CN115944868A - Control method for ship-borne fire water monitor - Google Patents
Control method for ship-borne fire water monitor Download PDFInfo
- Publication number
- CN115944868A CN115944868A CN202310078364.0A CN202310078364A CN115944868A CN 115944868 A CN115944868 A CN 115944868A CN 202310078364 A CN202310078364 A CN 202310078364A CN 115944868 A CN115944868 A CN 115944868A
- Authority
- CN
- China
- Prior art keywords
- fire
- water monitor
- fire water
- attitude
- adjustment 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 241000876443 Varanus salvator Species 0.000 title claims abstract description 54
- 230000003044 adaptive effect Effects 0.000 claims abstract description 14
- 238000013528 artificial neural network Methods 0.000 claims abstract description 10
- 238000005096 rolling process Methods 0.000 claims abstract description 8
- 230000015654 memory Effects 0.000 claims abstract description 4
- 241000282461 Canis lupus Species 0.000 claims description 59
- 230000008569 process Effects 0.000 claims description 17
- 239000002245 particle Substances 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 8
- 238000012549 training Methods 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 5
- 238000013139 quantization Methods 0.000 claims description 5
- 230000002146 bilateral effect Effects 0.000 claims description 4
- 230000005484 gravity Effects 0.000 claims description 4
- 230000007787 long-term memory Effects 0.000 claims description 4
- 230000006403 short-term memory Effects 0.000 claims description 4
- 230000002779 inactivation Effects 0.000 claims description 3
- 238000013519 translation Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 abstract description 16
- 206010034719 Personality change Diseases 0.000 description 9
- 230000000007 visual effect Effects 0.000 description 9
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 8
- 230000035772 mutation Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 206010000369 Accident Diseases 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 230000002262 irrigation Effects 0.000 description 1
- 238000003973 irrigation Methods 0.000 description 1
- 238000011068 loading method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- 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
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture
- Y02A40/28—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture specially adapted for farming
Landscapes
- Feedback Control In General (AREA)
Abstract
The invention provides a control method of a ship-borne fire monitor, which comprises the following steps: identifying the position of a fire source point and a jet flow drop point of the fire water monitor according to the image; acquiring motion attitude prediction information of a ship carrying a fire water monitor through a long-short term memory neural network, wherein the motion attitude prediction information comprises rolling and pitching; the interference of the motion attitude prediction information on the attitude angle of the fire water monitor is quantified by using a geometric method; inputting the position deviation of a fire source point and a jet flow drop point into a first self-adaptive fuzzy controller to obtain a primary adjustment value of the attitude angle of the fire water monitor; inputting the interference of the motion attitude prediction information on the attitude angle of the fire water monitor into a second self-adaptive fuzzy controller to obtain a compensation adjustment value of the attitude angle of the fire water monitor; and accumulating the primary adjustment value and the compensation adjustment value to obtain a total adjustment value of the attitude angle. The adaptive fuzzy controller can adaptively adjust the parameters along with the change of working conditions, thereby adapting to different environments.
Description
Technical Field
The invention relates to the field of fire fighting, in particular to a control method of a ship-borne fire water monitor.
Background
With the rapid development of the shipping industry in China, ship fire accidents frequently occur. Particularly, oil and chemical goods, cause enormous casualties, environmental pollution and economic loss in case of fire, and have bad social impact. Therefore, it is very important to perform the marine fire rescue rapidly, accurately and safely. The current main fire water monitor control methods comprise manual field control, manual remote control and closed-loop adjustment based on visual identification. The fire fighter site control fire water monitor carries out the operation and has higher flexibility ratio, but the operation performance of different fire fighters under different scenes differs greatly, and causes the injury of fire fighters and even sacrifices easily. The manual remote control mode can ensure the safety of fire fighters, but the accuracy and the efficiency of fire fighting operation are always poor due to the problems of delay, fuzziness and the like of returned pictures. The closed-loop autonomous control method based on visual identification is to identify the fire source and the jet flow drop point position through computer visual identification, and the deviation of the fire source and the jet flow drop point position is used as the reference for adjusting the attitude angle of the fire gun. However, in the time of adjusting the attitude of the fire water monitor, the change of the motion attitude of the ship caused by wind and wave can cause the relative attitude angle of the fire water monitor to shift, thereby causing operation errors, and the interference exists continuously and changes along with the working condition (sea condition), so the performance of the fire fighting operation of the method is not ideal when the sea condition grade is slightly larger. In order to solve the problem, the invention provides a ship-based fire monitor control method considering interference on the basis of a visual identification fire monitor closed-loop control method.
