CN112288155B - Security plan generation scheduling method and system based on machine learning and collaborative filtering - Google Patents
Security plan generation scheduling method and system based on machine learning and collaborative filtering Download PDFInfo
- Publication number
- CN112288155B CN112288155B CN202011144677.4A CN202011144677A CN112288155B CN 112288155 B CN112288155 B CN 112288155B CN 202011144677 A CN202011144677 A CN 202011144677A CN 112288155 B CN112288155 B CN 112288155B
- Authority
- CN
- China
- Prior art keywords
- plan
- unmanned aerial
- equipment
- matrix
- aerial vehicle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000001914 filtration Methods 0.000 title claims abstract description 23
- 238000010801 machine learning Methods 0.000 title claims abstract description 18
- 238000012544 monitoring process Methods 0.000 claims abstract description 15
- 238000005457 optimization Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 11
- 239000011159 matrix material Substances 0.000 claims description 54
- 238000004891 communication Methods 0.000 claims description 24
- 230000002159 abnormal effect Effects 0.000 claims description 23
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 7
- 230000009471 action Effects 0.000 claims description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 5
- 230000009545 invasion Effects 0.000 claims description 4
- 238000012417 linear regression Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 230000033001 locomotion Effects 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 claims description 3
- 238000013439 planning Methods 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims description 2
- 238000010276 construction Methods 0.000 claims description 2
- 230000004927 fusion Effects 0.000 claims description 2
- 238000010223 real-time analysis Methods 0.000 claims description 2
- 238000007689 inspection Methods 0.000 abstract description 2
- 238000001514 detection method Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- DMBHHRLKUKUOEG-UHFFFAOYSA-N diphenylamine Chemical compound C=1C=CC=CC=1NC1=CC=CC=C1 DMBHHRLKUKUOEG-UHFFFAOYSA-N 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 239000013589 supplement Substances 0.000 description 2
- 208000012661 Dyskinesia Diseases 0.000 description 1
- 230000001154 acute effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000008713 feedback mechanism Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 239000012925 reference material Substances 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
- 230000007306 turnover Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Tourism & Hospitality (AREA)
- Mathematical Analysis (AREA)
- Software Systems (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
Abstract
The invention discloses a security and protection plan generation scheduling method and system based on machine learning and collaborative filtering. Wherein: unmanned aerial vehicle equipment crowd comprises a plurality of unmanned aerial vehicles, ground mobile robot equipment crowd comprises a plurality of robots, the equipment deployment platform is provided with the regional wireless module of charging in ground, the monitoring center station can show monitoring information in real time and control unmanned aerial vehicle and ground mobile robot. The method comprises real-time state inspection, task receiving, task execution (such as fault processing, intrusion processing, plan generation, plan optimization and equipment scheduling) and task completion. The invention can carry out routine and emergent security protection task execution according to the original plan, and can improve the optimization scheme or directly generate a new plan according to the actual task execution condition, thereby improving the security protection task efficiency and the rationality of the plan.
Description
Technical Field
The invention relates to application of an artificial intelligence technology in security event processing, in particular to a security scheme generation scheduling method and a security scheme generation scheduling system based on machine learning and collaborative filtering, and belongs to the field of artificial intelligence.
Background
With the continuous development of security technologies, more and more modern technologies are applied to the security field.
At present, a specific jurisdiction such as a camp has a relatively perfect emergency response system and management method for an intrusion event, and a relatively complete intrusion plan is provided for common intrusion events. However, the intelligent security is a complex process, and not only needs to have accurate identification, but also needs to give reasonable plan suggestions. The plan refers to an emergency treatment plan which is made in advance according to evaluation analysis or experience and the category and the influence degree of the potential or possible emergency events. This is crucial to camp security. How to rapidly, timely and accurately compile a corresponding plan can play a very positive role in reducing loss and improving emergency response speed.
