bridge predicted pose, whether the bridge under the bridge predicted pose is safe or not is judged by comparing the theoretical attribute value with the predicted attribute value of each set mark point, if not, the abnormal pushing path at the next moment is indicated, the control parameters in the pushing process, including the pushing force and the pushing speed, are timely adjusted, smooth completion of the pushing of the bridge is ensured, and monitoring of the pushing path of the bridge is realized.
Further, in the process of adjusting control parameters in the pushing process, firstly initializing the expected pose of the bridge, and accurately outputting the predicted attribute value of each set marking point under the expected pose of the bridge by means of the trained attribute prediction network; the predicted attribute value of each set mark point is further used for participating in calculation of expected pose loss, the expected pose loss is used for reflecting rationality of expected pose of the bridge, the smaller the value of the expected pose loss is, the more the bridge under the expected pose of the bridge meets the safety requirement, and the smaller the difficulty of adjusting the bridge from the real-time pose of the bridge to the expected pose of the bridge at the current moment is, the more the rationality of the expected pose of the bridge is; and carrying out counter propagation on the expected pose loss to update the expected pose of the bridge, taking the updated expected pose of the bridge as a target bridge pose when the expected pose loss of the bridge corresponding to the updated expected pose of the bridge is not more than the maximum allowable loss, providing reference information for adjusting the jacking force and the pushing speed by the target bridge pose, ensuring the safety of the bridge at the next adjacent moment on the premise of ensuring the minimum adjustment quantity, and realizing the accurate control of control parameters.
Drawings
The above, as well as additional purposes, features, and advantages of exemplary embodiments of the present application will become readily apparent from the following detailed description when read in conjunction with the accompanying drawings. Several embodiments of the present application are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which like reference numerals refer to similar or corresponding parts and in which:
FIG. 1 is a flow chart of a method for monitoring a bridge pushing path based on deep learning according to an embodiment of the application;
FIG. 2 is a schematic diagram of a pose prediction network according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a structure of an attribute prediction network according to an embodiment of the present application;
fig. 4 is a block diagram of a deep learning based bridge pushing path monitoring system according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all, of the embodiments of the present application. All other embodiments, which can be made by those skilled in the art based on the embodiments herein without making any inventive effort, are intended to be within the scope of the present application.
It should be understood that when the terms "first," "second," and the like are used in the claims, specification, and drawings of this application, they are used merely for distinguishing between different objects and not for describing a particular sequential order. The terms "comprises" and "comprising," when used in the specification and claims of this application, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
According to a first aspect of the application, the application provides a bridge pushing path monitoring method based on deep learning, which is used for monitoring a pushing path in the process of bridge pushing construction by using a pushing device and timely adjusting control parameters when the pushing path is abnormal. In the pushing construction process, after the pushing device pushes at least one section of bridge which is poured in advance to a preset position on the bridge pier, the section of bridge and the completed Liang Tilian on the bridge pier are integrated, and the bridge is continuously pushed to the corresponding position on the bridge pier, so that the construction of all the bridges is completed.
Fig. 1 is a flowchart of a method for monitoring a bridge pushing path based on deep learning according to an embodiment of the application. As shown in fig. 1, the method 100 for monitoring a bridge pushing path based on deep learning includes steps S101 to S105, which are described in detail below.
S101, in the bridge pushing process, a pushing force sequence and a pushing speed sequence at the current moment are collected, wherein the pushing force sequence comprises pushing force at each moment between the bridge pushing starting moment and the current moment, and the pushing speed sequence comprises pushing speed at each moment between the bridge pushing starting moment and the current moment.
In one embodiment, in the pushing process of the bridge, the control parameters of the pushing device comprise pushing force and pushing speed, and the bridge moves to a preset position on the bridge pier under the action of the pushing device. That is, the pushing force and the pushing speed of the pushing device directly influence the pushing path in the bridge pushing process.
Collecting the jacking force of each moment between the initial moment and the current moment of the bridge jacking to obtain a jacking force sequence of the current moment; and collecting the pushing speed of each moment from the initial moment to the current moment of the bridge pushing, and obtaining a pushing speed sequence. Each moment can be every second or every minute, an implementer can set according to actual conditions, and the pushing speed is the pushing distance of the bridge between adjacent moments.