Disclosure of Invention
The invention provides a ship-borne fire monitor control method which is based on a fire monitor closed-loop autonomous control technology of visual recognition, predicts the motion attitude of a ship through a neural network method, quantifies the interference of attitude change on the fire monitor attitude control, and simultaneously adopts a double-structure controller to carry out closed-loop control on the total attitude angle adjustment value of the fire monitor.
In order to achieve the above object, the present invention provides a method for controlling a ship-based fire monitor, comprising:
identifying the position of a fire source point and a jet flow drop point of the fire water monitor according to the image;
acquiring motion attitude prediction information of a ship carrying a fire water monitor through a long-term and short-term memory neural network, wherein the motion attitude prediction information comprises rolling and pitching; the interference of the motion attitude prediction information on the attitude angle of the fire water monitor is quantified by using a geometric method; the attitude angle of the fire water monitor comprises a pitch angle and a horizontal deflection angle;
inputting the position deviation of a fire source point and a jet flow drop point into a first self-adaptive fuzzy controller to obtain a primary adjustment value of the attitude angle of the fire water monitor; inputting the interference of the motion attitude prediction information on the attitude angle of the fire water monitor into a second self-adaptive fuzzy controller to obtain a compensation adjustment value of the attitude angle of the fire water monitor; and accumulating the primary adjustment value and the compensation adjustment value to obtain a total adjustment value of the attitude angle.
The invention is further improved in that: in the process of identifying the fire source point, a YOLO v5s network is adopted as a flame detection network to carry out bilateral filtering on the image when the input image is preprocessed.
The invention is further improved in that: in the process of training the flame detection network, collecting fire images, labeling the fire images to obtain a data set, and performing data enhancement on the labeled data set, wherein the process of data enhancement comprises translation and rotation on the images.
The invention is further improved in that: the process of identifying the jet drop point from the image includes:
extracting a jet flow track from an image by adopting an improved U-net network, wherein the U-net network is obtained by adding an attention module between a lower sampling layer and an upper sampling layer on the basis of the existing U-net network;
and carrying out binarization processing on the extracted jet flow track, and extracting a point farthest from the jet flow starting point as a jet flow falling point.
The invention is further improved in that: carrying out periodic training and construction on the long-term and short-term memory neural network for obtaining the motion attitude prediction information; in each training and building process, the super parameters of the long-term and short-term memory neural network are adjusted by adopting an enhanced wolf optimization algorithm; the super parameters comprise hidden layer number, learning rate and random inactivation rate.
The invention is further improved in that: for the first adaptive fuzzy controller and the second adaptive fuzzy controller,
the input variables and the output variables are divided into 7 fuzzy sets, namely NB, NM, NS, ZE, PS, PM and PB which respectively and correspondingly represent negative large, negative middle, negative small, zero, positive small, positive middle and positive large;
mapping the input variable and the output variable to a fuzzy control membership function of a corresponding fuzzy set by adopting an isosceles triangle;
the defuzzification method selects a gravity center method.
The invention is further improved in that: and improving the fuzzy controller through an online particle swarm algorithm, designing a fitness function to evaluate the advantages and disadvantages of a proportional factor and a quantization factor of the current fuzzy controller, and performing self-adaptive adjustment on the proportional factor and the quantization factor.
The method has the beneficial effects that:
1. the fuzzy control algorithm is improved through the online particle swarm algorithm, and the adaptive fuzzy controller can adaptively adjust parameters along with the change of working conditions, so that the better control performance is kept under different environments;
2. the control interference of the fire monitor caused by the change of the ship motion attitude is further considered, the double-adaptive fuzzy controller is designed to combine the interference value with the visual identification information for information fusion control by predicting the ship motion attitude and quantizing the interference degree of the ship motion attitude to the attitude angle control of the fire monitor, the interference is effectively resisted, and the operation precision and efficiency of offshore fire fighting are improved.
