CN116027673A - Equipment control autonomous decision-making method based on fuzzy neural network - Google Patents
Equipment control autonomous decision-making method based on fuzzy neural network Download PDFInfo
- Publication number
- CN116027673A CN116027673A CN202310318098.4A CN202310318098A CN116027673A CN 116027673 A CN116027673 A CN 116027673A CN 202310318098 A CN202310318098 A CN 202310318098A CN 116027673 A CN116027673 A CN 116027673A
- Authority
- CN
- China
- Prior art keywords
- fuzzy
- neural network
- control
- equipment
- equipment control
- 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 50
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 35
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 9
- 230000006870 function Effects 0.000 claims description 36
- 238000003062 neural network model Methods 0.000 claims description 10
- 230000008901 benefit Effects 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 5
- 230000007123 defense Effects 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims 1
- 230000008859 change Effects 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 7
- 238000010801 machine learning Methods 0.000 description 5
- 230000003044 adaptive effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000001052 transient effect Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012805 post-processing Methods 0.000 description 1
- 239000002243 precursor Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
- 230000035899 viability Effects 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Feedback Control In General (AREA)
Abstract
The invention discloses an equipment control autonomous decision-making method based on a fuzzy neural network, which comprises the following steps: judging the current situation domain of the equipment to form an equipment control rule base, converting the equipment control rule into an equipment control fuzzy rule to form a fuzzy sample by a fuzzification method, constructing a fuzzy logic reasoning model to complete input fuzzification processing, fuzzy reasoning and output defuzzification, realizing real-time quick decision of maneuvering control, radar control, weapon control and electronic countermeasure control, and carrying out self-adaptive learning training on the structure and parameters of the equipment control fuzzy rule. According to the invention, a neural network learning scheme based on fuzzy logic is adopted, the equipment is rapidly provided with autonomous decision-making capability based on expert combat experience, self-adaptive learning training is carried out through a neural network learning algorithm, the efficient real-time evolution of the equipment control autonomous decision-making capability is completed, the learning period is effectively shortened, and the adaptability and the intelligent level of the equipment to external environment change are improved.
Description
Technical Field
The invention relates to the technical field of equipment control autonomous decision making, in particular to an equipment control autonomous decision making method based on a fuzzy neural network.
Background
The modern combat equipment is developed in the directions of high automation, informatization and intelligence, the information provided for commanders or equipment operators is complex and various, and particularly, the modern combat equipment has multiple input and output elements aiming at a plurality of control factors such as maneuvering control, radar control, weapon control, electronic countermeasure control and the like, the equipment control problem has multiple input and output elements, and the elements are mutually related, so that the modern combat equipment is a highly complex, strong nonlinear and constrained dynamic optimization solving problem, and is almost impossible to make optimal decision control by only relying on people. Therefore, an intelligent autonomous decision-making method for equipment control needs to be studied to assist an equipment operator to make real-time decisions on maneuvering control, radar control, weapon control, electronic countermeasure control and other aspects in face of complex battlefield situations, reduce the operational burden of the operator, improve the grasping ability of the operator to external environments, enable the performance characteristics of the equipment to meet special task targets and environment requirements, and improve the viability and task efficiency of the equipment.
For the control of mobile equipment in a complex environment, an autonomous decision method with good performance is not yet adopted. The existing research is mainly focused on the aspects of traditional artificial intelligence methods such as rule-based expert systems, machine learning and the like, and mainly has the following problems.
1. The expert system based on rules makes reasoning directly according to expert knowledge and rules, compared with other decision methods, the expert system method does not need to perform a large amount of calculation, can react faster for real battlefield environments with changeable transient conditions, and has the characteristics of high response speed and simple structure. Then, as the knowledge stored in the expert system knowledge base is solidified data, once the target scene which is not stored in the system knowledge base appears in the using process of the equipment, the expert system fails and cannot adapt to the battlefield environment which is changed in a transient manner.
2. The traditional machine learning algorithm is based on massive learning samples, a machine learning model similar to a black box is built, and a large amount of input and output data is subjected to learning training, so that the mode has high requirements on sample data, long learning training period, low learning speed, poor generalization performance and unsecured effect, and is a great challenge for the operation time and the storage space of a computer. Aiming at the equipment autonomous decision-making problem, autonomous decision-making is carried out on multiple dimensions such as maneuvering control, radar control, weapon control, electronic countermeasure control and the like according to complex state information of the friend and foe, the input and output elements are multiple, the elements are mutually related, the complexity is high, the data size is large, meanwhile, the real combat sample data acquisition difficulty is high, the cost is high, and the traditional machine learning algorithm is difficult to effectively process and analyze.
