CN112734111A - AGV dynamic time estimation method for horizontal transportation task - Google Patents
AGV dynamic time estimation method for horizontal transportation task Download PDFInfo
- Publication number
- CN112734111A CN112734111A CN202110027744.2A CN202110027744A CN112734111A CN 112734111 A CN112734111 A CN 112734111A CN 202110027744 A CN202110027744 A CN 202110027744A CN 112734111 A CN112734111 A CN 112734111A
- Authority
- CN
- China
- Prior art keywords
- data
- agv
- time estimation
- dynamic time
- estimation method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000010801 machine learning Methods 0.000 claims abstract description 32
- 238000013528 artificial neural network Methods 0.000 claims abstract description 27
- 238000012545 processing Methods 0.000 claims abstract description 18
- 238000003062 neural network model Methods 0.000 claims abstract description 16
- 230000008569 process Effects 0.000 claims abstract description 6
- 230000003068 static effect Effects 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000012216 screening Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 5
- 238000004422 calculation algorithm Methods 0.000 claims description 5
- 238000003066 decision tree Methods 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000007637 random forest analysis Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 3
- 238000012417 linear regression Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000001133 acceleration Effects 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Abstract
The invention discloses an AGV dynamic time estimation method for a horizontal transportation task in the technical field of dynamic time estimation, which comprises the following steps: step 1: selecting characteristics; step 2: processing data; and step 3: machine learning; and 4, step 4: a neural network; and 5: verifying the dynamic time estimation model; according to the horizontal transport task AGV dynamic time estimation method, a machine learning model and a neural network model are used for dynamically predicting the task completion time of the AGV, and the influence of various factors on the AGV in the horizontal transport process of a wharf is described. Therefore, the time estimation result is improved, and more accurate basic data is provided for other application scenes.
Description
Technical Field
The invention relates to the technical field of dynamic time estimation, in particular to an AGV dynamic time estimation method for a horizontal transportation task.
Background
Agvs (automated Guided vehicles), also known as automated Guided vehicles, laser Guided vehicles, are distinguished by being unmanned. AGV unmanned vehicle loading and unloading operation is widely used in the current port industrial field. In the process of AGV operation, the AGV time estimation application scenes are wide, the AGV operation time is the key basic data of a scheduling system, and the AGV time estimation method can be applied to AGV task selection and arrangement, actual task AGV operation time estimation, port overall efficiency optimization, selection of traffic conflict priorities during the operation of the AGV, automatic TOS task distribution of a port, matching decision of vehicles and tasks, conflict optimization of vehicle path planning and the like, and is wide in application scenes and important in research significance.
The static time matrix estimation is widely used in the current wharf time estimation scene, the static time estimation is carried out by simply considering an acceleration and deceleration principle through a starting point and a terminal point of an AGV, and the time required by the starting point to the terminal point is directly calculated by a strategy without any collision and avoidance. The static time matrix estimation step is as follows: 1. splitting a route, namely splitting the planned route into a plurality of different sections; 2. time calculation, namely calculating the time consumed by the AGV passing each section respectively by utilizing the physical scene of the acceleration and deceleration principle as the starting speed and the ending speed of the AGV of each section of path are known, and adding the time to obtain the final static time
In the prior art, a static time matrix is used for automatic wharf port operation to estimate AGV traffic time, other vehicles on a wharf are not considered, a physical rule is directly used, and speed planning calculation is performed. The deviation between the prediction result and the actual result exceeds 35%, and the too high time prediction deviation causes the phenomenon of resource idling or low efficiency in each link of port operation, and is not improved for a long time.
Disclosure of Invention
The invention aims to provide an AGV dynamic time estimation method for a horizontal transportation task, which uses a machine learning model and a neural network model to dynamically predict the completion task time of the AGV and solves the technical problem that in the prior art, the deviation between a static time matrix prediction result and an actual result is great, so that the phenomenon of resource idling or low efficiency is generated in each link of port operation.