Drawings
FIG. 1 is a flow chart of a fire source location identification process;
FIG. 2 is a flow chart of a jet drop location identification process;
FIG. 3 is a diagram of an improved U-net network architecture;
FIG. 4 is a flow chart of a jet drop point identification process;
FIG. 5 is a flow chart of a dynamic vessel attitude prediction model based on EGWO-LSTM;
FIG. 6 is a schematic diagram of a crossover operator;
FIG. 7 is a schematic diagram of a mutation operator;
FIG. 8 is a flow chart of an online particle swarm algorithm to adaptively adjust fuzzy controller parameters;
FIG. 9 is a flow chart of a method of controlling a ship-based fire monitor;
FIG. 10 is a schematic of a fuzzy control membership function.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention, and the drawings only show the components related to the present invention rather than the number, shape and size of the components in actual implementation, and the type, quantity and proportion of the components in actual implementation may be changed freely, and the layout of the components may be more complicated.
Some exemplary embodiments of the invention have been described for illustrative purposes, and it is to be understood that the invention may be practiced otherwise than as specifically described.
The invention provides a ship-borne fire water monitor control method, which comprises the steps of firstly, predicting the motion attitude of a ship by a neural network method, and quantifying the interference of the attitude change on the attitude control of the fire water monitor; secondly, identifying and obtaining the positions of a fire source point and a jet flow drop point by using a deep learning method, and calculating to obtain the deviation between the fire source point and the jet flow drop point; and then, improving fuzzy control by using an online particle swarm algorithm and designing a double-adaptive fuzzy controller, wherein the two adaptive fuzzy controllers respectively obtain a primary adjustment value and a compensation value of two attitude angles of the fire water monitor, and finally, the information fusion control of the fire water monitor with strong anti-interference, strong adaptive capacity, high operation precision and high response speed is realized. The method comprises the following specific steps:
an improved enhanced wolf optimization algorithm (EGWO) is used for carrying out periodic training and construction on LSTM (long-short term memory neural network), so that high-precision ship motion attitude (rolling and pitching) prediction information is obtained by adapting to working condition changes and using a current optimal prediction model. Based on the prediction information, after determining the type value parameters of the fire fighting ship and the relative installation position of the fire fighting water monitor on the fire fighting ship, the interference of the rolling and pitching on the pitch angle and the horizontal deflection angle of the fire fighting water monitor is quantified by using a geometric method.
In the aspect of fire source point identification, a ship fire data set is self-made, the problem that images are not clear due to sea fog and flame combustion is solved by carrying out image enhancement processing on the obtained data set images, a YOLO v5 network is trained, and the images collected by a camera are identified to obtain fire source point coordinates.
And establishing a two-stage jet flow drop point identification method. In the second stage, the extracted jet flow track is subjected to binarization processing, under the condition that a jet flow starting point is known, pixel points of the jet flow track are traversed, and the point farthest from the jet flow starting point is a jet flow falling point.
And improving fuzzy control by using an online particle swarm algorithm, designing a double-structure controller, and fusing quantized ship attitude change interference and visual identification information to control the fire water monitor. The input of the self-adaptive fuzzy controller 1 is the deviation between the positions of a fire source point and a jet flow falling point, the output of the self-adaptive fuzzy controller is a primary adjustment value of two attitude angles of the fire water monitor, the input of the self-adaptive fuzzy controller 2 is predicted and quantized ship attitude change interference, the output of the self-adaptive fuzzy controller is a compensation adjustment value of the two attitude angles of the fire water monitor, and the sum of the primary adjustment value and the compensation value is used as a total adjustment value of the attitude angles of the fire water monitor at the next moment for closed-loop control.
In the specific implementation process:
1) Fire source point and jet flow drop point position identification
(1) Fire source point identification
And (3) making a data set aiming at the ship fire, analyzing the characteristics of the marine fire, preprocessing the images in the data set, and realizing the rapid and accurate identification of the fire source point. Fig. 1 is a flowchart of a fire source location identification process.