Disclosure of Invention
Aiming at the problems of poor adaptability, high complexity, large required data quantity, poor timeliness and high cost of the traditional equipment control decision method, the invention provides an equipment control autonomous decision method based on a fuzzy neural network, which adopts a neural network learning scheme based on fuzzy logic, forms an equipment control fuzzy rule based on expert combat experience under the condition of limited number of combat samples, rapidly enables the equipment to have autonomous decision capability through fuzzy reasoning, carries out self-adaptive learning training on fuzzy rule structures and parameters through the neural network learning algorithm, completes the efficient real-time evolution of the equipment control autonomous decision capability, effectively shortens the learning period, and improves the adaptability and the intelligent level of the equipment to external environment change.
To achieve the above object, the present invention provides an equipment control autonomous decision method based on a fuzzy neural network, the method comprising the steps of:
s1: acquiring target data of a battlefield enemy, and judging the current situation domain of equipment;
s2: information extraction is carried out on equipment combat experience knowledge or sample data according to an input-output structure to form an equipment control rule base, the equipment control rule base is converted into a fuzzy rule through a fuzzification method to form a fuzzy sample, and a fuzzy logic reasoning model is constructed;
s3: the membership function is selected to carry out fuzzification processing on the input condition domain, the target distance, the target azimuth angle and the target entry angle data, a neural network learning model is constructed based on a fuzzy sample, and a BP algorithm is adopted to carry out learning training on the front part parameters;
s4: obtaining fuzzy output through fuzzy reasoning, optimizing the front part parameters through the S3, and learning and training the rear part parameters based on fuzzy samples by utilizing a least square algorithm in a neural network learning model;
s5: and (3) defuzzifying the fuzzy output to obtain a control signal, controlling corresponding equipment to work, judging whether the control requirement is met, ending if the control requirement is met, and returning to the step (S1) otherwise.
Optionally, the case domain includes information assurance, advantage assurance, attack, defense, and defending and attacking; wherein, the information assurance is: no target is found, or only direction finding information; the advantage is ensured as follows: the target has been found and there is direction-finding ranging information.
Optionally, the step S1 specifically includes: and acquiring target data of a battlefield enemy, and judging the current situation domain of the equipment according to the target data and a preset situation domain rule.
Optionally, the equipment control rule in the equipment control rule base is a relation from observing input to conclusion output, and specifically includes: a number of observation item condition information and a conclusion item associated with the number of condition information.
Optionally, the observation items include a case field, a target distance, a target azimuth, and a target entry angle; the conclusion items include maneuver direction, maneuver speed, radar control, weapon control, and electronic countermeasure control.
Optionally, the fuzzy logic reasoning model comprises a fuzzification layer, an applicability layer, a normalization layer, a fuzzy output layer and a total output layer; the input of the fuzzy logic inference model comprises a target distance=[-500,500]Target azimuth->= [-180,180]Target entry angle->= [-180,180]Case Domain->= { 1-info guarantee, 2-advantage guarantee, 3-attack, 4-defense, 5-defense and attack target }; the output includes the maneuver direction->= { 1-pull-up, 2-dive, 3-turn left, 4-turn right }, maneuver speed +.>= { 1-acceleration, 2-deceleration, 3-average speed }, radar control +.>= { 1-not open, 2-search, 3-track }, weapon control +.>= { 1-no emission, 2-emission }, electronic countermeasure control +.>= { 1-scout, 2-interference }.
Optionally, in step S3, the membership function is selected to perform the blurring processing on the input data of the case domain, the target distance, the target azimuth angle, and the target entry angle, specifically: numerical transformation of input variables for autonomous decision making of equipment controlForming the language value of the fuzzy language variable; according to the size of the input variable data, the target distance is calculatedTarget azimuth->Target entry angleAnd case Domain->Dividing into 5 fuzzy subsets: positive large, positive small, zero, negative small, negative large; wherein:
the membership function uses a gaussian function:
wherein ,for inputting variable +.>Is (are) fuzzy set, < >> and />Representing the center and width of the membership function respectively,membership function value, representing the input variable +.>Belonging to fuzzy set->To a degree of (3). />
Optionally, in step S3, a neural network learning model is constructed based on the fuzzy sample, and a BP algorithm is used to learn and train the parameters of the front part, which specifically includes: based on the fuzzy sample, a BP algorithm in a neural network model is adopted to learn and train the front piece parameters such as the center, the width and the like of the membership function, the front piece parameters are fed back to a fuzzy logic reasoning model, and the shape of the fuzzy membership function is optimized and updated.