In order to achieve the technical purpose, the invention adopts the following technical scheme:
according to one aspect of the invention, a method for estimating the AGV dynamic time of a horizontal transportation task is provided, which comprises the following steps:
step 1: selecting characteristics: selecting and screening AGV operation instruction characteristics of the horizontal transportation task by using a computer algorithm, and integrating characteristic data;
step 2: data processing: screening and processing the characteristic data in the step 1 by using a computer algorithm to form standardized data;
and step 3: machine learning: adopting various machine learning models to carry out reference prediction on the standardized data in the step 2, and searching for optimal parameters;
and 4, step 4: a neural network: building a full connection layer neural network, and predicting the standardized data in the step 2 by using a neural network model;
and 5: verifying the dynamic time estimation model: comparing the machine learning model predicted value, the neural network model predicted value and the static time matrix, and evaluating the performance of the dynamic time estimation model;
according to the AGV dynamic time estimation method for the horizontal transportation task in the aspect of the invention, the characteristic selection in the step 1 comprises the following steps:
s11, extracting data from the order table; selecting a record related to the AGV operation instruction from an AGV _ ORDERS table;
s12, extracting data from the track table; finding one or more indexes, screening key points in the driving process of the vehicle according to one Order, and dividing the Track corresponding to each Order data according to the key points;
s13, integrating data: and counting the number of AGV path conflicts to generate a total characteristic.
According to the AGV dynamic time estimation method of the horizontal transportation task of the present invention, the step S12 includes the following steps:
s121: according to the Movefrequency field of the data table after the characteristic extraction, counting and calculating the times of straight movement, turning and oblique movement of the AGV trolley;
s122: calculating the path length of the track according to the data in the step S121;
s123: processing the start and end times of the data in the above step S121;
s124: and calculating the running speed of the trolley according to the starting time, the ending time and the distance of the track.
According to the AGV dynamic time estimation method of the horizontal transportation task in the above aspect of the present invention, the general characteristics in step S13 include vehicle driving distance, number of straight-going runs, number of turns, number of inclined runs, number of colliding vehicles, ship unloading, ship loading, pile turning, box handling, box fetching, a PB segment to PB segment, a PB segment to WSTP segment, a WSTP segment to PB segment, and a WSTP segment to WSTP segment.
According to the AGV dynamic time estimation method for the horizontal transportation task, the step 2 comprises the following steps:
s21: analyzing the characteristic data, and removing abnormal values from the obvious large amount of error data and invalid data;
s22: further screening abnormal values, and clustering two direct factors of time consumption and total distance;
s23: after the clustering is finished, selecting data meeting the conditions as final training data;
s24: the data was normalized, normalized and one-hot encoded.
According to the AGV dynamic time estimation method for the horizontal transportation task in the aspect of the invention, in the step 3, various machine learning models comprise XGB, GBDT, RandomForest, DecisionTree, Ridge and Linear regression.
According to the AGV dynamic time estimation method for the horizontal transportation task in the aspect of the invention, a grid search method is adopted in the method for searching the optimal parameters in various models by machine learning in the step 3.
According to the AGV dynamic time estimation method for the horizontal transportation task in the above aspect of the invention, the first layer of the fully-connected layer neural network in the step 4 is an input, the last layer is an output, and the hidden layer has a 128-by-128 structure.
According to the AGV dynamic time estimation method for the horizontal transportation task in the aspect of the invention, the value range of each layer of output of the fully-connected layer neural network is compressed to (0, 1) by using the activation function, and the output value of each level is used as the input value of the next level.
According to the AGV dynamic time estimation method for the horizontal transportation task in the aspect of the invention, in the step 5, the evaluation standard for verifying the prediction of the machine learning model on the time is the root mean square error, the root mean square error between the predicted value, the static time matrix value and the average value of the computer learning model and the actual value is respectively calculated, and the performance of the machine learning model is evaluated through comparison of calculation results.