The existing security protection plan generation mainly depends on the planning of experienced disposal personnel such as security guards, for example, the security protection personnel decides which measures to be taken specifically to dispose the event according to the received event information and the resources which can be used at present by depending on experience, or requests the corresponding resources, personnel and disposal methods to help from the surrounding area and the superior department. This method has a great drawback:
1) Because the plan is basically made by depending on the experience of the personnel, some human judgment errors often occur, the generated plan is extremely inefficient, and sometimes in order to design a reasonable plan, the security personnel need to repeatedly modify the plan, which takes much time. A small number of special events, including highly specialized events, difficult events, major accidents, sensitive events, fuzzy events, etc., are troublesome once occurring;
2) In a highly urgent situation, security personnel need to be able to make a correct decision quickly for an emergency, which requires mastering a lot of professional and background knowledge of the event, such as profession, geography, laws and regulations, emergency resource deployment, etc., but these lots of knowledge cannot be realized for emergency treatment personnel by means of memory alone.
Every time a new emergency is met, every person needs to have some reference materials, and the existing scheme system cannot achieve the purpose by taking the experience of the previous person as a reference. With a large amount of data, it is a significant risk for emergency disposal if one is always experienced. Because things are diverse, never one can say all the cases. And even if people do not have defects in knowledge, experience and intelligence, the people are feared and hesitant to bear risks when encountering emergencies, so that the people lose time and cast mistakes.
Some systems play a role of expert emergency decision, and provide alternative schemes and technical support for emergency decision through expert consultation and emergency consultation to form scientific emergency action schemes. This problem is particularly acute because emergency decisions have more stringent time requirements. The real problem is that the expert is required to be scheduled to be in place for the first time and the consultation takes much time. If the decision time is delayed in order to enhance the scientificity of the decision so as to miss the best solution opportunity, it will be irrevocable and defeat the purpose of the scientific decision.
In the past, security problems of specific districts such as camps and the like are emphasized, and with the introduction of ground robots and unmanned aerial vehicles in fixed camps, traditional manpower security inspection is gradually replaced. However, an intelligent scheme generation method and system for security protection are still lacking at present.
Disclosure of Invention
The invention aims to provide a hybrid intelligent plan method based on machine learning and collaborative filtering. The system is based on the favorable degree in the prior camp security plan knowledge base, and the traditional collaborative filtering algorithm, the matrix factorization method, the linear regression model and the XGboost model are improved to be fused, so that the hybrid intelligent plan recommendation system based on machine learning and collaborative filtering is invented, and a feedback mechanism optimization recommendation system is provided for users. The prediction components of the two models are combined and finally a recommendation list is generated according to the algorithm of TOP-N.
The technical scheme of the invention is as follows:
a security protection plan generation scheduling system based on machine learning and collaborative filtering comprises an aerial unmanned aerial vehicle equipment group, a ground mobile robot equipment group, at least one equipment deployment platform, a monitoring center station, a positioning navigation module and a communication network system.
Aerial unmanned aerial vehicle equipment crowd comprises a plurality of unmanned aerial vehicles, unmanned aerial vehicle is provided with camera and location communication module, the camera includes high definition and infrared camera, location communication module includes big dipper satellite positioning module or regional network orientation module, unmanned aerial vehicle has automatic obstacle avoidance function.
Ground mobile robot equipment crowd comprises a plurality of robots, the robot is crawler-type ground mobile robot, can remove by full topography, is provided with qxcomm technology camera, location communication module and speaker, the camera includes high definition and infrared camera, location communication module includes big dipper satellite positioning module or regional network orientation module, qxcomm technology camera has level and every single move angle rotation function, ground mobile robot has the automatic barrier function of keeping away.
The equipment deployment platform is provided with a ground area wireless charging module which can wirelessly charge the unmanned aerial vehicle and the ground mobile robot, and the unmanned aerial vehicle and the ground mobile robot can automatically charge when stopping to a specific range area; the equipment deployment platform is also provided with an automatic lifting door for the unmanned aerial vehicle and the ground mobile robot to enter and exit.
The monitoring center station comprises a comprehensive monitoring platform which can display monitoring information in real time and comprises a manual control console and other control workstations for manually controlling the unmanned aerial vehicle and the ground mobile robot; further:
the system also comprises a control subsystem of the unmanned aerial vehicle and the ground mobile robot, and the unmanned aerial vehicle and the ground mobile robot (trolley) are controlled;
the system also comprises a multi-task cooperative execution subsystem, which can realize cooperative scheduling of equipment among various tasks and meet the requirement of each equipment for carrying out the tasks;
the system also comprises a plan subsystem, wherein the plan comprises an initial plan, an emergency plan and the like, and can be used for learning the building area security service modeling and plan management according to the actual task execution condition, improving the original plan and generating a new plan; the plan subsystem can also carry out autonomous learning according to information such as a route, time, manual access and the like of an actual plan execution task, supplement and optimize an original plan or generate a new plan, and submit the new plan to manual review.