It can be appreciated that the jacking force sequence and the jacking speed sequence can reflect the control parameters of the jacking device at each moment between the initial moment and the current moment of the bridge jacking, and further can reflect the jacking path between the initial moment and the current moment of the bridge jacking.
S102, acquiring an initial bridge pose, inputting the initial bridge pose, the pushing force sequence and the pushing speed sequence into a trained pose prediction network to output a bridge prediction pose, wherein the bridge prediction pose comprises three-dimensional coordinates of a plurality of set mark points on a bridge at the next adjacent moment, and the initial bridge pose comprises three-dimensional coordinates of a plurality of set mark points on the bridge at the initial moment of the bridge pushing process.
In one embodiment, a plurality of set mark points are marked on the bridge in advance, the pose of the bridge in the space can be accurately positioned according to the three-dimensional coordinates of all the set mark points, and the more the number of the set mark points is, the more the pose of the bridge in the space can be accurately represented. As can be appreciated, the three-dimensional coordinates of a plurality of set marking points on the bridge at the starting moment of the bridge pushing process are the starting positions of the bridge pushing paths, and the pose of the bridge in space is continuously changed under the action of control parameters (pushing force and pushing speed); and acquiring three-dimensional coordinates of all set marking points at each moment from the initial moment to the current moment of the bridge pushing, and determining a bridge pushing path. And taking the three-dimensional coordinates of a plurality of set marking points on the bridge at the initial moment of the bridge pushing process as the initial pose of the bridge.
In one embodiment, the initial pose of the bridge, the top thrust sequence and the top pushing speed sequence at the current moment are input into a trained pose prediction network, and the bridge prediction pose is output, wherein the bridge prediction pose comprises three-dimensional coordinates of a plurality of set marking points on the bridge at the next adjacent moment, namely the bridge prediction pose is used for reflecting the pose of the bridge at the next adjacent moment in space.
In one embodiment, please refer to fig. 2, which is a schematic structural diagram of a pose prediction network according to an embodiment of the present application. The pose prediction network comprises a first time sequence module, a second time sequence module and a regression module; the first timing module is used for extracting timing sequence characteristics of the thrust sequence to obtain thrust timing sequence characteristics; the second time sequence module is used for extracting time sequence characteristics of the pushing speed sequence and acquiring time sequence characteristics of the pushing speed; and inputting the jacking force time sequence characteristic, the jacking speed time sequence characteristic and the initial pose of the bridge into the regression module after splicing so as to output the predicted pose of the bridge.
The first time sequence module and the second time sequence module can adopt LSTM or GRU and other circulating neural networks, and the regression module adopts fully-connected neural networks. The bridge prediction pose is a vector of 3×N rows and 1 columns, N is the number of set marking points on the bridge, and one set marking point corresponds to one three-dimensional coordinate.
In one embodiment, the pose prediction network needs to be trained in order for the pose prediction network to be able to output accurate bridge predicted poses. Specifically, the training method of the pose prediction network comprises the following steps: in a historical bridge pushing process, collecting the initial pose of the bridge in the historical bridge pushing process, and a pushing force sequence and a pushing speed sequence at any moment, as a set of pose samples, and collecting the three-dimensional coordinates of all set marking points on the bridge at the next adjacent moment at any moment to obtain pose labels of the pose samples; inputting the pose sample into the pose prediction network to obtain an output result; calculating a mean square error loss function value between the output result and the pose label as pose loss; updating the pose prediction network by using a gradient descent method so as to reduce the pose loss; and continuously collecting pose samples, and iteratively updating the pose prediction network until the pose loss is smaller than a set loss value or the iteration number is larger than the set number, so as to obtain the trained pose prediction network.
Wherein the value of the set loss value is 0.001; the value of the set times is 10000.
Thus, the pose of the bridge in the space at the next adjacent moment is accurately predicted according to the trained pose prediction network, and the pose of the bridge in the space at the next adjacent moment comprises three-dimensional coordinates of all set marking points.
S103, inputting the bridge predicted pose into a trained attribute prediction network to obtain a predicted attribute value of each set mark point, wherein the predicted attribute value of one set mark point comprises a predicted stress value and a predicted strain value.