By adopting the technical scheme, the invention has the following advantages:
the invention provides an AGV dynamic time estimation method for a horizontal transportation task, which uses a machine learning model and a neural network model to dynamically predict the completion task time of the AGV, combines the consideration of the path, characteristics, traffic condition in a field, a congestion area and the like, combines time information and field information according to a model trained by historical information, greatly optimizes a model result of dynamic time estimation, and can ensure that the deviation of the prediction result and an actual result can reach within 20 percent so as to solve the technical problem that factors such as AGV traffic conflict, time and the like are difficult to directly depict, improve the phenomenon of resource idling or low efficiency of each link of port operation, improve the result of time prediction and provide more accurate basic data for other application scenes.
Drawings
Other features, objects and advantages of the present invention will become more apparent upon reading of the following detailed description of non-limiting embodiments thereof with reference to the accompanying drawings.
FIG. 1 is a flow chart of a method for estimating AGV dynamic time for a horizontal transportation task according to the present invention;
FIG. 2 is a data screenshot of the invention extracting an order table;
FIG. 3 is a data screenshot of the invention extracting the track table;
FIG. 4 is a data table screenshot after feature extraction of the present invention;
FIG. 5 is a screen shot of a graph of the time profile of a completed task before data processing according to the present invention;
FIG. 6 is a screen shot of a graph of the time distribution of completed tasks after data processing according to the present invention;
FIG. 7 is a screen shot of a machine learning prediction results graph of various models of the present invention;
FIG. 8 is a screen shot of a segmented RMS error table of machine learning predictions for various models of the present invention;
FIG. 9 is a screen shot of a segmented root mean square error table of neural network predictions in accordance with the present invention;
FIG. 10 is a screen shot of a comparison of the static time matrix and the dynamic time estimate prediction results of the present invention.
Detailed Description
The technical scheme of the invention is specifically explained in the following with the accompanying drawings of the specification.
The detailed features and advantages of the present invention are described in detail in the detailed description which follows, and will be sufficient for anyone skilled in the art to understand the technical content of the present invention and to implement the present invention, and the related objects and advantages of the present invention will be easily understood by those skilled in the art from the description, claims and drawings disclosed in the present specification.
Examples
Fig. 1 shows a flowchart of the horizontal transportation task AGV dynamic time estimation method of the present invention, and in combination with consideration of the path, characteristics, traffic conditions in the yard, etc., the horizontal transportation task AGV dynamic time estimation method specifically includes the following steps as shown in fig. 1:
step 1: selecting characteristics: selecting and screening AGV operation instruction characteristics of the horizontal transportation task by using a computer program algorithm, and integrating characteristic data;
FIG. 2 shows a data screenshot of an extraction process order (order) table; FIG. 3 shows a data screenshot of the extract processing track table; FIG. 4 shows a data table screenshot after feature extraction;
wherein: the feature selection in step 1 comprises the following steps:
and S11, processing data in the order table, as shown in the figure 2, selecting a record related to an AGV operation instruction from an AGV _ ORDERS table (a sequence table of AGV trolleys), only considering key points PB (sea side operation area) and WSTP (sea side switching area) in the vehicle driving process at present, and screening the data according to whether a starting position (H _ AGO _ ST _ LOCATION) and a target position (H _ AGO _ PLANNED _ DESTINATION) are PB, QC (bridge crane) and WSTP (container yard sea side switching area).
And S12, processing data in a Track table (a motion Track table of the AGV trolley), finding one or more indexes, measuring the characteristics of a route from a current departure point to a destination point of the AGV, screening key points in the driving process of the vehicle according to a sequence Order, such as QC to PB, PB to QC, PB to PB, PB to WSTP, WSTP to PB, and WSTP to WSTP, screening corresponding Order data according to AGV _ ID (information of the AGV trolley), and dividing the Track corresponding to each Order data according to key points to obtain a data table after characteristic extraction shown in FIG. 4.
The step S12 includes the steps of:
s121: counting and calculating the times of straight running, turning and inclined running of the AGV trolley according to the Movefrequency field of the data;
s122: calculating the path length of the track according to the data;
s123: processing the start and end times of the data;
s124: and calculating the running speed of the trolley according to the starting time, the ending time and the distance of the track.