The system also comprises an image real-time analysis subsystem which can analyze the returned images in real time and alarm in time when abnormity is found.
The positioning navigation module can adopt Beidou satellite positioning and/or regional network positioning and the like.
The communication network system may employ various communication means such as wired or wireless network communication, wired or wireless asynchronous communication, and the like.
The unmanned aerial vehicle equipment group, the ground mobile robot equipment group, the equipment deployment platform, the monitoring center station and the positioning navigation module are in communication connection through the communication network system.
A security protection plan generation scheduling method based on machine learning and collaborative filtering comprises the following steps:
step 1 real-time status checking
The ground robot equipment group and the unmanned aerial vehicle equipment group which are positioned at the equipment deployment stations are in communication connection with the monitoring center station, the state information of the ground robot equipment group and the unmanned aerial vehicle equipment group is uploaded to the monitoring center in real time, the monitoring center station checks whether the state of the ground robot equipment group and the unmanned aerial vehicle equipment group is normal or not, if the state of the ground robot equipment group and the unmanned aerial vehicle equipment group is abnormal, the monitoring center station sets the state of the ground robot equipment group and the unmanned aerial vehicle equipment group to be abnormal and informs monitoring personnel of the information.
Step 2 task reception
The "abnormity" judged by the monitoring center can be divided into two types: equipment failure and intrusion discovery. If the equipment is abnormal in fault, the maintainer receives the task and carries out equipment maintenance work; if the target is abnormal caused by the invasion of the foreign target, the plan subsystem receives the abnormal information and carries out corresponding plan generation work.
Step 3. Task execution
Step 3.1 failure handling
In the process of executing tasks by the task groups, if sudden conditions such as loss of connection of equipment among the task groups, abnormal movement, failure of planning routes and the like occur, information such as the current position, abnormal conditions, images and the like is sent to the monitoring center station, and the monitoring center station personnel judge processing modes such as task continuation, task stop, manual operation, abnormal equipment stop, automatic calling of idle equipment supplement and the like.
Step 3.2 intrusion handling
When the monitoring central station finds an illegal target in a task group and information transmission, the monitoring central station can transmit video, audio and other information transmitted by the task group into the plan subsystem, and an algorithm in the subsystem utilizes the existing information to generate a plan and optimize the plan.
Step 3.2.1 plan Generation
1) Mixed collaborative filtration
Firstly, a plan scoring database based on users and a plan is built, according to the feature scores of the users and the plan, a user feature matrix similar among the users and a plan feature matrix similar among the plans are built respectively, and the scoring degree represents the preference degree. And respectively taking the corresponding matrixes to perform matrix operation, selecting the nearest set of the user and the plan, predicting the scoring of the user on the existing plan by the traditional recommendation algorithm according to the nearest set, and combining the predicted scoring and the real scoring to form a new training set. And inputting a new training set as a model of linear regression, predicting unknown scores according to the trained good model, and generating a recommendation list by adopting a TOP-N algorithm.
A user characteristic matrix:
a pre-arranged pattern characteristic matrix:
2) Matrix decomposition
The matrix decomposition is used for extracting features, and the user feature matrix and the plan feature matrix can be decomposed into a user matrix and a plan matrix by decomposing the scoring matrix.
For the user feature matrix, each row may be a user vector and each column may be an invisible vector of users; for the pre-pattern feature matrix, each row may be a pre-pattern vector and each column may be a stealth vector of the pre-pattern.
Each column of the user's feature matrix and a column of the plan feature matrix, their inner product sum representing the user's preference for the plan. By doing so, not only can the spatial complexity be reduced, but also the invisible features of the user or the plan can be found, so that the prediction scores can be well interpreted. As shown in formula (1):
R m*n ≈U m*k *I k*n formula (1)
Wherein R represents a prediction scoring matrix, m represents the number of users, n represents the number of plans, U is a user characteristic matrix, I is a plan characteristic matrix, and k represents the kth row of the matrix;
predicting the score of the user U on the plan I by using the user characteristic matrix U and the plan characteristic matrix I; solving the matrix L (P, Q) by minimizing the sum of the squared errors, the model construction of L (P, Q) is represented as follows:
wherein r is u，i Is true value, r' u，i Is a predicted value, k isA hidden feature number.