(1) Data set collection and annotation
The data set determines the quality of a target detection result to a great extent, multi-scene fire images such as ships and the like are collected from the Internet, a labeling tool Labelimg is adopted for labeling, data enhancement operations such as translation and rotation are carried out on the labeled data set to finally obtain 5005 images, and the images are divided into a training set and a verification set according to the proportion of 8:2.
(2) Image pre-processing
The method is applied to water area scenes such as the sea, the environment is complex, water vapor is more abundant compared with the ground, dense fog is more easily generated, a target to be detected is flame, the smoke is diffused when a fire disaster occurs, particles, dust and the like are generated, a camera is blurred, the characteristic that the visual field is not clear is generated, the definition of an image is low, the resolution ratio is reduced, the effect influences the identification accuracy, and therefore bilateral filtering operation is performed on the image in data concentration.
The bilateral filtering belongs to nonlinear filtering, the filtering idea is similar to Gaussian filtering, and the difference is that the Gaussian filtering only considers the airspace and value domain information of pixel points, and can well retain the image edge information.
(3) Network selection
The YOLO (You Only Look one) algorithm is a single-step end-to-end target detection algorithm, integrates candidate region extraction and recognition classification, and has the advantages of high detection speed, small model file and the like. Through the update iteration, the YOLO v5 model is now updated. The YOLO v5 model has 4 versions of YOLO v5s, YOLO v5m, YOLO v5l and YOLO v5x, wherein a YOLO v5s network is the minimum depth in a YOLO v5 series, the minimum model file and the basis of other versions of networks. The real-time performance and the accuracy of detection are integrated, and the YOLO v5s network is used as a flame detection network in the experiment.
(2) Jet drop point extraction
And (3) making a jet flow track semantic segmentation data set, loading the jet flow track semantic segmentation data set into an improved semantic segmentation network for training to obtain a clear and complete jet flow track, analyzing the rule of the jet flow track, and deciding an outgoing flow drop point coordinate. Fig. 2 is a flowchart of a jet drop location identification process.
(1) Data collection and annotation
848 jet pictures are collected from the internet and subjected to data enhancement, including land fire protection, sea law enforcement, indoor fire protection, agricultural irrigation and the like, and the universality of the model is enhanced by covering as many scenes as possible. The jet area was manually extracted using Labelme labeling software.
(2) Network selection
Selecting a U-net network to extract the jet flow track, adding an attention module between down sampling and up sampling to fully fuse the context extraction information, and inputting the fused information into an up sampling network, wherein fig. 3 is a structural diagram of the improved U-net network.
(3) Jet drop point extraction
Analyzing the characteristics of the jet flow trajectory: as the fire water monitor has overlarge pressure and longer jet distance, the jet flow track keeps better directionality in a certain distance. Therefore, under the condition that the jet starting point of the water cannon is known, the extracted jet trajectory is subjected to binarization processing, the point farthest from the jet starting point is the landing point of the jet trajectory, and fig. 4 is a flow chart of the jet landing point identification process.
2) Predicting and quantifying ship motion attitude interference
(1) Ship motion attitude prediction
In order to cope with the situation that the accuracy of the prediction model is reduced due to the change of the environment, the optimal hidden layer number, the learning rate and the random inactivation rate in the current environment are searched through an improved enhanced wolf optimization algorithm (EGWO), and the ship attitude prediction model is dynamically constructed, so that the current optimal prediction model is adaptively constructed along with the change of the working conditions. FIG. 5 is a flow chart of the EGWO-LSTM method, and the improved enhanced gray-optimized wolf algorithm is specifically as follows:
(1) crossover, mutation operators
The traditional GWO design is used for solving the continuous optimization problem and cannot be directly used for processing the discrete problem, and meanwhile, in order to improve the population diversity and enhance the intra-population and inter-population communication, the cross and mutation operators are designed by referring to the genetic algorithm to enhance the capacity of the wolf search. The crossover operation may be represented by equation (1), where α wolf, β wolf, and δ wolf are also used to guide the search process.