S13: the data in fig. 4 are integrated, the number of path conflicts is counted, and total features are generated, and the description of the total feature parameters and the data sources are shown in table 1 below:
table 1: total characteristic parameter description and data source
Name of field | Field description | Data source mode |
Distance | Distance traveled by vehicle | According to path calculation |
DirectFrequency | Number of straight lines | Reading from the path |
TurnFrequency | Number of turns | Reading from the path |
SlideFrequency | Number of diagonal strokes | Reading from the path |
ConflictVehicle | Number of conflicting vehicles | Computing with path keypoints |
AroundVehicle | Number of vehicles in yard at start of mission | The system gives |
DSCH | Ship unloading | Reading from order |
LOAD | Shipment of ships | Reading from order |
SHIFT | Rotating pile | Reading from order |
DELIVER | Cross box | Reading from order |
RECEIVE | Box taking device | Reading from order |
REPOSITION | Redirection | Reading from order |
PB2PB | From PB segment to PB segment | According to setting fill-in |
PB2WSTP | From PB segment to WSTP segment | According to setting fill-in |
WSTP2PB | From WSTP segment to PB segment | According to setting fill-in |
WSTP2WSTP | From WSTP segment to WSTP segment | According to setting fill-in |
earlymorning | 0 point to 4 points | According to setting fill-in |
morning | 4 to 8 points | According to setting fill-in |
Noon | 8 o 'clock to 12 o' clock | According to setting fill-in |
Afternoon | 12 o 'clock to 16 o' clock | According to setting fill-in |
Night | 16 o 'clock to 20 o' clock | According to setting fill-in |
Latenight | 20 o 'clock to 24 o' clock | According to setting fill-in |
Distance_no_path | Distance without path | Calculated starting point and end point |
FIG. 5 shows a graph of the task completion time profile before data processing; fig. 6 shows a graph of the task completion time profile after data processing.
Step 2: data processing: screening and processing the characteristic data in the step 1 by using a computer algorithm to form standardized data;
the step 2 comprises the following steps:
s21: the tasks before data processing take time as shown in fig. 5, and it is obvious that there are a large amount of error data and invalid data. Performing boxplot analysis on Distance data, and eliminating abnormal values;
s22: since the distribution of the screened data is still more dispersed, the outliers need to be further screened. Two direct factors of time consumption and total distance are considered to be clustered;
s23: after clustering is completed, selecting classes meeting the online of the box line graph as final training data;
s24: the data were normalized, and encoded for one hot, resulting in the time taken for the task to complete as shown in fig. 7.
And step 3: machine learning: adopting various machine learning models to participate in the prediction of the standardized data in the step 2, and searching for optimal parameters;
FIG. 7 illustrates a machine learning various model prediction results graphical screen shot; FIG. 8 illustrates a machine learning various model prediction results segmented root mean square error table screenshot;
in step 3, the various machine learning models include XGB (a tree model), GBDT (gradient boosting decision tree), RandomForest, DecisionTree, Ridge, and linear regression.
The prediction is performed in step 2 by using the above prediction models, the prediction results are shown in fig. 7, and it can be seen from fig. 7 that the prediction results XGB > GBDT > randomfortest > DecisionTree > Ridge ═ linear regression
And finding the optimal parameters by using a grid search method. Setting the XGB n _ estimator (parameter tuning code) to 200, the XGB model predicts the data results as shown in fig. 9, calculated as:
using a neural network model, the coefficients were determined to 0.66585,
the results of the piecewise view model at the root mean square error of each segment, regardless of the special cases 70-90 and 210-230, which are not representative and have small data size, are as follows:
the root mean square error ratio of the neural network in 90-110s is about 16.72/100-16.7%
The root mean square error ratio of the neural network in 110-
The ratio of root mean square error of the neural network in 130-
The root mean square error ratio of the neural network in 150-170s is about 23.08/160-14.4%
The root mean square error ratio of the neural network in 170-190s is about 27.96/180-15.6%
The ratio of root mean square error of the neural network in 190-210s is about 34.11/200-17.1%
The prediction error is about 10% -20%.