Step 3.2.2 plan optimization
And retrieving typical cases similar to the current case from the case base according to the at least one group of selected emergency plans, and correcting the at least one group of emergency plans to generate an emergency action scheme.
If there are two or more groups of emergency plans, different departments and different schemes need to be coordinated and combined to form a fusion optimized emergency action scheme.
In the invention, all the emergency plans have expected corresponding event hypotheses, if the current event is the same as the event hypotheses in the emergency plans, the departments and schemes scheduled in the emergency plans can be directly adopted, because the invasion events stored in the case library are various and the corresponding event hypotheses are also various, it is unrealistic to independently generate an emergency plan for each event, and only one emergency plan can be formulated for a certain class of similar events, so under most conditions, the event hypotheses corresponding to the current event and the emergency plans are not completely consistent, and at the moment, the plan serving as the scheme is not in accordance with the actual condition, and the plan is required to be corrected according to the specific situation of the current invasion event, thereby generating a feasible scheme.
Step 3.2.3 device scheduling
The multitask coordination subsystem schedules the corresponding types and the number of the devices in the plan, and the unmanned aerial vehicle and the ground mobile robot control subsystem generate a motion scheme for each device according to the scheme.
Step 4. Task completion
And according to the actual situation, arranging the equipment to a nearest equipment deployment platform for charging, self-checking and information reporting.
A computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the security scheme generation scheduling method based on machine learning and collaborative filtering. The computer readable storage medium is disposed within a comprehensive monitoring platform of the monitoring central station.
The invention has the advantages of
(1) The invention can carry out routine and emergent security protection task execution according to the original plan, and can improve the optimization scheme or directly generate a new plan according to the actual task execution condition, thereby improving the security protection task efficiency and the rationality of the plan.
(2) The invention discloses a multi-task cooperative execution mode, and solves the problems that the original security intrusion detection unmanned aerial vehicle and a ground robot only execute a single task, the equipment utilization rate is low, and global unified scheduling cannot be realized.
(3) The invention enables the security intrusion detection task to be automatically and intelligently executed, greatly improves the working efficiency, liberates manpower and is particularly suitable for security intrusion detection tasks in large areas.
(4) Unmanned aerial vehicle adds ground mobile robot's detection mode, can satisfy the ground-air integration, and the advantage of full play unmanned aerial vehicle and ground robot satisfies various scenes that detect, improves the mode of many cameras of multi-device of security protection, can improve intrusion detection's efficiency greatly.
Drawings
Fig. 1 is a schematic diagram of a system configuration in embodiment 1 of the present invention.
Fig. 2 is a schematic diagram of a monitoring center station according to embodiment 1 of the present invention.
FIG. 3 is a flowchart of the entire method of embodiment 1 of the present invention.
FIG. 4 is a flow chart of the hybrid collaborative filtering system according to example 1 of the present invention.
Detailed Description
The invention is further illustrated by the following figures and examples.
Example 1
And (3) intrusion condition: in rainy days, suspicious people turn over the wall and enter the camp; the number of intruders: 3 persons
As shown in fig. 1, the system for processing intrusion into a fixed camp based on an intelligent plan provided by the present invention includes an aerial unmanned aerial vehicle device group, a ground mobile robot device group, a device deployment station, a monitoring center station, a beidou satellite navigation system and a communication network system, wherein the aerial unmanned aerial vehicle device group and the ground mobile robot device group form a corresponding task execution group according to scheduling when executing tasks.
Step 1 real-time status checking
Each device transmits self information and captured camp information to a monitoring central station according to preset frequency, and if the device is not connected or an abnormal target appears, the system sets the state as abnormal and informs monitoring personnel of the information.
Step 2. Task reception
And after the monitoring center station worker finds the abnormal condition, if the abnormal condition is caused by machine failure and the like, the maintainer receives the failure processing task and carries out maintenance detection on corresponding equipment, and if the abnormal condition is caused by external target intrusion, the plan subsystem receives abnormal information and processes the abnormal information.