Optionally, in step S4, fuzzy output is obtained through fuzzy reasoning, specifically:
wherein ,for the total number of fuzzy rules, +.>For inputting a number of numbers>For the number of outputs, ->Representing the applicability of each fuzzy rule, +.>Normalized fitness representing each fuzzy rule, < >>Fuzzy outputs representing each fuzzy rule, < >>The conclusion parameters are also called back-piece parameters for fuzzy reasoning.
Optionally, in step S4, based on the fuzzy sample, learning and training the back-part parameter by using a least squares algorithm in the neural network model specifically includes: based on the fuzzy sample, the least square algorithm in the neural network model is adopted to learn and train the back-piece parameters, and the back-piece parameters are transmitted to the fuzzy logic reasoning model to update the fuzzy reasoning parameters in an optimized mode.
Compared with the prior art, the equipment control autonomous decision method based on the fuzzy neural network has the following beneficial effects: (1) Aiming at the problems of more input and output elements, mutual correlation among the elements and high complexity of the control decision-making problem of modern equipment, the invention firstly judges the current situation domain according to target data of a battlefield enemy, then takes the situation domain, the target distance, the target azimuth angle and the target entry angle as fuzzy input, solves the problems by using a fuzzy logic reasoning model, assists an equipment operator to make real-time and quick decisions on maneuvering control, radar control, weapon control and electronic countermeasure control in the face of complex battlefield situation, reduces the problem range and the calculated quantity and improves timeliness; (2) Aiming at the problems that the traditional machine learning method needs to be based on massive learning samples, has long learning training period and poor generalization performance, and real combat sample data acquisition difficulty is high, expert combat experience knowledge is introduced, and fuzzy rules are formed by combining the expert combat experience knowledge with a fuzzy reasoning model, so that the existing combat knowledge and strategies can be efficiently utilized, equipment is provided with autonomous decision making capability rapidly, and the learning period is effectively shortened; (3) Aiming at the problem that the traditional equipment control method cannot adapt to the transient and changeable battlefield environment, a learning mechanism of a neural network is introduced into a fuzzy logic control model, and self-adaptive learning training is carried out on a fuzzy rule structure and parameters through a neural network learning algorithm, so that the efficient real-time evolution of the equipment control autonomous decision-making capability is completed, and the adaptability and the intelligent level of the equipment to external environment change are improved.
Drawings
FIG. 1 is a flow diagram of an overall implementation of the present invention.
Fig. 2 is a schematic diagram of an adaptive fuzzy logic inference model of the present invention.
Fig. 3 is a schematic diagram of a learning training process of the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The design scheme provided by the invention is a device control autonomous decision method based on a fuzzy neural network, the implementation flow of which is shown in figure 1, and the specific method is as follows:
s1: based on battlefield enemy target data acquired by equipment such as radar, electronic countermeasures and the like, the current situation domain of the equipment is judged according to the situation domain rule, different situation domains are divided for solving respectively, and the problem solving space is reduced.
S2: information extraction is carried out on equipment combat experience knowledge or sample data according to an input-output structure to form an equipment control rule base, the equipment control rule base is converted into a fuzzy rule through a fuzzification method to form a fuzzy sample, and a fuzzy logic reasoning model is constructed;
s3: and (3) selecting a membership function to carry out fuzzification processing on the input condition domain, the target distance, the target azimuth angle and the target entry angle data, and based on a fuzzy sample, adopting a BP algorithm in a neural network model to learn and train the front part parameters, so as to improve the description accuracy of the fuzzification membership function on the input information.
S4: and obtaining fuzzy output through fuzzy reasoning, optimizing the front part parameters through the S3, and based on the fuzzy sample, utilizing a least square algorithm in a neural network model to learn and train the rear part parameters, so that the accuracy of the output is improved.
S5: and (3) defuzzifying the equipment control fuzzy output to obtain real maneuvering control, radar control, weapon control and electronic countermeasure control output, controlling corresponding equipment to work, judging whether the control requirement is met, ending if the control requirement is met, and returning to S1 otherwise.
The following describes the procedure in detail with reference to specific examples.
Step S1, aiming at the autonomous decision-making problem of equipment control, firstly judging the current situation domain of the equipment based on the acquired battlefield data, and then determining control outputs of maneuvering control, radar control, weapon control, electronic countermeasure control and the like of the equipment according to input parameters such as the situation domain, the target distance, the target azimuth angle, the target entry angle and the like.