And 4, step 4: a neural network: and (3) building a fully-connected layer neural network for the multiple prediction models in the step (3), wherein the first layer is input, the last layer is output, the hidden layer with a 128 x 128 structure is provided, the value range of the output of each layer is compressed to (0, 1) by using an activation function, the value output at each level is used as the input of the next level, and the condition that certain input is infinite and other inputs are invalid is avoided. And finally, realizing accurate expression of the model on the data by using a solution method of back propagation and gradient descent. The neural network computes a series of transformations that change the similarity of the samples. The activity of each layer of neurons is a non-linear function of the activity of the previous layer.
Predicting data using a neural network model; fig. 9 shows a screen shot of a segmented mean square error table of a neural network prediction result, a fully-connected layer neural network is built, an activation function softmax, training algebras 300, training numbers 200 in each generation, an evaluation index MSE, and optimization (optimizer) Adam (updating of an optimizer, dynamically adjusting a learning rate of each parameter by using first moment estimation and second moment estimation of gradient), and a result of neural network model prediction data is calculated as shown in fig. 9:
using a neural network model, the coefficients were determined to 0.814517,
the results of the piecewise view model at the root mean square error of each segment, regardless of the special cases 70-90 and 210-230, which are not representative and have small data size, are as follows:
the root mean square error ratio of the neural network in 90-110s is about 16.57/100-16.6%
The root mean square error ratio of the neural network in 110-
The ratio of root mean square error of the neural network in 130-150s is about 19.99/140-14.3%
The root mean square error ratio of the neural network in 150-170s is about 23.17/160-14.5%
The root mean square error ratio of the neural network in 170-190s is about 28.36/180-15.7%
The ratio of root mean square error of the neural network in 190-210s is about 34.69/200-17.3%
The prediction error is about 10% -20%.
Figure 10 shows a comparison of the static time matrix and dynamic time estimate predictions of the present invention,
and 5: verifying the dynamic time estimation model: comparing the machine learning model predicted value with the neural network model predicted value and the static time matrix, and evaluating the performance of the dynamic time estimation model;
in order to verify the reliability of the machine learning model, the predicted value of the neural network model of the predicted value of the machine learning model is compared with a static time matrix, the prediction of time belongs to a regression problem, the typical regression performance evaluation standard is Mean Square Error (MSE), the root mean square error of the predicted value of the machine learning model, the static time matrix value and the predicted value of the neural network model is respectively calculated, and the performance of the machine learning model is evaluated through comparison of calculation results. For example, as shown in fig. 10, in the prediction result pair, as can be seen from fig. 10, by calculating the root mean square error between the actual consumed time and the mean value and the root mean square error between the actual consumed time and the static time matrix, the comparison data shows that, from the overall angle analysis, the root mean square error of the static time matrix and the mean value is much larger than the predicted value of the machine learning model, the root mean square error between the predicted value and the actual value of the neural network model, and the time prediction effect is significantly improved by the machine learning model and the neural network model. The model trained according to the historical information combines the time information and the field information, the model result of dynamic time estimation is greatly optimized, the deviation between the prediction result and the actual result can reach within 20%, and the phenomenon that resources are idle or the efficiency is low in each link of port operation is improved.
Finally, it should be noted that while the present invention has been described with reference to the specific embodiments, it should be understood by those skilled in the art that the above embodiments are only for illustrating the present invention and are not to be construed as limiting the present invention, and various equivalent changes and substitutions may be made without departing from the spirit of the present invention, therefore, the changes and modifications to the above embodiments within the spirit of the present invention are within the scope of the appended claims.
Claims (10)
1. A horizontal transportation task AGV dynamic time estimation method is characterized by comprising the following steps:
step 1: selecting characteristics: selecting and screening AGV operation instruction characteristics of the horizontal transportation task by using a computer algorithm, and integrating characteristic data;
step 2: data processing: screening and processing the characteristic data in the step 1 by using a computer algorithm to form standardized data;
and step 3: machine learning: adopting various machine learning models to participate in prediction on the standardized data in the step 2, and searching for optimal parameters;
and 4, step 4: a neural network: building a full connection layer neural network, and predicting the standardized data in the step 2 by using a neural network model;
and 5: verifying the dynamic time estimation model: and comparing the machine learning model predicted value, the neural network model predicted value and the static time matrix, and evaluating the performance of the dynamic time estimation model.