Wherein, X k Is the proposal of the kth wolf in the wolf group,
is a new scheme of the kth wolf in the wolf group, X α 、X β And X δ Denote the schemes of α wolf, β wolf and δ wolf respectively, f is the proposed crossover operator, rand stands for [0,1 ]]The random number of (2). The proposed crossover operator f is shown in fig. 6. The specific crossing scheme is to randomly generate the crossing point positions and then replace the values of the same positions in the current gray wolf scheme with the values of the crossing points in the leading wolf.
To enhance the global search capability of the algorithm to avoid premature convergence, variant operations are performed using equation (2).
Wherein g is the mutation operator proposed, P m Is the mutation probability.
FIG. 7 shows a mutation operator g. The specific variation scheme is to randomly generate the positions of 2 variation points, and then randomly generate the values of the 2 variation points of the current grayish wolf scheme within the constraint range.
(2) Multi-wolf group collaboration
One key issue in improving the performance of the GWO algorithm is how to balance exploration and development. Therefore, the invention divides the whole wolf group into three sub wolf groups, which are a main wolf group, a global auxiliary wolf group and a local auxiliary wolf group. By designing different variation strategies and search tasks, the balance of exploration and development is achieved.
Different wolf groups have different tasks, so three variation probabilities are designed to match with the functions of the wolf groups, and the specific calculation of the three variation probabilities is as follows.
Where t is the current iteration number, t max Is the maximum number of iterations. The three wolf groups are described as follows:
1. the main wolf group: the main wolf group is a main body optimizing wolf group in the whole algorithm execution process, the total number of individuals is N1, a local auxiliary wolf group is constructed, and a better scheme can be obtained from two auxiliary wolf groups. The iteration of the median of the main wolf group can increase the diversity of particles, and meanwhile, the inferior scheme in the main wolf group is replaced by the superior scheme in the two auxiliary wolf groups, so that the search range can be expanded, and the local optimum is avoided.
2. Global auxiliary wolf group: in order to improve the global optimizing capability of GWO, a global auxiliary wolf pack is constructed by randomly generating values in a constraint range, and the total number of individuals is N2. After each iteration is finished, the individual wolf with the fitness degree of the top N1 in all individuals in the global auxiliary wolf group is replaced into the main wolf group.
3. Local auxiliary wolf group: in order to improve the local optimizing capability of GWO, the local auxiliary wolf group is randomly constructed by taking the values of alpha wolf and delta wolf in the main wolf group as the corresponding upper and lower limits, and the total number of individuals is N3. And after each iteration is finished, replacing the individual wolf with fitness ranking top N1 in all individuals in the local auxiliary wolf group into the main wolf group.
The position updating formula of each wolf group is as follows.
Wherein, X m,k Is the current individual protocol, X m,k+1 Is a new scheme for the current individual,
a wolf alpha, a wolf beta, a wolf delta of a wolf group m (main, global auxiliary or local auxiliary wolf group), a is a convergence factor that drops linearly from 2 to 0, is/is>
Respectively, X calculated by the formula (5) m,k The wolf located in the wolf group guides the wolf->
(alpha wolf, beta wolf, delta wolf).
(2) Ship motion attitude change interference quantification
After the predicted ship motion attitude is obtained through EGWO-LSTM, in order to quantify the ship attitude change interference on the fire monitor by using a geometric method, the ship motion attitude and the fire monitor control quantity are defined as table 1, wherein the ship is inclined to the left at maximum of-180 degrees, the ship is inclined to the back at maximum of-90 degrees, and the horizontal angle when the fire monitor points to the beam is 0 degree.
TABLE 1 Ship motion attitude and fire monitor control quantity definition
After the ship type value parameters and the installation position of the fire water monitor are determined, the coupling motion of rolling and pitching is ignored, and the interference of the rolling and pitching on the pitch angle and the horizontal deflection angle of the fire water monitor is quantified by using a geometric method.
The pitch angle θ is disturbed mainly by roll. When boats and ships roll attitude changes, the angle of pitch of fire water monitor relative horizontal plane can corresponding production change, the angle of pitch of fire water monitor expectation at this moment will be different with boats and ships before rolling attitude changes, then the interference quantity theta of angle of pitch at next moment d (t + 1), which can be expressed by using the roll angle difference between the next time and the current time, as follows:
θ d (t+1)＝α roll (t+1)-α roll (t) (6)
wherein alpha is roll (t) is the roll angle at the present time, α roll (t + 1) is the roll angle at the next time.