Step 3. Task execution
Step 3.1 failure handling
In the process of daily patrol or task execution, emergent conditions such as network transmission abnormity, equipment running route abnormity and the like exist, when monitoring center station workers receive corresponding abnormal information, the conditions are further judged to be misjudgment or real abnormity, if the conditions are misjudgment, the state is changed into normal and the task is continued, and if the conditions are real abnormity, the task needs to be stopped and specific personnel needs to be arranged for processing.
Step 3.2.1 plan Generation
Inputting: the method comprises the following steps of preprocessing a user characteristic matrix UF, a plan characteristic moment IF, a user set U, a plan set I, a user plan scoring matrix R, a nearest neighbor number K and an invisible characteristic number K.
And (3) outputting: and (5) a training set is produced.
1) Respectively initializing a user matrix U and a plan matrix I according to the user characteristic matrix UF and the plan characteristic matrix IF
2) Finding a set of user scores in R
3) Based on matrix decomposition, calculating the prediction score r' (u, i) of the user u to the plan set i according to a formula of a matrix decomposition technology
4) And (4) continuously calculating the deviation between the two matrixes and the real value according to formulas (1) to (3) in the summary of the invention, and circularly executing calculation to calculate the difference between the real value and the prediction score in all the plans evaluated by the user u.
5) And circularly executing the steps, recording the scored plan of the user based on collaborative filtering and matrix prediction as X1, X2 and X3, recording the real value of the scored plan as y, combining the y and the y into a new training data set for linear regression and XGboost model training, finally scoring the unscored plan by using a final generation model, and calculating a recommendation list by using a TOPN algorithm.
Step 3.2.2 plan optimization
Once an emergency intrusion event occurs, an effective plan must be quickly and accurately retrieved, but since the actual situation is complicated and variable and cannot be completely the same as the assumption of the plan, most cases need to modify the plan retrieved and processed from the case base, and the reference basis for the modification is the existing experience, the past case, the field situation, and the like. On the basis of the selected emergency plan, the plan can be modified by:
1) Enhancing the data processing capacity through the model, and analyzing the manpower and material resources and the like scheduled for the current event;
2) Providing necessary common knowledge for the current intrusion event by utilizing a knowledge base so as to make up for the deficiency of a candidate plan;
3) The man-machine interaction can expand the information integration capability, is convenient for a decision maker to recognize and apply decision information, and simultaneously introduces the only action of the man-machine interaction.
Step 3.2.3 device scheduling
The multitask coordination subsystem schedules equipment of corresponding types and quantity in the plan, and if the scheduling is successful, the multitask coordination subsystem moves according to a route planned by the unmanned aerial vehicle and the ground mobile robot control subsystem; and if the scheduling fails, returning error information to wait for manual modification of the scheme.
Step 4. Task completion
After the corresponding task is finished, corresponding processing is carried out according to the corresponding condition of the following table, the relevant information of the processing process is reported, and finally whether the plan is stored in the existing plan library or not is determined manually.
Claims (9)
1. A security protection plan generation scheduling method based on machine learning and collaborative filtering is characterized by comprising the following steps:
step 1, real-time status checking
The ground robot equipment group and the unmanned aerial vehicle equipment group positioned at each equipment deployment platform are in communication connection with the monitoring center station, the self state information is uploaded to the monitoring center in real time, the monitoring center station checks whether the state is normal or not, if the abnormal condition occurs, the monitoring center station sets the state to be abnormal and informs monitoring personnel of the information;
step 2, task reception
The "abnormity" judged by the monitoring center can be divided into two types: equipment failure and intrusion discovery; if the equipment is abnormal in fault, the maintainer receives the task and carries out equipment maintenance work; if the target is abnormal caused by the invasion of the foreign target, the abnormal information is received by the plan subsystem, and corresponding plan generation work is carried out;
step 3, task execution
Step 3.1, fault handling