2. The AGV dynamic time estimation method according to claim 1, wherein said step 1 of selecting characteristics comprises the steps of:
s11, extracting data from the order table; selecting a record related to the AGV operation instruction from an AGV _ ORDERS table;
s12, extracting data from the track table; finding one or more indexes, screening key points in the driving process of the vehicle according to one Order, and dividing the Track corresponding to each Order data according to the key points;
s13, integrating data: and counting the number of AGV path conflicts to generate a total characteristic.
3. The AGV dynamic time estimation method according to claim 2, wherein said step S12 comprises the steps of:
s121: according to the Movefrequency field of the data table after the characteristic extraction, counting and calculating the times of straight movement, turning and oblique movement of the AGV trolley;
s122: calculating the path length of the track according to the data in the step S121;
s123: processing the start and end times of the data in the above step S121;
s124: and calculating the running speed of the trolley according to the starting time, the ending time and the distance of the track.
4. The AGV dynamic time estimation method according to claim 2, wherein the total characteristics in step S13 include vehicle driving distance, number of straight-going, number of turning, number of diagonal-going, number of colliding vehicles, ship unloading, ship loading, pile turning, box handing, box taking, PB segment to PB segment, PB segment to WSTP segment, WSTP segment to PB segment, and WSTP segment to WSTP segment.
5. The AGV dynamic time estimation method according to claim 1, wherein said step 2 comprises the steps of:
s21: analyzing the characteristic data, and removing abnormal values from the obvious large amount of error data and invalid data;
s22: further screening abnormal values, and clustering two direct factors of time consumption and total distance;
s23: after the clustering is finished, selecting data meeting the conditions as final training data;
s24: the data was normalized, normalized and one-hot encoded.
6. The AGV dynamic time estimation method according to claim 1, wherein the various machine learning models in step 3 include XGB, GBDT, RandomForest, DecisionTree, Ridg and Linear Reggression.
7. The AGV dynamic time estimation method according to claim 6, wherein the method for the machine to learn the optimal parameters in the models in step 3 is a grid search method.
8. The AGV dynamic time estimation method according to claim 1, wherein the first layer of the fully-connected neural network in step 4 is an input, the last layer is an output, and there are hidden layers of 128-by-128 structure.
9. The AGV dynamic time estimation method according to claim 8, wherein the value range of each layer output of said fully-connected layer neural network is compressed to (0, 1) by an activation function, and the value of each level output is used as the input value of the next level.
10. The AGV dynamic time estimation method according to claim 1, wherein the evaluation criterion for verifying the prediction of the machine learning model on time in the step 5 is root mean square error, the root mean square error between the predicted value, the static time matrix value, the average value and the actual value of the machine learning model is calculated respectively, and the performance of the machine learning model is evaluated by comparing the calculation results.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110027744.2A CN112734111B (en) | 2021-01-11 | 2021-01-11 | Horizontal transport task AGV dynamic time prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110027744.2A CN112734111B (en) | 2021-01-11 | 2021-01-11 | Horizontal transport task AGV dynamic time prediction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112734111A true CN112734111A (en) | 2021-04-30 |
CN112734111B CN112734111B (en) | 2023-12-22 |
Family
ID=75589936
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110027744.2A Active CN112734111B (en) | 2021-01-11 | 2021-01-11 | Horizontal transport task AGV dynamic time prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112734111B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114137978A (en) * | 2021-11-29 | 2022-03-04 | 佛山市毕佳索智能科技有限公司 | Speed planning and control method for storage AGV |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106971545A (en) * | 2017-05-16 | 2017-07-21 | 青岛大学 | A kind of bus arrival time Forecasting Methodology |