The horizontal declination angle beta is disturbed mainly by pitching. When the ship pitching attitude changes, the fire water monitor can generate fore-aft direction displacement, the expected horizontal deflection angle of the fire water monitor at the moment can change, and the connecting line between the fire water monitor and the target point at the next moment and the current moment and the connecting line between the fire water monitor and the target point can be used
The difference of the included angles in the negative direction is used as the interference quantization value beta of the horizontal deflection angle at the next moment d (t + 1), calculated specifically as follows:
Δx＝hsinα(t+1)-hsinα pitch (t) (7)
δ(t+1)＝arctan(SRsinδ(t),Δx+SRcosδ(t))
β d (t+1)＝δ(t+1)-δ(t)
wherein, deltax is the X-axis direction displacement caused by the change of the pitching attitude, h is the distance between the fire water monitor and the horizontal plane of the ship gravity center, and alpha pitch (t) and alpha pitch (t + 1) are the current times, respectivelyMoment and next moment longitudinal rocking angle, SR is the linear distance between the fire monitor at the current moment and the target point, and delta (t) and delta (t + 1) are respectively the connecting line between the fire monitor at the current moment and the next moment and the target point and
the angle in the negative direction.
3) Fire water monitor control
In order to improve the self-adaptive capacity of the controller, a variable-discourse-domain fuzzy control theory is introduced, a self-adaptive module is added in the fuzzy control, the parameters of the fuzzy controller are solved by using an online particle swarm algorithm, and the solved result is fed back to the fuzzy controller, so that the self-adaptive adjustment of the fuzzy controller is realized.
The design idea of the self-adaptive fuzzy controller is as follows: the formula (8) is used as an online particle swarm algorithm Fitness function Fitness, parameters (including a scale factor and a quantization factor) of the fuzzy controller are adjusted in real time, in order to solve the problem of overshoot, a weight coefficient is added, and through testing, the performance is better when the weight coefficient is set to be 1:4, and finally, the quick, accurate and stable offshore fire fighting operation is realized.
Wherein, the delta SR is the difference of the linear distance between the current water flow falling point and the coordinates of the target point and the fire water monitor, and the delta is the connecting line between the current water flow falling point and the coordinates of the target point and the fire water monitor and the coordinates of the fire water monitor
The difference in negative direction angles. The flow chart of the online particle swarm algorithm for adaptively adjusting the fuzzy controller parameters is shown in fig. 8.
In addition, in order to resist the control error of the ship-based fire monitor caused by the interference of wind and waves, the adaptive fuzzy controller 2 is added into the adaptive fuzzy control system based on visual recognition. The output variable of the adaptive fuzzy controller 1 is a pitch angle primary adjustment value delta theta 1 The primary adjustment value delta beta of the horizontal deflection angle 1 (ii) a Input deviceThe variables are the difference delta SR between the current water flow falling point and the linear distance between the target point and the fire monitor and the change rate d delta SR thereof, and the connecting line between the current water flow falling point and the target point and the fire monitor and the change rate d delta SR
The included angle difference delta in the negative direction and the change rate d delta thereof. The output variable of the adaptive fuzzy controller 2 is a pitch angle compensation value delta theta 2 Horizontal deflection angle compensation value delta beta 2 (ii) a Input variable being pitch angle disturbance theta d (t + 1) and its rate of change d θ d Horizontal declination interference beta d (t + 1) and the rate of change d β thereof d . And adding the primary adjustment value and the compensation value to be used as a total adjustment value of the attitude angle of the fire water monitor at the next moment, and carrying out closed-loop adjustment on the two attitude angles of the fire water monitor so as to deal with the interference of the change of the motion attitude of the ship. A flow chart of the control method of the carrier-based fire monitor is shown in fig. 9.
The fuzzy control membership function is isosceles triangle, as shown in FIG. 10. The input variables and the output variables are divided into 7 fuzzy sets, namely NB, NM, NS, ZE, PS, PM and PB, which respectively represent negative large, negative medium, negative small, zero, positive small, positive medium and positive large.