In one embodiment, the bridge prediction pose comprises three-dimensional coordinates of all set marking points on the bridge when the bridge prediction pose comprises the next adjacent moment at the current moment; based on the three-dimensional coordinates of all the set mark points, the pose information of the bridge can be reflected, and then the attribute value of each mark point under the pose information can be predicted, wherein the attribute value comprises a stress value and a strain value.
And inputting the bridge predicted pose into a trained attribute prediction network, and outputting a predicted attribute value of each set mark point under the bridge predicted pose, wherein the predicted attribute value comprises a predicted stress value and a predicted strain value.
In one embodiment, please refer to fig. 3, which is a schematic diagram of a structure of an attribute prediction network according to an embodiment of the present application. The attribute prediction network is a fully-connected neural network and comprises an input layer, a hidden layer and an output layer; the input layer includes 3×n neurons for receiving the bridge prediction poses; the hidden layer is used for carrying out dimension transformation on the bridge prediction pose; the output layer comprises 2 XN neurons, the output result of the hidden layer is mapped into a predicted attribute value of each set mark point, the predicted attribute value of one set mark point comprises a predicted stress value and a predicted strain value, and N is the number of all set mark points on the bridge.
In one embodiment, in order for the attribute prediction network to accurately predict the predicted attribute value for each set marker point, the attribute prediction network needs to be trained such that the attribute prediction network learns the mapping relationship between the bridge predicted pose and the predicted attribute value for each set marker point.
Specifically, the training method of the attribute prediction network comprises the following steps: in the history bridge pushing process, collecting three-dimensional coordinates of each set mark point on a bridge at any moment to serve as a group of attribute samples, and collecting actual measurement attribute values of each set mark point at any moment to serve as attribute labels of the attribute samples through a preset sensor, wherein the actual measurement attribute values comprise actual measurement stress values and actual measurement strain values; inputting the attribute sample into the attribute prediction network to obtain a prediction result corresponding to the attribute sample; calculating a mean square error loss function value between the prediction result and the attribute tag as attribute loss; updating the attribute prediction network by using a gradient descent method to reduce the attribute loss; and continuously collecting attribute samples, and iteratively updating the attribute prediction network until the attribute loss is smaller than a set loss value or the iteration number is larger than the set number, so as to obtain the trained attribute prediction network.
Wherein the value of the set loss value is 0.001; the value of the set times is 10000. The preset sensor comprises a stress sensor and a strain sensor, wherein the stress sensor is used for collecting actual measurement stress values, and the strain sensor is used for collecting actual measurement strain values.
Therefore, the trained attribute prediction network can learn the mapping relation between the bridge prediction pose and the predicted attribute value of each set mark point, and further accurately output the predicted attribute value of each set mark point under the bridge prediction pose.
S104, comparing the predicted attribute value with a theoretical attribute value to obtain a safety discrimination result, wherein the safety discrimination result comprises safety and unsafe, and the theoretical attribute value comprises a maximum allowable stress value and a maximum allowable strain value of the bridge.
In one embodiment, the theoretical property value is related to the material strength and structure of the bridge, and the theoretical property value can be set by those skilled in the art according to the specific situation of the bridge.
Specifically, comparing the predicted attribute value with the theoretical attribute value to obtain the security discrimination result includes: for any one set mark point, in response to the predicted stress value of the set mark point being greater than the maximum allowable stress value or the predicted strain value of the set mark point being greater than the maximum allowable strain value, marking the set mark point as an outlier; responding to the number of the abnormal points being equal to 0, wherein the safety discrimination result is safety; and responding to the number of the abnormal points to be at least one, wherein the safety judgment result is unsafe.
When the safety judging result is safe, the position and the posture of the bridge at the next adjacent moment can meet the safety requirement, and the pushing path at the next adjacent moment is not abnormal; when the safety judgment result is unsafe, the position and the posture of the bridge at the next adjacent moment cannot meet the safety requirement, the predicted stress value or the maximum strain value is larger than the corresponding maximum allowable value, and the pushing road at the next adjacent moment is abnormal.
And S105, adjusting the jacking force and the jacking speed to monitor the jacking path of the bridge in response to the fact that the safety judging result is unsafe.