For autonomous decision-making of single air combat, typical situation domains of equipment control comprise information assurance, advantage assurance, attack, defense and attack, and the factors mainly considered in the decision of the situation domains are as follows: whether the target is in the radar detection range; whether the target is within the attack area; whether the missile can be launched; whether locked by the counterpart missile; whether within the attack area of the target machine, etc. The situation domain determination is mainly performed according to the target detectable state (c is discoverable, n is not discoverable) and the missile attacked state (a is attackeable, n is not attackeable). The conditions and effects to be achieved are shown in Table 1.
Table 1: typical case domain of equipment control decisions
Step S2, the input information of the equipment control decision problem mainly comprises distance=[-500,500]Azimuth of target= [-180,180]Target entry angle->= [-180,180]Case Domain->= { 1-info guarantee, 2-dominance guarantee, 3-attack4-defensive, 5-defensive and attack target }, maneuver control output comprises maneuver direction +.>= { 1-pull-up, 2-dive, 3-turn left, 4-turn right }, maneuver speed +.>= { 1-acceleration, 2-deceleration, 3-average speed }, radar control +.>= { 1-not open, 2-search, 3-track }, weapon control +.>= { 1-no emission, 2-emission }, electronic countermeasure control +.>= { 1-scout, 2-interference }. According to the size of the input variable data, the target distance +.>Target azimuth->And target entry angle->Dividing into 5 fuzzy subsets: positive Large (PL), positive Small (PS), zero (ZE), negative Small (NS), negative Large (NL).
Typical equipment control rules for single machine air combat equipment control decisions are shown in table 2.
Table 2: equipment control rules
According to the equipment control rules, an equipment control fuzzy rule base can be formed through fuzzification processing, and the basic format is as follows:
as shown in FIG. 2, the adaptive fuzzy logic inference model is adopted for solving, and the method mainly comprises five parts, namely layer 1 fuzzification, layer 2 fitness, layer 3 normalization, layer 4 fuzzy output and layer 5 total output, as shown in FIG. 2. For equipment control decision problem, its input is,/>For the number of input quantity, the output is +.>,/>And m is the total number of fuzzy rules for the number of output quantities.
wherein, the condition domain, the target distance, the target azimuth angle and the target entry angle are 4 input variablesThe number of fuzzy linguistic variables is 5, < ->Is a fuzzy set of input variables. />The membership function value of A is the input variable +.>Belonging to fuzzy set->The shape of the membership function is entirely determined by parameters called the precursor parameters. Here, the membership function is a gaussian function, and then:
wherein , and />Representing the center and width of the membership functions, respectively, the shape of the membership functions being determined primarily by them.
Layer 2 applicability: each node of the layer represents a fuzzy rule, which is used for matching with the fuzzy rule front piece to calculate each nodeThe suitability of the bar rule, the output of this layer being multiplication of the input signals, "in the figure""means multiplication of its corresponding input information.
wherein ,,/>the output of each node is +.>Representing the fitness of the rule. Here "/->"may be any AND operator that meets the specification. For a given input, only those language variable values near the input point have a greater degree of membership, and language variable values far from the input point have a smaller degree of membership, and are taken as 0 when the degree of membership is less than 0.01. Thus, only a small number of nodes output non-0's in the inference layer, while the majority of nodes output 0's.
Layer 3 normalization: normalization processing for each rule fitness of the layer, the firstThe personal node calculates->Normalized fitness of bar rules
Layer 4 fuzzy output: calculating the output of each fuzzy rule of the layer, and each node of the layerIs an adaptive node, th->The output of the individual node is +.>
Here, theFor the third layer output, +>The conclusion parameters are also referred to as the back-piece parameters.
Layer 5 total output: for a fixed node, the total output of all input signals is calculated:
the equipment control autonomous decision making of the fuzzy neural network is advantageous in that the modeling method based on knowledge data, the modeling membership function in the system and the establishment of the fuzzy rule are obtained by post-learning adjustment optimization of known sample data and expert experience knowledge, not just based on experience or intuitionally arbitrary given. The method combines the actual control problem characteristics of the autonomous decision of the equipment control, adopts a mixed algorithm combining a BP algorithm and a least square algorithm to learn, realizes the learning and adjustment of rule front-part parameters through the BP algorithm, realizes the learning and adjustment of rule back-part parameters through the least square algorithm, and is shown in a figure 3 in the autonomous decision learning and training flow of the equipment control based on the fuzzy neural network.