According to the experience of adjusting the horizontal deflection angle and the pitch angle, for the controller 2, when the pitch angle interferes with theta d (t + 1) is large, and d θ d And theta d When the directions of (t + 1) are opposite, the pitch angle compensation value is larger; when the horizontal deflection angle interferes with beta d (t + 1) is large, and d β d And beta d When the (t + 1) direction is opposite, the horizontal declination compensation value should be larger. The same rules apply to the controller 1, the fuzzy rule design is shown in tables 2 and 3.
TABLE 2. DELTA.theta. 1 、Δθ 2 Fuzzy rule table of
TABLE 3. DELTA.beta 1 、Δβ 2 Fuzzy rule table of
For the adaptive fuzzy controllers 1 and 2, the fuzzy decision adopts a Mamdani type fuzzy reasoning method, and the defuzzification method adopts a gravity center method.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (7)
1. A control method for a ship-based fire water monitor is characterized by comprising the following steps:
identifying the position of a fire source point and a jet flow drop point of the fire water monitor according to the image;
acquiring motion attitude prediction information of a ship carrying a fire water monitor through a long-short term memory neural network, wherein the motion attitude prediction information comprises rolling and pitching; the interference of the motion attitude prediction information on the attitude angle of the fire water monitor is quantified by using a geometric method; the attitude angle of the fire water monitor comprises a pitch angle and a horizontal deflection angle;
inputting the position deviation of a fire source point and a jet flow drop point into a first self-adaptive fuzzy controller to obtain a primary adjustment value of the attitude angle of the fire water monitor; inputting the interference of the motion attitude prediction information on the attitude angle of the fire water monitor into a second self-adaptive fuzzy controller to obtain a compensation adjustment value of the attitude angle of the fire water monitor; and accumulating the primary adjustment value and the compensation adjustment value to obtain a total adjustment value of the attitude angle.
2. The control method for the ship-based fire monitor according to claim 1, characterized in that: in the process of identifying the fire source point, a YOLO v5s network is adopted as a flame detection network to carry out bilateral filtering on the image when the input image is preprocessed.
3. The method as claimed in claim 2, wherein during the training of the flame detection network, the fire images are collected, the fire images are labeled to obtain a data set, and the labeled data set is subjected to data enhancement, wherein the data enhancement includes translation and rotation of the images.
4. The method as claimed in claim 1, wherein the process of identifying the jet landing point from the image comprises:
extracting a jet flow track from an image by adopting an improved U-net network, wherein the U-net network is obtained by adding an attention module between a lower sampling layer and an upper sampling layer on the basis of the existing U-net network;
and carrying out binarization processing on the extracted jet flow track, and extracting a point farthest from the jet flow starting point as a jet flow falling point.
5. The method for controlling the shipborne fire water monitor as recited in claim 1, wherein a long-short term memory neural network for acquiring motion attitude prediction information is periodically trained and constructed; in each training and building process, the super parameters of the long-term and short-term memory neural network are adjusted by adopting an enhanced wolf optimization algorithm; the super parameters comprise hidden layer number, learning rate and random inactivation rate.
6. The method of claim 1, wherein for the first adaptive fuzzy controller and the second adaptive fuzzy controller,
the input variables and the output variables are divided into 7 fuzzy sets, namely NB, NM, NS, ZE, PS, PM and PB which respectively and correspondingly represent negative large, negative middle, negative small, zero, positive small, positive middle and positive large;
mapping the input variable and the output variable to a fuzzy control membership function of a corresponding fuzzy set by adopting an isosceles triangle;
the defuzzification method selects a gravity center method.