KR101897407B1 (en) * | 2017-06-14 | 2018-10-18 | 국방과학연구소 | Method of Adaptive Dynamic Model-base]d Optimal Path Planning for Autonomous Navigation of Unmanned Ground Vehicle and Appratus thereof |
CN110570678A (en) * | 2019-10-23 | 2019-12-13 | 厦门大学 | Method and device for predicting total travel time of bus from starting point to end point |
CN112149888A (en) * | 2020-09-08 | 2020-12-29 | 北京踏歌智行科技有限公司 | Real-time travel time prediction method for unmanned mine card for strip mine |
-
2021
- 2021-01-11 CN CN202110027744.2A patent/CN112734111B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106971545A (en) * | 2017-05-16 | 2017-07-21 | 青岛大学 | A kind of bus arrival time Forecasting Methodology |
KR101897407B1 (en) * | 2017-06-14 | 2018-10-18 | 국방과학연구소 | Method of Adaptive Dynamic Model-base]d Optimal Path Planning for Autonomous Navigation of Unmanned Ground Vehicle and Appratus thereof |
CN110570678A (en) * | 2019-10-23 | 2019-12-13 | 厦门大学 | Method and device for predicting total travel time of bus from starting point to end point |
CN112149888A (en) * | 2020-09-08 | 2020-12-29 | 北京踏歌智行科技有限公司 | Real-time travel time prediction method for unmanned mine card for strip mine |
Non-Patent Citations (1)
Title |
---|
PEILAN HE ET AL: "Learning heterogeneous traffic patterns for travel timeprediction of bus journeys", INFORMATION SCIENCES, vol. 512, pages 1394 - 1406, XP085933495, DOI: 10.1016/j.ins.2019.10.073 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114137978A (en) * | 2021-11-29 | 2022-03-04 | 佛山市毕佳索智能科技有限公司 | Speed planning and control method for storage AGV |
Also Published As
Publication number | Publication date |
---|---|
CN112734111B (en) | 2023-12-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107179769B (en) | AGV quantity configuration method based on real-time scheduling simulation and queuing theory | |
CN111860989B (en) | LSTM neural network short-time traffic flow prediction method based on ant colony optimization | |
CN116739323B (en) | Intelligent evaluation method and system for emergency resource scheduling | |
CN114723125A (en) | Intercity vehicle order allocation method combining deep learning and multi-task optimization | |
CN114595607A (en) | Digital twin textile sliver barrel conveying method and system | |
CN116629424A (en) | Novel automatic double-yard bridge intelligent scheduling method for container yard operation mode | |
Gaidar et al. | Mathematical method for optimising the transport and logistics industry | |
Hani et al. | Simulation based optimization of a train maintenance facility | |
CN115981264A (en) | AGV scheduling and quantity combined optimization method considering conflicts | |
CN117541197A (en) | Intelligent building method and system based on BIM and AIOT data driving | |
CN114444809A (en) | Data-driven multi-target strip mine card path optimization method | |
CN112734111A (en) | AGV dynamic time estimation method for horizontal transportation task | |
CN117093009B (en) | Logistics AGV trolley navigation control method and system based on machine vision | |
Asghari et al. | Disruption management for the electric vehicle routing problem in a geographically flexible network | |
Jun et al. | Evolutionary neural network for learning of scalable heuristics for pickup and delivery problems with time windows | |
Xia et al. | A multi-AGV optimal scheduling algorithm based on particle swarm optimization | |
CN116901981A (en) | Online self-learning Markov vehicle speed prediction method | |
CN116882883A (en) | Power grid material distribution path planning method utilizing big data | |
CN113687651B (en) | Path planning method and device for dispatching vehicles on demand | |
Abbas et al. | A constrained fuzzy knowledge-based system for the management of container yard operations | |
Ji et al. | A research based on k-means clustering and artificial fish-swarm algorithm for the vehicle routing optimization | |
Gan et al. | Scheduling problems of automated guided vehicles in automated container terminals using a genetic algorithm | |
CN113222285A (en) | Strip mine charging pile site selection method based on adaptive disturbance goblet sea squirt algorithm | |
CN108985526B (en) | Transportation capacity prediction method and device, computer readable storage medium and terminal | |
Wang et al. | DSS approach for sustainable system design of shuttle-based storage and retrieval systems |
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 |