And S3, determining membership functions through a fuzzy logic reasoning model, initializing rule front part parameters, and carrying out rule fuzzification processing on input data of a fuzzy rule or a sample to obtain fuzzy input. Meanwhile, the BP algorithm in the neural network model learns and trains the front part parameters through fuzzy rules or sample data, and continuously transmits the learned and optimized front part parameters to the fuzzy logic reasoning model to update the shape of the fuzzy membership function, so that the description accuracy of the fuzzy membership function on the input information is improved.
The fuzzy neural network model can be described as follows from the foregoing analysis
Where m is the number of fuzzy rules already determined,for input, & lt + & gt>For outputting (I)>,/> and />Is a freely variable parameter that needs to be learned, wherein, < -> and />For the front piece parameter +.>For the back-piece parameter-> and />Parameters of Gaussian membership function, table->Show->The rule center average receives the center value of the blur er.
The system can be regarded as a feed-forward network system, so that the task is changed to the problem of training the parameters of the feed-forward network by using a gradient descent method, wherein the algorithm is also called an error back propagation algorithm, namely a BP algorithm, which is also a basic algorithm in a neural network system, and the BP algorithm is used for learning the parameters of the front part and />. Setting an objective function as follows:
According to the gradient descent method, the following steps can be obtained:
wherein ,,/>is the learning step size. According to the derivation rule of the composite function, the following can be obtained:
wherein ,
and S4, determining fuzzy operators related to the operation of the reasoning rules, initializing the parameters of the post-processing products, obtaining fuzzy output through fuzzy reasoning, and researching the fuzzy theory to provide a plurality of fuzzy operators, wherein the common reasoning operation methods are two operators of maximum-minimum (MAX-MIN) and maximum-product (MAX-PROD). Meanwhile, the front part parameters are optimized through the sample data and the BP algorithm, the rule back part parameters are learned and trained through the least square algorithm, the learned and optimized back part parameters are continuously transmitted to the fuzzy logic reasoning model to update the fuzzy reasoning parameters in an iterative mode, the expected mapping relation of input and output is achieved, and the accuracy of output is improved.
Learning of back-piece parameters by least squares. The recursive least squares method minimizes the sum of the fitting errors of all input-output data pairs of all 1 to p. That is, the objective function is set as:
for the fuzzy neural network system, the fuzzy rule number m and the front piece parameters are determined firstly by adopting recursive least square method learning and />Thereby forming a fuzzy basis function->。/>
Then, setting:
wherein :
the fuzzy neural network system may be described as:
And S5, performing defuzzification on the fuzzy output to obtain real output, judging whether the control requirement is met, ending if the control requirement is met, otherwise, returning to the first step to continue training and learning until the requirement is met.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) as described above, comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (10)
1. An equipment control autonomous decision making method based on a fuzzy neural network, which is characterized by comprising the following steps:
s1: acquiring target data of a battlefield enemy, and judging the current situation domain of equipment;
s2: information extraction is carried out on equipment combat experience knowledge or sample data according to an input-output structure to form an equipment control rule base, the equipment control rule base is converted into a fuzzy rule through a fuzzification method to form a fuzzy sample, and a fuzzy logic reasoning model is constructed;
s3: the membership function is selected to carry out fuzzification processing on the input condition domain, the target distance, the target azimuth angle and the target entry angle data, a neural network learning model is constructed based on a fuzzy sample, and a BP algorithm is adopted to carry out learning training on the front part parameters;
s4: obtaining fuzzy output through fuzzy reasoning, optimizing the front part parameters through the S3, and learning and training the rear part parameters based on fuzzy samples by utilizing a least square algorithm in a neural network learning model;
s5: and (3) defuzzifying the fuzzy output to obtain a control signal, controlling corresponding equipment to work, judging whether the control requirement is met, ending if the control requirement is met, and returning to the step (S1) otherwise.
2. The fuzzy neural network-based equipment control autonomous decision making method of claim 1, wherein the case domain includes information assurance, advantage assurance, attack, defense and defensive and attack; wherein, the information assurance is: no target is found, or only direction finding information; the advantage is ensured as follows: the target has been found and there is direction-finding ranging information.
3. The autonomous decision making method for equipment control based on a fuzzy neural network according to claim 1, wherein the step S1 is specifically: and acquiring target data of a battlefield enemy, and judging the current situation domain of the equipment according to the target data and a preset situation domain rule.
4. The autonomous decision making method for equipment control based on a fuzzy neural network according to claim 1, wherein the equipment control rules in the equipment control rule base are relationships from observed inputs to conclusion outputs, and specifically comprise: a number of observation item condition information and a conclusion item associated with the number of condition information.