In one embodiment, in response to the safety determination result being unsafe, the abnormal condition of the pushing path at the next adjacent moment is indicated, and the control parameters are required to be adjusted to ensure the smooth completion of the bridge pushing process, so that the monitoring of the bridge pushing path is realized.
Wherein, in the process of adjusting the pushing force and the pushing speed, an alarm can be sent out, and the adjustment of the pushing force and the pushing speed is carried out by a person skilled in the art.
In another embodiment, adjusting the jacking force and the jacking speed includes: initializing a bridge expected pose, wherein the bridge expected pose is a three-dimensional coordinate of a plurality of set marking points on a bridge at the next adjacent moment after the jacking force and the pushing speed are adjusted; inputting the bridge expected pose into a trained attribute prediction network to output a predicted attribute value of each set mark point under the bridge expected pose; calculating expected pose loss based on the theoretical attribute value and the predicted attribute value of each set marking point under the expected pose of the bridge, wherein the expected pose loss is used for reflecting the rationality of the expected pose of the bridge; in response to the expected pose loss being greater than the maximum allowable loss, back-propagating according to the expected pose loss to update the bridge expected pose, until the expected pose loss corresponding to the updated bridge expected pose is not greater than the maximum allowable loss, taking the updated bridge expected pose as a target bridge pose; and adjusting the pushing force and the pushing speed according to the target bridge pose and the real-time bridge pose at the current moment so that the three-dimensional coordinates of each set marking point on the bridge at the next adjacent moment meet the target bridge pose.
The process of initializing the expected pose of the bridge can be regarded as a process of randomly generating a vector of 3×n rows and 1 columns. The maximum allowable loss is 0.001. The bridge real-time pose at the current moment is the three-dimensional coordinates of a plurality of set marking points on the bridge at the current moment, and the bridge real-time pose at the current moment can be obtained through a position sensor deployed at each set marking point.
The predicted attribute value of each set mark point under the expected pose of the bridge can be accurately output by means of the trained attribute prediction network; further, whether the bridge is safe under the expected position and posture of the bridge can be judged by comparing the theoretical attribute value with the predicted attribute value of each set mark point, and calculation of expected position and posture loss is participated, that is, the closer the expected position and posture loss is to 0, the predicted attribute value of each set mark point does not exceed the theoretical attribute value, and the greater the rationality of the expected position and posture of the bridge is.
Specifically, the expected pose loss satisfies the relation:
wherein,is->Predicted stress values of each set mark point, +.>Is->Predicted strain values for each set mark point,And->Maximum allowable stress value and maximum allowable strain value, respectively, < >>For all the number of set mark points, +.>For the bridge the desired pose +.>For the real-time pose of the bridge at the current moment +.>In order to preset the adjustment coefficient, the adjustment coefficient is set,representation->Corresponding to the set mark pointsMaximum value of>For the desired pose loss, +.>Is thatActivating function->The activation function satisfies the relation:
。
wherein, the value of the preset adjusting coefficient is 0.3.Indicates the corresponding +.>The maximum value of the set mark point is used for representing the magnitude relation between the theoretical attribute value and the predicted attribute value of each set mark point, and when the value of the maximum value is equal to 0, the predicted attribute value of each set mark point does not exceed the theoretical attribute value, namely, the bridge in the expected pose of the bridge meets the safety requirement;the method is used for representing the difference between the bridge expected pose and the bridge real-time pose at the current moment, the larger the numerical value is, the larger the difficulty of adjusting the bridge from the bridge real-time pose to the bridge expected pose at the current moment is, the larger the damage to the pushing device is, and conversely, the smaller the numerical value is, the smaller the difficulty of adjusting the bridge from the bridge real-time pose to the bridge expected pose at the current moment is. Therefore, the smaller the value of the expected pose loss is, the more the bridge under the expected pose of the bridge meets the safety requirement, and the smaller the difficulty of adjusting the bridge from the real-time pose of the bridge to the expected pose of the bridge at the current moment is, the more the rationality of the expected pose of the bridge is.