7. The method as claimed in claim 6, wherein the fuzzy controller is improved through an online particle swarm algorithm, a fitness function is designed to evaluate the merits of the scale factor and the quantization factor of the current fuzzy controller, and the adaptive adjustment is performed on the fitness function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310078364.0A CN115944868B (en) | 2023-01-16 | 2023-01-16 | Ship-borne fire water monitor control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310078364.0A CN115944868B (en) | 2023-01-16 | 2023-01-16 | Ship-borne fire water monitor control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115944868A true CN115944868A (en) | 2023-04-11 |
| CN115944868B CN115944868B (en) | 2024-04-05 |
Family
ID=87285967
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310078364.0A Active CN115944868B (en) | 2023-01-16 | 2023-01-16 | Ship-borne fire water monitor control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115944868B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116842381A (en) * | 2023-06-13 | 2023-10-03 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely-short-term prediction model generalization optimization method based on data fusion |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113877124A (en) * | 2021-11-15 | 2022-01-04 | 应急管理部天津消防研究所 | Intelligent control system for jet flow falling point of fire monitor |
| KR20220001124A (en) * | 2020-06-29 | 2022-01-05 | 충남대학교산학협력단 | Method for reinforced learning of fire suppression models related to autonomous ship with motion in irregular waves |
| CN114082568A (en) * | 2021-11-19 | 2022-02-25 | 江苏科技大学 | Large ship body segmented spraying operation offline digital-analog planning system and method thereof |
-
2023
- 2023-01-16 CN CN202310078364.0A patent/CN115944868B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20220001124A (en) * | 2020-06-29 | 2022-01-05 | 충남대학교산학협력단 | Method for reinforced learning of fire suppression models related to autonomous ship with motion in irregular waves |
| CN113877124A (en) * | 2021-11-15 | 2022-01-04 | 应急管理部天津消防研究所 | Intelligent control system for jet flow falling point of fire monitor |
| CN114082568A (en) * | 2021-11-19 | 2022-02-25 | 江苏科技大学 | Large ship body segmented spraying operation offline digital-analog planning system and method thereof |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116842381A (en) * | 2023-06-13 | 2023-10-03 | 青岛哈尔滨工程大学创新发展中心 | Ship motion extremely-short-term prediction model generalization optimization method based on data fusion |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115944868B (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107909082B (en) | Sonar image target identification method based on deep learning technology | |
| CN110232350A (en) | A kind of real-time water surface multiple mobile object detecting and tracking method based on on-line study | |
| CN107818302A (en) | Non-rigid multiple dimensioned object detecting method based on convolutional neural networks | |
| CN107229904A (en) | A kind of object detection and recognition method based on deep learning | |
| CN112836893B (en) | Method for predicting ship oil consumption under severe sea conditions based on sea condition and ship navigation condition | |
| CN109785385A (en) | Visual target tracking method and system | |
| Li et al. | Incorporation of AIS data-based machine learning into unsupervised route planning for maritime autonomous surface ships | |
| CN110189304A (en) | Remote sensing image target on-line quick detection method based on artificial intelligence | |
| CN113011430B (en) | Large-scale point cloud semantic segmentation method and system | |
| CN108334859B (en) | Optical remote sensing ship model identification crowdsourcing system based on fine-grained characteristics | |
| CN110991257B (en) | Polarized SAR oil spill detection method based on feature fusion and SVM | |
| CN115944868B (en) | Ship-borne fire water monitor control method | |
| CN111652271A (en) | Nonlinear feature selection method based on neural network | |
| CN109670660A (en) | A kind of fleet dynamic air defense threat estimating method based on intuitionistic fuzzy TOPSIS | |
| CN111079790A (en) | Image classification method for constructing class center | |
| CN113589272A (en) | Automatic generation method for target tracking equipment on-duty log | |
| CN111325259A (en) | Remote sensing image classification method based on deep learning and binary coding | |
| CN116108383A (en) | Ship track prediction method based on improved goblet sea squirt multi-output support vector | |
| CN117237902B (en) | Robot character recognition system based on deep learning | |
| Mirakhorlo et al. | Integration of SimWeight and Markov Chain to Predict Land Use of Lavasanat Basin | |
| CN110111358B (en) | Target tracking method based on multilayer time sequence filtering | |
| CN116167902A (en) | Multi-attribute analysis-based water surface target safety evaluation algorithm model and method | |
| CN115082734A (en) | Aluminum electrolysis cell fire eye video inspection system and superheat degree deep learning identification method | |
| Zhang et al. | Semantic segmentation of point clouds of field obstacle-crossing terrain for multi-legged rescue equipment based on random forest | |
| Shi et al. | Optimal kinematics design of macpherson suspension: Integrated use of grey relational analysis and improved entropy weight method |
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 |