5. The fuzzy neural network based equipment control autonomous decision method of claim 4, wherein the observation term includes a case domain, a target distance, a target azimuth, and a target entry angle; the conclusion items include maneuver direction, maneuver speed, radar control, weapon control, and electronic countermeasure control.
6. The fuzzy neural network based equipment control autonomous decision method of claim 5, wherein the fuzzy logic inference model includes a fuzzification layer, a fitness layer, a normalization layer, a fuzzy output layer, and a total output layer; the input of the fuzzy logic inference model comprises a target distance=[-500,500]Target azimuth->= [-180,180]Target entry angle->= [-180,180]Case Domain->= { 1-info guarantee, 2-advantage guarantee, 3-attack, 4-defense, 5-defense and attack target }; the output includes the maneuver direction->= { 1-pull-up, 2-dive, 3-turn left, 4-turn right }, maneuver speed +.>= { 1-acceleration, 2-deceleration, 3-average speed }, radar control +.>= { 1-not open, 2-search, 3-track }, weapon control +.>= { 1-no emission, 2-emission }, electronic countermeasure control +.>= { 1-scout, 2-interference }.
7. The autonomous decision making method of equipment control based on a fuzzy neural network as claimed in claim 6, wherein in step S3, a membership function is selected to perform fuzzy processing on the input data of the case domain, the target distance, the target azimuth and the target entry angle, specifically: transforming the input variable values of the equipment control autonomous decision into linguistic values of fuzzy linguistic variables; according to the size of the input variable data, the target distance is calculatedTarget azimuth->Target(s)Entrance angle->And case Domain->Dividing into 5 fuzzy subsets: positive large, positive small, zero, negative small, negative large; wherein:
the membership function uses a gaussian function:
8. The method for autonomous decision making of equipment control based on a fuzzy neural network according to claim 7, wherein in step S3, a neural network learning model is constructed based on the fuzzy sample, and the learning training of the front part parameters is performed by using a BP algorithm, specifically comprising: based on the fuzzy sample, the BP algorithm in the neural network model is adopted to learn and train the center and width front piece parameters of the membership function, the front piece parameters are fed back to the fuzzy logic reasoning model, and the shape of the fuzzy membership function is optimized and updated.
9. The equipment control autonomous decision making method based on the fuzzy neural network as claimed in claim 8, wherein in step S4, fuzzy output is obtained through fuzzy reasoning, specifically:
wherein ,for the total number of fuzzy rules, +.>For inputting a number of numbers>For the number of outputs, ->Representing the applicability of each fuzzy rule, +.>Normalized fitness representing each fuzzy rule, < >>Fuzzy outputs representing each fuzzy rule, < >>The conclusion parameters are also called back-piece parameters for fuzzy reasoning.
10. The method for autonomous decision making of equipment control based on a fuzzy neural network according to claim 9, wherein in step S4, based on the fuzzy samples, the learning and training are performed on the back-part parameters by using a least squares algorithm in the neural network model, specifically comprising: based on the fuzzy sample, the least square algorithm in the neural network model is adopted to learn and train the back-piece parameters, and the back-piece parameters are transmitted to the fuzzy logic reasoning model to update the fuzzy reasoning parameters in an optimized mode.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310318098.4A CN116027673B (en) | 2023-03-29 | 2023-03-29 | Equipment control autonomous decision-making method based on fuzzy neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310318098.4A CN116027673B (en) | 2023-03-29 | 2023-03-29 | Equipment control autonomous decision-making method based on fuzzy neural network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116027673A true CN116027673A (en) | 2023-04-28 |
CN116027673B CN116027673B (en) | 2023-06-06 |
Family
ID=86089742