After determining the pose of the target bridge, the pose of the target bridge comprises target three-dimensional coordinates of each set mark point, and the pose of the target bridge provides reference information for adjusting the jacking force and the pushing speed; the jacking force and the pushing speed can be adjusted by a person skilled in the art according to the difference between the target bridge pose and the real-time pose of the bridge at the current moment, so that the three-dimensional coordinates of each set mark point on the bridge at the next adjacent moment meet the target bridge pose, that is, the three-dimensional coordinates of each set mark point are equal to the corresponding target three-dimensional coordinates; on the premise of ensuring the minimum adjustment quantity, the bridge safety at the next adjacent moment is ensured, and the accurate control of the control parameters is realized.
Technical principles and implementation details of the bridge pushing path monitoring method based on deep learning are introduced through the specific embodiments. According to the technical scheme provided by the application, in the bridge pushing process, the initial pose, the pushing force sequence and the pushing speed sequence of the bridge are input into the trained pose prediction network to output the predicted pose of the bridge, and the predicted pose of the bridge can reflect the pose of the bridge at the next adjacent moment; further, the bridge predicted pose is input into a trained attribute prediction network to obtain a predicted attribute value of each set mark point under the bridge predicted pose, whether the bridge under the bridge predicted pose is safe or not is judged by comparing the theoretical attribute value with the predicted attribute value of each set mark point, if not, the abnormal pushing path at the next moment is indicated, the control parameters in the pushing process, including the pushing force and the pushing speed, are timely adjusted, smooth completion of the pushing of the bridge is ensured, and monitoring of the pushing path of the bridge is realized.
According to a second aspect of the application, the application further provides a bridge pushing path monitoring system based on deep learning. Fig. 4 is a block diagram of a deep learning based bridge pushing path monitoring system according to an embodiment of the present application. As shown in fig. 4, the system 50 includes a processor and a memory storing computer program instructions that when executed by the processor implement a method for deep learning based bridge pushing path monitoring according to the first aspect of the present application. The system further comprises other components known to those skilled in the art, such as a communication bus and a communication interface, the arrangement and function of which are known in the art and are therefore not described in detail herein.
In the context of this application, the foregoing memory may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, or device. For example, the computer readable storage medium may be any suitable magnetic or magneto-optical storage medium, such as, for example, resistance change Memory RRAM (Resistive Random Access Memory), dynamic Random Access Memory DRAM (Dynamic Random Access Memory), static Random Access Memory SRAM (Static Random-Access Memory), enhanced dynamic Random Access Memory EDRAM (Enhanced Dynamic Random Access Memory), high-Bandwidth Memory HBM (High-Bandwidth Memory), hybrid storage cube HMC (Hybrid Memory Cube), etc., or any other medium that may be used to store the desired information and that may be accessed by an application, a module, or both. Any such computer storage media may be part of, or accessible by, or connectable to, the device. Any of the applications or modules described herein may be implemented using computer-readable/executable instructions that may be stored or otherwise maintained by such computer-readable media.
The technical features of the above-described embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above-described embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the claims. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.
Claims (9)
1. The method for monitoring the pushing path of the bridge based on deep learning is characterized by comprising the following steps of:
in the bridge pushing process, a pushing force sequence and a pushing speed sequence at the current moment are collected, wherein the pushing force sequence comprises pushing force at each moment from the beginning moment to the current moment of the bridge pushing, and the pushing speed sequence comprises pushing speed at each moment from the beginning moment to the current moment of the bridge pushing;
acquiring an initial bridge pose, inputting the initial bridge pose, the pushing force sequence and the pushing speed sequence into a trained pose prediction network to output a bridge prediction pose, wherein the bridge prediction pose comprises three-dimensional coordinates of a plurality of set marking points on a bridge at the next adjacent moment, and the initial bridge pose comprises three-dimensional coordinates of a plurality of set marking points on the bridge at the initial moment of the bridge pushing process;
inputting the bridge predicted pose into a trained attribute prediction network to obtain a predicted attribute value of each set mark point, wherein the predicted attribute value of one set mark point comprises a predicted stress value and a predicted strain value;
comparing the predicted attribute value with a theoretical attribute value to obtain a safety discrimination result, wherein the safety discrimination result comprises safety and unsafe, and the theoretical attribute value comprises a maximum allowable stress value and a maximum allowable strain value of the bridge;
responding to the safety judging result as unsafe, adjusting the pushing force and the pushing speed to realize monitoring of a bridge pushing path;
the pose prediction network comprises a first time sequence module, a second time sequence module and a regression module; the attribute prediction network is a fully-connected neural network;
in the pose prediction network, the first timing module is used for extracting timing sequence characteristics of the thrust sequence to obtain thrust timing sequence characteristics; the second time sequence module is used for extracting time sequence characteristics of the pushing speed sequence and acquiring time sequence characteristics of the pushing speed; and inputting the jacking force time sequence characteristic, the jacking speed time sequence characteristic and the initial pose of the bridge into the regression module after splicing so as to output the predicted pose of the bridge.