In the process of executing tasks by the task groups, if the devices among the task groups lose connection, and the conditions of abnormal motion and sudden failure of planning a route are met, sending current position, abnormal conditions and image information to a monitoring center station, and judging the continuation of the tasks, stopping the tasks, manually operating, stopping abnormal devices and automatically calling a supplementary processing mode of idle devices by personnel at the monitoring center station;
step 3.2, intrusion handling
When the monitoring central station finds an illegal target in a task group and information transmission, the video and audio information transmitted by the task group can be transmitted to the plan subsystem, and an algorithm in the subsystem utilizes the existing information to generate a plan and optimizes the plan;
step 3.2.1, generation of a plan
1) Hybrid collaborative filtration
Firstly, constructing a plan scoring database based on users and plans, and respectively constructing a user characteristic matrix similar among the users and a plan characteristic matrix similar among the plans according to the characteristic scores of the users and the plans, wherein the scoring height represents the preference degree; respectively taking corresponding matrixes to perform matrix operation, selecting a nearest set of the user and the plan, predicting the user to score the existing plan according to the nearest set through a traditional recommendation algorithm, and combining the predicted score and the real score to form a new training set; inputting a new training set as a model of linear regression, predicting unknown scores according to the trained model, and generating a recommendation list by adopting a TOP-N algorithm;
2) Matrix decomposition
The matrix decomposition is used for extracting features, and the user feature matrix and the plan feature matrix can be decomposed into a user matrix and a plan matrix by decomposing the scoring matrix;
each column of the user's feature matrix and a certain column of the plan feature matrix, their inner product sum representing the user's preference for the plan; as shown in equation (1):
R m*n ≈U m*k *I k*n formula (1)
Wherein R represents a prediction scoring matrix, m represents the number of users, n represents the number of plans, U is a user characteristic matrix, I is a plan characteristic matrix, and k represents the kth row of the matrix;
predicting the score of the user U on the plan I by using the user characteristic matrix U and the plan characteristic matrix I; solving the matrix L (P, Q) by minimizing the sum of the squared errors, the model construction of L (P, Q) is represented as follows:
wherein r is u，i Is true value, r' u，i Is a predicted value, k is a hidden feature number;
step 3.2.2, plan optimization
Retrieving typical cases similar to the current situation from a case library according to the at least one group of selected emergency plans, and correcting the at least one group of emergency plans to generate an emergency action scheme;
step 3.2.3, device scheduling
The multitask coordination subsystem schedules the equipment with corresponding types and quantity in the plan, and the unmanned aerial vehicle and ground mobile robot control subsystem generates a motion scheme for each equipment according to the scheme;
step 4, task completion
And according to the actual situation, arranging the equipment to a nearest equipment deployment platform for charging, self-checking and information reporting.
2. The machine learning and collaborative filtering based security scheme generation scheduling method of claim 1, wherein: the user feature matrix in the generation of the plan is expressed as:
step 3.2.1, the plan feature matrix in plan generation is represented as:
3. the security scheme generation and scheduling method based on machine learning and collaborative filtering according to claim 1, wherein:
in the plan optimization, if two or more groups of emergency plans exist, different departments and different schemes need to be coordinated and combined to form an emergency action scheme for fusion optimization.
4. The security scheme generation and scheduling method based on machine learning and collaborative filtering according to claim 1, wherein:
in the plan optimization, all emergency plans have corresponding expected event hypotheses, and if the current event is the same as the event hypotheses in the emergency plans, departments and schemes scheduled in the emergency plans can be directly adopted.
5. The security scheme generation and scheduling method based on machine learning and collaborative filtering according to any one of claims 1 to 4, wherein:
in the plan optimization, when the current event and the event assumption corresponding to the emergency plan are not completely consistent, the plan needs to be corrected according to the specific situation of the current intrusion event, so as to generate a feasible scheme.
6. A computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the machine learning and collaborative filtering based security scheme generation scheduling method of any of claims 1 to 5.