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310318098.4A Active CN116027673B (en) | 2023-03-29 | 2023-03-29 | Equipment control autonomous decision-making method based on fuzzy neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116027673B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116819964A (en) * | 2023-06-20 | 2023-09-29 | 小米汽车科技有限公司 | Model optimization method, model optimization device, electronic device, vehicle and medium |
Citations (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101025618A (en) * | 2006-12-28 | 2007-08-29 | 上海电力学院 | Power plant thermal equipment intelligent state diagnosing and analyzing system |
CN101464096A (en) * | 2008-12-11 | 2009-06-24 | 江苏金源锻造股份有限公司 | Multi-point sensing and intelligent control method for temperature of gas heating stove |
CN101650568A (en) * | 2009-09-04 | 2010-02-17 | 湖南大学 | Method for ensuring navigation safety of mobile robots in unknown environments |
CN102339055A (en) * | 2011-09-27 | 2012-02-01 | 深圳市全智聚能科技有限公司 | Traffic energy management and control method and system |
US20120179326A1 (en) * | 2011-01-06 | 2012-07-12 | Eurocopter | Predicting time to maintenance by fusion between modeling and simulation for electronic equipment on board an aircraft |
CN105136469A (en) * | 2015-07-23 | 2015-12-09 | 江苏大学 | Unmanned vehicle speed control method based on PSO and RBF neutral network |
CN106647781A (en) * | 2016-10-26 | 2017-05-10 | 广西师范大学 | Neural-fuzzy PID control method of four-rotor aircraft based on repetitive control compensation |
CN108284442A (en) * | 2017-01-24 | 2018-07-17 | 中国北方车辆研究所 | A kind of mechanical arm flexible joint control method based on fuzzy neural network |
DE102018204684A1 (en) * | 2017-04-10 | 2018-10-11 | Hitachi, Ltd. | PLANT CONTROL AND ITS CONTROL METHOD, ROLLING STREET CONTROL AND ITS CONTROL PROCEDURE AND PROGRAM |
CN110297503A (en) * | 2019-07-08 | 2019-10-01 | 中国电子科技集团公司第二十九研究所 | A kind of method of more unmanned systems collaboratively searching danger sources |
CN111026146A (en) * | 2019-12-24 | 2020-04-17 | 西北工业大学 | Attitude control method for composite wing vertical take-off and landing unmanned aerial vehicle |
CN111736461A (en) * | 2020-06-30 | 2020-10-02 | 西安电子科技大学 | Unmanned aerial vehicle task collaborative allocation method based on Q learning |
JP2021061019A (en) * | 2020-12-17 | 2021-04-15 | 株式会社日立製作所 | Plant controller, method for controlling the same, roller controller, method for controlling the same, and program |
CN113312411A (en) * | 2021-06-22 | 2021-08-27 | 中国电子科技集团公司第十四研究所 | Equipment fault cause diagnosis method based on knowledge graph and rule constraint |
CN114913095A (en) * | 2022-06-08 | 2022-08-16 | 西北工业大学 | Depth deblurring method based on domain adaptation |
-
2023
- 2023-03-29 CN CN202310318098.4A patent/CN116027673B/en active Active
Patent Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101025618A (en) * | 2006-12-28 | 2007-08-29 | 上海电力学院 | Power plant thermal equipment intelligent state diagnosing and analyzing system |
CN101464096A (en) * | 2008-12-11 | 2009-06-24 | 江苏金源锻造股份有限公司 | Multi-point sensing and intelligent control method for temperature of gas heating stove |
CN101650568A (en) * | 2009-09-04 | 2010-02-17 | 湖南大学 | Method for ensuring navigation safety of mobile robots in unknown environments |
US20120179326A1 (en) * | 2011-01-06 | 2012-07-12 | Eurocopter | Predicting time to maintenance by fusion between modeling and simulation for electronic equipment on board an aircraft |
CN102339055A (en) * | 2011-09-27 | 2012-02-01 | 深圳市全智聚能科技有限公司 | Traffic energy management and control method and system |
CN105136469A (en) * | 2015-07-23 | 2015-12-09 | 江苏大学 | Unmanned vehicle speed control method based on PSO and RBF neutral network |
CN106647781A (en) * | 2016-10-26 | 2017-05-10 | 广西师范大学 | Neural-fuzzy PID control method of four-rotor aircraft based on repetitive control compensation |
CN108284442A (en) * | 2017-01-24 | 2018-07-17 | 中国北方车辆研究所 | A kind of mechanical arm flexible joint control method based on fuzzy neural network |
DE102018204684A1 (en) * | 2017-04-10 | 2018-10-11 | Hitachi, Ltd. | PLANT CONTROL AND ITS CONTROL METHOD, ROLLING STREET CONTROL AND ITS CONTROL PROCEDURE AND PROGRAM |
CN108687137A (en) * | 2017-04-10 | 2018-10-23 | 株式会社日立制作所 | Complete equipment control device, rolling mill control apparatus, control method and storage medium |
CN110297503A (en) * | 2019-07-08 | 2019-10-01 | 中国电子科技集团公司第二十九研究所 | A kind of method of more unmanned systems collaboratively searching danger sources |