2. The method for monitoring the pushing path of the bridge based on deep learning according to claim 1, wherein the training method of the pose prediction network comprises the following steps:
in a historical bridge pushing process, collecting the initial pose of the bridge in the historical bridge pushing process, and a pushing force sequence and a pushing speed sequence at any moment, as a set of pose samples, and collecting the three-dimensional coordinates of all set marking points on the bridge at the next adjacent moment at any moment to obtain pose labels of the pose samples;
inputting the pose sample into the pose prediction network to obtain an output result;
calculating a mean square error loss function value between the output result and the pose label as pose loss;
updating the pose prediction network by using a gradient descent method so as to reduce the pose loss;
and continuously collecting pose samples, and iteratively updating the pose prediction network until the pose loss is smaller than a set loss value or the iteration number is larger than the set number, so as to obtain the trained pose prediction network.
3. The method for monitoring a pushing path of a bridge based on deep learning according to claim 1, wherein the attribute prediction network comprises an input layer, a hidden layer and an output layer;
the input layer includes 3×n neurons for receiving the bridge prediction poses;
the hidden layer is used for carrying out dimension transformation on the bridge prediction pose;
the output layer comprises 2×n neurons, and is used for mapping the output result of the hidden layer into a predicted attribute value of each set mark point, where N is the number of all set mark points on the bridge.
4. The method for monitoring a pushing path of a bridge based on deep learning according to claim 3, wherein the training method of the attribute prediction network comprises the following steps:
in the history bridge pushing process, collecting three-dimensional coordinates of each set mark point on a bridge at any moment to serve as a group of attribute samples, and collecting actual measurement attribute values of each set mark point at any moment to serve as attribute labels of the attribute samples through a preset sensor, wherein the actual measurement attribute values comprise actual measurement stress values and actual measurement strain values;
inputting the attribute sample into the attribute prediction network to obtain a prediction result corresponding to the attribute sample;
calculating a mean square error loss function value between the prediction result and the attribute tag as attribute loss;
updating the attribute prediction network by using a gradient descent method to reduce the attribute loss;
and continuously collecting attribute samples, and iteratively updating the attribute prediction network until the attribute loss is smaller than a set loss value or the iteration number is larger than the set number, so as to obtain the trained attribute prediction network.
5. The method for monitoring a pushing path of a bridge based on deep learning according to claim 1, wherein comparing the predicted attribute value with the theoretical attribute value to obtain a safety discrimination result comprises:
for any one set mark point, in response to the predicted stress value of the set mark point being greater than the maximum allowable stress value or the predicted strain value of the set mark point being greater than the maximum allowable strain value, marking the set mark point as an outlier;
responding to the number of the abnormal points being equal to 0, wherein the safety discrimination result is safety;
and responding to the number of the abnormal points to be at least one, wherein the safety judgment result is unsafe.
6. The method of claim 1, wherein adjusting the jacking force and the jacking speed comprises:
initializing a bridge expected pose, wherein the bridge expected pose is a three-dimensional coordinate of a plurality of set marking points on a bridge at the next adjacent moment after the jacking force and the pushing speed are adjusted;
inputting the bridge expected pose into a trained attribute prediction network to output a predicted attribute value of each set mark point under the bridge expected pose;
calculating expected pose loss based on the theoretical attribute value and the predicted attribute value of each set marking point under the expected pose of the bridge, wherein the expected pose loss is used for reflecting the rationality of the expected pose of the bridge;
in response to the expected pose loss being greater than the maximum allowable loss, back-propagating according to the expected pose loss to update the bridge expected pose, until the expected pose loss corresponding to the updated bridge expected pose is not greater than the maximum allowable loss, taking the updated bridge expected pose as a target bridge pose;
and adjusting the pushing force and the pushing speed according to the target bridge pose and the real-time bridge pose at the current moment so that the three-dimensional coordinates of each set marking point on the bridge at the next adjacent moment meet the target bridge pose.