7. The utility model provides a security protection scheme generates dispatch system based on machine learning and collaborative filtering which characterized in that:
the system comprises an aerial unmanned aerial vehicle equipment group, a ground mobile robot equipment group, at least one equipment deployment platform, a monitoring center station, a positioning navigation module and a communication network system;
the aerial unmanned aerial vehicle equipment group consists of a plurality of unmanned aerial vehicles, each unmanned aerial vehicle is provided with a camera and a positioning communication module, each camera comprises a high-definition camera and an infrared camera, each positioning communication module comprises a Beidou satellite positioning module or a regional network positioning module, and each unmanned aerial vehicle has an automatic obstacle avoidance function;
the ground mobile robot equipment group consists of a plurality of robots, the robots are crawler-type ground mobile robots, can move in all terrain, and are provided with omnidirectional cameras, positioning communication modules and loudspeakers, the cameras comprise high-definition cameras and infrared cameras, the positioning communication modules comprise Beidou satellite positioning modules or area network positioning modules, the omnidirectional cameras have horizontal and pitching angle rotation functions, and the ground mobile robots have automatic obstacle avoidance functions;
the equipment deployment platform is provided with a ground area wireless charging module which can wirelessly charge the unmanned aerial vehicle and the ground mobile robot, and the unmanned aerial vehicle and the ground mobile robot can automatically charge when stopping to a specific range area; the equipment deployment platform is also provided with an automatic lifting door for the unmanned aerial vehicle and the ground mobile robot to enter and exit;
the monitoring center station comprises a comprehensive monitoring platform which can display monitoring information in real time and comprises a manual control console and other control workstations for manually controlling the unmanned aerial vehicle and the ground mobile robot; further comprising the computer-readable storage medium of claim 6, the computer-readable storage medium disposed within the integrated monitoring platform;
the positioning navigation module can adopt Beidou satellite positioning and/or regional network positioning;
the communication network system adopts wired or wireless network communication and wired or wireless asynchronous communication modes;
the unmanned aerial vehicle equipment group, the ground mobile robot equipment group, the equipment deployment platform, the monitoring center station and the positioning navigation module are in communication connection through the communication network system.
8. The machine learning and collaborative filtering based security scheme generation scheduling system of claim 7, wherein:
the system also comprises an image real-time analysis subsystem which can analyze the returned images in real time and alarm in time when abnormity is found.
9. The machine learning and collaborative filtering based security scheme generation scheduling system of claim 7, wherein:
the system also comprises a control subsystem of the unmanned aerial vehicle and the ground mobile robot, and the unmanned aerial vehicle and the ground mobile robot are controlled.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011144677.4A CN112288155B (en) | 2020-10-23 | 2020-10-23 | Security plan generation scheduling method and system based on machine learning and collaborative filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011144677.4A CN112288155B (en) | 2020-10-23 | 2020-10-23 | Security plan generation scheduling method and system based on machine learning and collaborative filtering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112288155A CN112288155A (en) | 2021-01-29 |
| CN112288155B true CN112288155B (en) | 2022-12-09 |
Family
ID=74425040
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011144677.4A Active CN112288155B (en) | 2020-10-23 | 2020-10-23 | Security plan generation scheduling method and system based on machine learning and collaborative filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112288155B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113382413B (en) * | 2021-06-07 | 2022-09-27 | 西安电子科技大学 | Intelligent unmanned equipment networking anomaly detection method and system based on machine learning |
| CN113612976B (en) * | 2021-10-11 | 2021-12-17 | 成都派沃特科技股份有限公司 | Security plan optimization method and security plan optimization device |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102013083A (en) * | 2010-12-01 | 2011-04-13 | 深圳市天维尔通讯技术有限公司 | Method and system for generating emergency action plan based on pre-arranged plan |
| CN102915484A (en) * | 2012-10-12 | 2013-02-06 | 重庆亚德科技股份有限公司 | Intelligent predetermined plan system based on collaborative filtering |
| CN106339829A (en) * | 2016-11-10 | 2017-01-18 | 国网山东省电力公司济南供电公司 | Big data, Cloud, IoT and mobile internet technologies based active maintenance panorama monitoring system of power distribution network |
| CN108629719A (en) * | 2017-03-17 | 2018-10-09 | 南京聚特机器人技术有限公司 | Public safety emergency command and control system based on multi-robot Cooperation and method |