CN111026146A (en) * | 2019-12-24 | 2020-04-17 | 西北工业大学 | Attitude control method for composite wing vertical take-off and landing unmanned aerial vehicle |
CN111736461A (en) * | 2020-06-30 | 2020-10-02 | 西安电子科技大学 | Unmanned aerial vehicle task collaborative allocation method based on Q learning |
JP2021061019A (en) * | 2020-12-17 | 2021-04-15 | 株式会社日立製作所 | Plant controller, method for controlling the same, roller controller, method for controlling the same, and program |
CN113312411A (en) * | 2021-06-22 | 2021-08-27 | 中国电子科技集团公司第十四研究所 | Equipment fault cause diagnosis method based on knowledge graph and rule constraint |
CN114913095A (en) * | 2022-06-08 | 2022-08-16 | 西北工业大学 | Depth deblurring method based on domain adaptation |
Non-Patent Citations (5)
Title |
---|
崔洋培等: "基于自适应神经模糊推理系统的空中目标意图预测", 《第15届中国系统仿真技术及其应用学术会议(15TH CCSSTA2014)》, pages 277 - 281 * |
张子明等: "未来空战中的意图理解技术综述", 《第十九届中国航空测控技术年会论文集》, pages 392 - 396 * |
王钦钊等: "装甲装备作战效能评估的模糊Petri网模型研究", 《火力与指挥控制》, pages 63 - 67 * |
董小丽: "基于云模型的自适应神经模糊推理系统的装备保障态势评估方法的研究", 《中国优秀硕士学位论文全文数据库信息科技辑》, pages 140 - 106 * |
陈晓山,马曲立: "基于神经网络集成的军事装备保障体系控制研究", 《2011 3RD WORLD CONGRESS IN APPLIED COMPUTING,COMPUTER SCIENCE, AND COMPUTER ENGINEERING》, pages 260 - 265 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116819964A (en) * | 2023-06-20 | 2023-09-29 | 小米汽车科技有限公司 | Model optimization method, model optimization device, electronic device, vehicle and medium |
CN116819964B (en) * | 2023-06-20 | 2024-02-06 | 小米汽车科技有限公司 | Model optimization method, model optimization device, electronic device, vehicle and medium |
Also Published As
Publication number | Publication date |
---|---|
CN116027673B (en) | 2023-06-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Yang et al. | UAV air combat autonomous maneuver decision based on DDPG algorithm | |
CN113435644B (en) | Emergency prediction method based on deep bidirectional long-short term memory neural network | |
CN112036556B (en) | Target intention inversion method based on LSTM neural network | |
CN116027673B (en) | Equipment control autonomous decision-making method based on fuzzy neural network | |
Luo et al. | A quantum encoding bat algorithm for uninhabited combat aerial vehicle path planning | |
Пэй et al. | Mobile robot automatic navigation control algorithm based on fuzzy neural network in industrial Internet of things environment | |
Turkmen et al. | Cheap joint probabilistic data association with adaptive neuro-fuzzy inference system state filter for tracking multiple targets in cluttered environment | |
Yihong et al. | Threat assessment based on adaptive intuitionistic fuzzy neural network | |
Pekaslan et al. | Noise parameter estimation for non-singleton fuzzy logic systems | |
Wu et al. | Dynamic multitarget assignment based on deep reinforcement learning | |
Liu et al. | Evolutionary algorithm-based attack strategy with swarm robots in denied environments | |
CN115130523A (en) | Flight target behavior intention prediction method based on hidden Markov model | |
Wang et al. | Local Sensing based Multi-agent Pursuit-evasion with Deep Reinforcement Learning | |
Al-Talabi et al. | A two stage learning technique using PSO-based FLC and QFIS for the pursuit evasion differential game | |
CN108446858B (en) | Air-to-ground threat assessment method based on specific network structure | |
Wang et al. | Parameter Identification of a PN-Guided Incoming Missile Using an Improved Multiple-Model Mechanism | |
Zheng et al. | UAV maneuver decision-making via deep reinforcement learning for short-range air combat | |
Xiao et al. | A robust target intention recognition method based on dynamic bayesian network | |
Lu et al. | Strategy Generation Based on DDPG with Prioritized Experience Replay for UCAV | |
Fan et al. | A novel dynamic Bayesian network based threat assessment algorithm | |
Fang et al. | Adaptive Decision-Making in Attack-Defense Games With Bayesian Inference of Rationality Level | |
Nan et al. | A Quasi-Greedy Strategy based on Reinforcement Learning in Intelligent Control Technology | |
Hu et al. | Hierarchical fusion evaluation and optimization of radar intelligent tracking algorithm via hybrid weight design mechanism | |
Rane et al. | Analysis of mamdani and sugeno fuzzy inference system for destroying multiple target at high angle of attack using simulink | |
Chen et al. | The optimal combat maneuver decision of the autonomous underwater vehicle based on the fuzzy comprehensive evaluation 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 |