7. The method for monitoring a pushing path of a bridge based on deep learning of claim 6, wherein the expected pose loss satisfies the relationship:
wherein,is->Predicted stress values of each set mark point, +.>Is->Predicted strain values for each set mark point, +.>And->Maximum allowable stress value and maximum allowable strain value, respectively, < >>For all the number of set mark points, +.>For the bridge the desired pose +.>For the real-time pose of the bridge at the current moment +.>In order to preset the adjustment coefficient, the adjustment coefficient is set,representation->Corresponding to the set mark pointsMaximum value of>For the desired pose loss, +.>Is thatActivating function->The activation function satisfies the relation:
。
8. the method for monitoring the pushing path of the bridge based on deep learning according to claim 6, wherein the real-time pose of the bridge at the current moment comprises three-dimensional coordinates of a plurality of set marking points on the bridge at the current moment, and the real-time pose of the bridge at the current moment can be obtained through a position sensor deployed at each set marking point.
9. A deep learning based bridge pushing path monitoring system, comprising a processor and a memory, the memory storing computer program instructions which when executed by the processor implement a deep learning based bridge pushing path monitoring method according to any one of claims 1 to 8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410065927.7A CN117591984B (en) | 2024-01-17 | 2024-01-17 | Bridge pushing path monitoring method and system based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410065927.7A CN117591984B (en) | 2024-01-17 | 2024-01-17 | Bridge pushing path monitoring method and system based on deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117591984A CN117591984A (en) | 2024-02-23 |
CN117591984B true CN117591984B (en) | 2024-04-09 |
Family
ID=89920445
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410065927.7A Active CN117591984B (en) | 2024-01-17 | 2024-01-17 | Bridge pushing path monitoring method and system based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117591984B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117870776B (en) * | 2024-03-11 | 2024-06-11 | 四川公路桥梁建设集团有限公司 | Bridge pushing real-time detection method and detection device |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20130040409A (en) * | 2011-10-14 | 2013-04-24 | 코벡주식회사 | System and method for launching upper structure of bridege |
CN111749201A (en) * | 2020-06-22 | 2020-10-09 | 上海中交水运设计研究有限公司 | Normal-submerged pontoon anchoring stress air bag preposed arresting system for preventing ship collision of bridge body |
CN114790702A (en) * | 2022-05-16 | 2022-07-26 | 中国水利水电第七工程局有限公司 | Steel box girder integral synchronous pushing installation method based on cooperation of BIM and FEM |
CN115292780A (en) * | 2022-07-14 | 2022-11-04 | 哈尔滨工业大学 | Large-span bridge buffeting response prediction method based on deep learning |
CN115408931A (en) * | 2022-08-16 | 2022-11-29 | 哈尔滨工业大学 | Vortex vibration response prediction method based on deep learning |
KR20230021272A (en) * | 2021-08-05 | 2023-02-14 | 주식회사 뉴로다임 | Bridge health monitoring system using artificial intelligence |
CN116479727A (en) * | 2023-06-21 | 2023-07-25 | 中交二公局工程检测技术有限公司 | Road surface construction quality monitoring device |
CN116927764A (en) * | 2023-08-03 | 2023-10-24 | 长沙理工大学 | Measurement nipple for sensing lithology and geological structure of surrounding rock stratum while drilling |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11687805B2 (en) * | 2019-10-31 | 2023-06-27 | Shoreline Iot, Inc. | Systems and methods for self-learning artificial intelligence of things (AIOT) devices and services |
-
2024
- 2024-01-17 CN CN202410065927.7A patent/CN117591984B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20130040409A (en) * | 2011-10-14 | 2013-04-24 | 코벡주식회사 | System and method for launching upper structure of bridege |
CN111749201A (en) * | 2020-06-22 | 2020-10-09 | 上海中交水运设计研究有限公司 | Normal-submerged pontoon anchoring stress air bag preposed arresting system for preventing ship collision of bridge body |