| CN110647145A (en) * | 2019-09-05 | 2020-01-03 | 新疆大学 | Ground mobile robot and unmanned aerial vehicle cooperative operation system and method based on security |
| CN110838157A (en) * | 2019-10-10 | 2020-02-25 | 青岛海信网络科技股份有限公司 | Method and device for generating thematic map of emergency burst scene |
| CN111178778A (en) * | 2020-01-02 | 2020-05-19 | 中冶赛迪重庆信息技术有限公司 | Security activity scheme generation method and system based on machine learning and security activity management system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11445152B2 (en) * | 2018-08-09 | 2022-09-13 | Cobalt Robotics Inc. | Security automation in a mobile robot |
| CN111784191A (en) * | 2020-07-17 | 2020-10-16 | 塔盾信息技术(上海)有限公司 | Safety management aid decision-making system based on industrial intelligence |
-
2020
- 2020-10-23 CN CN202011144677.4A patent/CN112288155B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102013083A (en) * | 2010-12-01 | 2011-04-13 | 深圳市天维尔通讯技术有限公司 | Method and system for generating emergency action plan based on pre-arranged plan |
| CN102915484A (en) * | 2012-10-12 | 2013-02-06 | 重庆亚德科技股份有限公司 | Intelligent predetermined plan system based on collaborative filtering |
| CN106339829A (en) * | 2016-11-10 | 2017-01-18 | 国网山东省电力公司济南供电公司 | Big data, Cloud, IoT and mobile internet technologies based active maintenance panorama monitoring system of power distribution network |
| CN108629719A (en) * | 2017-03-17 | 2018-10-09 | 南京聚特机器人技术有限公司 | Public safety emergency command and control system based on multi-robot Cooperation and method |
| CN110647145A (en) * | 2019-09-05 | 2020-01-03 | 新疆大学 | Ground mobile robot and unmanned aerial vehicle cooperative operation system and method based on security |
| CN110838157A (en) * | 2019-10-10 | 2020-02-25 | 青岛海信网络科技股份有限公司 | Method and device for generating thematic map of emergency burst scene |
| CN111178778A (en) * | 2020-01-02 | 2020-05-19 | 中冶赛迪重庆信息技术有限公司 | Security activity scheme generation method and system based on machine learning and security activity management system |
Non-Patent Citations (3)
| Title |
|---|
| Analysis of Machine Learning Methods for Wildfire Security Monitoring with an Unmanned Aerial Vehicles;Dmitriy Alexandrov,etc;《 2019 24th Conference of Open Innovations Association (FRUCT)》;20190513;正文第1-7页 * |
| 基于5G的空地一体化智慧新安防应用;范天伟 等;《信息通信技术》;20191231(第S1期);第40-44页 * |
| 基于矩阵分解协同过滤算法的评分预测;刘佳;《科技创新与应用》;20160229(第5期);第86页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112288155A (en) | 2021-01-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108319234B (en) | Security management system, method, device, storage medium and information processing cloud platform | |
| CN111582697B (en) | Power distribution network fault assessment and scheduling method and system | |
| CN108762224A (en) | A kind of wisdom power plant comprehensive monitoring management system | |
| CN108833831A (en) | A kind of power construction intelligent safety monitor system | |
| CN112288155B (en) | Security plan generation scheduling method and system based on machine learning and collaborative filtering | |
| CN107134851A (en) | A kind of intelligent O＆M method and system of the photovoltaic plant based on BIM | |
| CN113706355A (en) | Method for building intelligent emergency system of chemical industry park | |
| CN112907389A (en) | Land, air and space integrated intelligent construction site system and management method | |
| CN105812752A (en) | Monitoring center applied to intelligent environmental sanitation system | |
| CN110883772B (en) | Method and system for processing potential safety hazard of railway station by using robot | |
| CN111882474B (en) | FDS function design method for automatic driving vehicle cluster scheduling | |
| CN116405101B (en) | Monitoring, controlling and disposing system and method for low-altitude digital resource and digital infrastructure | |
| CN112488898B (en) | Emergency handling system and method based on intelligent building management and control | |
| CN115278542A (en) | Ultra-wideband technology-based construction site personnel positioning management system and method | |
| CN111753780A (en) | Transformer substation violation detection system and violation detection method | |
| Indri et al. | Sen3Bot Net: A meta-sensors network to enable smart factories implementation | |
| Sha et al. | Smart city public safety intelligent early warning and detection | |
| CN112650335A (en) | Intelligent energy consumption management and control system and method based on big data mining | |
| CN115574945B (en) | Control system and method for inspection robot of submerged arc furnace | |
| CN111553264A (en) | Campus non-safety behavior detection and early warning method suitable for primary and secondary school students | |
| CN116523492A (en) | Hydropower station supervision method and system, electronic equipment and storage medium | |
| Bitton et al. | Hydra: A framework and algorithms for mixed-initiative UAV-assisted search and rescue | |
| CN114529438A (en) | Rescue site risk assessment and resource management system | |
| CN113867401A (en) | Multi-unmanned aerial vehicle patrol scheduling method based on optical fiber disturbance monitoring system | |
| CN108399389B (en) | Multi-machine monitoring system and method for machine vision, client, server and storage medium |
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 |