KR20230021272A (en) * | 2021-08-05 | 2023-02-14 | 주식회사 뉴로다임 | Bridge health monitoring system using artificial intelligence |
CN114790702A (en) * | 2022-05-16 | 2022-07-26 | 中国水利水电第七工程局有限公司 | Steel box girder integral synchronous pushing installation method based on cooperation of BIM and FEM |
CN115292780A (en) * | 2022-07-14 | 2022-11-04 | 哈尔滨工业大学 | Large-span bridge buffeting response prediction method based on deep learning |
CN115408931A (en) * | 2022-08-16 | 2022-11-29 | 哈尔滨工业大学 | Vortex vibration response prediction method based on deep learning |
CN116479727A (en) * | 2023-06-21 | 2023-07-25 | 中交二公局工程检测技术有限公司 | Road surface construction quality monitoring device |
CN116927764A (en) * | 2023-08-03 | 2023-10-24 | 长沙理工大学 | Measurement nipple for sensing lithology and geological structure of surrounding rock stratum while drilling |
Non-Patent Citations (3)
Title |
---|
BP神经网络在桥梁施工控制中的应用;习会峰等;《南昌大学学报(理科版)》;20071031;1-4页 * |
Dynamic prediction for attitude and position in shield tunneling: A deep learning method;周承洲等;《Automation in Construction》;20190930;全文 * |
大跨度连续梁拱组合桥施工控制关键技术研究;金森;《中国优秀硕士学位论文电子期刊》;20230215;1-119 * |
Also Published As
Publication number | Publication date |
---|---|
CN117591984A (en) | 2024-02-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN117591984B (en) | Bridge pushing path monitoring method and system based on deep learning | |
CN106981198B (en) | Deep learning network model for travel time prediction and establishing method thereof | |
Zhang et al. | Improving Dam Seepage Prediction Using Back‐Propagation Neural Network and Genetic Algorithm | |
CN113836999A (en) | Tunnel construction risk intelligent identification method and system based on ground penetrating radar | |
CN110162895A (en) | A kind of two stage high energy efficiency ship form optimization design method | |
Liu et al. | A deep reinforcement learning approach for ramp metering based on traffic video data | |
CN112597694A (en) | Neural network-based surrounding rock deformation prediction system and prediction method | |
Zhang et al. | Time-frequency fusion features-based incremental network for smartphone measured structural seismic response classification | |
CN112883478B (en) | Steel structure displacement prediction method, device, terminal equipment and system | |
CN114997036A (en) | Network topology reconstruction method, device and equipment based on deep learning | |
Zhang et al. | AI-based modeling and data-driven identification of moving load on continuous beams | |
CN116758767B (en) | Traffic signal lamp control method based on multi-strategy reinforcement learning | |
Liu et al. | 3 System Reliability | |
CN115238394B (en) | Multi-source uncertainty hybrid reliability digital twin modeling method for composite material structure | |
CN116910534A (en) | Space-time intelligent prediction method and device for ocean environmental elements in different sea areas | |
JP4853997B2 (en) | Agent learning apparatus, agent learning method, and agent learning program | |
Gao et al. | An improved assessment method for urban growth simulations across models, regions, and time | |
CN105808845A (en) | Structural damage identification method based on glowworm group intelligent algorithm | |
CN117571358B (en) | Method and system for detecting performance of pushing device for bridge | |
CN114153216A (en) | Lunar surface path planning system and method based on deep reinforcement learning and block planning | |
KR20220111215A (en) | Apparatus and method for predicting drug-target interaction using deep neural network model based on self-attention | |
Li | Evaluation of financial performance of port enterprises based on radial basis function neural network | |
CN112435464A (en) | Q reinforcement learning-based takeover time prediction system and prediction method for vehicles running on main line of highway | |
CN117216846B (en) | Reinforced concrete member hysteresis curve prediction method, system, equipment and medium | |
Lawniczak et al. | Decisions and success of heterogeneous population of agents in learning to cross a highway |
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 |