CN113627081A - Unmanned intelligent cluster confrontation control method based on bionic eagle hunting behavior - Google Patents
Unmanned intelligent cluster confrontation control method based on bionic eagle hunting behaviorDownload PDFInfo
- Publication number
- CN113627081A CN113627081ACN202110883706.7ACN202110883706ACN113627081ACN 113627081 ACN113627081 ACN 113627081ACN 202110883706 ACN202110883706 ACN 202110883706ACN 113627081 ACN113627081 ACN 113627081A
- Authority
- CN
- China
- Prior art keywords
- enemy
- cluster
- intelligent
- control method
- bionic
- 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
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computer Hardware Design (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Geometry (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
Technical Field
The invention relates to an unmanned intelligent cluster countermeasure control method, in particular to a cluster countermeasure control method of a simple intelligent agent with high maneuverability and low durability.
Background
Currently popular agent clustering strategies can be mainly classified into three categories: the system comprises a traditional algorithm based on a physical-mathematical model, a bionic algorithm simulating a biological cluster strategy and an enhanced learning algorithm based on a strategy reward and punishment mechanism. Among the three, the traditional algorithm has too many restrictive assumptions, a small application range and too rigid strategies, and the reinforcement learning algorithm has poor interpretability and large application hidden danger. Therefore, a bionic algorithm for reducing both advantages and disadvantages is a suitable research direction at present.
In many biological bionic possibilities, eagle clusters better meet the requirements of mass winning and decentralization of unmanned cluster confrontation. In particular, eagle groups often capture prey similar to their own size by team in actual hunting. Meanwhile, the cooperation of the team does not depend on the tedious communication among individuals, and more, the cooperation of the team is constructed by independent analysis of the situation of the individuals under the unified operational policy. These features make hunting behaviors of eagle groups provide better templates for their corresponding bionic control methods.
At present, related research based on eagle swarm hunting behavior bionic has achieved good results in optimizing the problem, but has not yet been fully applied in the aspect of unmanned swarm confrontation control. The invention discloses a mathematic model aiming at the unmanned cluster confrontation behavior for a mathematic model based on a Harris eagle optimization algorithm, and is reconstructed from the actual characteristics of the confrontation problem, so that an unmanned intelligent cluster confrontation control method based on eagle hunting behavior bionics is provided.
Disclosure of Invention
The invention provides an unmanned intelligent cluster confrontation control method based on eagle hunting behavior bionics and an implementation method thereof, aiming at solving the cluster confrontation problem of low-ammunition amount and low-armored intelligent bodies. The control method can be applied to military occasions such as timely defense in a site, enemy target destruction and the like, and is particularly suitable for the situation without obvious attack stage division (namely two groups of formation are separated by longer unit time). The control method can make the intelligent cluster form effective enclosure and extinguishment for the enemy; corresponding effective striking can be performed on scattered individuals around the enclosure.
The bionic unmanned intelligent cluster confrontation control method based on eagle hunting behaviors is a bionic confrontation control method without role differentiation but with strategic stage differentiation being emphasized. Each agent in the intelligent cluster can independently analyze the current battlefield situation and determine real-time countermeasure action according to the low-frequency timing global battlefield information of satellite communication and the real-time perceived battlefield information in the visual field; the strategic stage differentiation comprises three stages of wandering enemy seeking, soft surrounding enemy disturbing and hard surrounding enemy fighting, and the individual in the cluster analyzes and independently switches action stages according to the information sensed by the individual.
In terms of actual confrontation background, the wandering foraging enemy corresponds to the beginning stage of confrontation, and the cluster of the local parties needs to keep approximate positions of enemies depicted by the loose matrix to the timing information to increase the probability of meeting enemies, contact enemy troops quickly and keep the contact positions far away from the rear of the guardian; the soft surrounding enemy disturbing stage corresponds to a stage of forming a surrounding rudiment after meeting enemies, and the intelligent cluster of our party improves the target density of the enemy in the surrounding ring through small-scale disturbing blocking and disturbing the action of the enemy, so that the hitting effect is optimized; the hard surrounding fighter is corresponding to the final surrounding action of the team, and the intelligent agent of the team rapidly folds the surrounding ring and rapidly strikes and fights the target of the enemy.
The invention is characterized in that:
(1) the bionic countermeasure control method has lower armored and ammunition amount level for the intelligent body, and better reduces the countermeasure cost;
(2) the bionic confrontation control method has the advantages that the design of combining the soft enclosure and the hard enclosure limits the action of an enemy, and simultaneously, the use efficiency of intelligent ammunition is improved;
(3) on the premise of ensuring the threat of the cluster to the enemy, the behavior of the intelligent agent reserves higher randomness to the individual action, better disturbs the countermeasure decision of the enemy and improves the grasp of the cluster to the countermeasure initiative.
(4) The behavior of the intelligent agent does not need high-frequency data interactive sharing and does not depend on the existence or non-existence of other intelligent agents, and the decentralization degree is strong;
(5) the behavior of the intelligent agent follows the same control logic, the behavior between the intelligent agents has certain cooperativeness, and the overall strategy with certain interpretability is presented.
Drawings
FIG. 1 is a flow chart of an unmanned intelligent cluster confrontation control method based on eagle swarm hunting behavior bionics
Detailed Description
The eagle hunting behavior bionic unmanned intelligent cluster confrontation control method and the implementation method thereof provided by the invention are described in detail below with reference to the accompanying drawings.
The invention provides an unmanned intelligent cluster confrontation control method based on eagle swarm hunting behavior bionics. The flow of the bionic unmanned cluster confrontation control method is shown in figure 1. In the initial wandering and enemy seeking stage, the cluster selects any 5 enemy targets according to the current stored satellite information, and anchors the central positions of the enemy targets as the current wandering direction; meanwhile, the distance between the intelligent agent and other intelligent agents in the visual field is gradually increased to increase the probability of meeting enemies and better develop multi-directional impact disturbance at the next stage. When the intelligent agent meets enemies, the intelligent agent is switched to a soft surrounding enemy disturbing stage, the intelligent agent continuously moves towards the center of an enemy target in the visual field, and the intelligent agent keeps a distance with other intelligent agents continuously; at this time, the intelligence can randomly take a fast bump as a nuisance with a small probability. When the enemy density in the field of view of the intelligent agent is large enough or the enemy serving time is too long, the intelligent agent enters a hard surrounding enemy fighting mode, does not keep a distance any longer, but approaches to each other, rapidly approaches to the enemy target, and strikes any target in an attack range.
The following describes in detail the mathematical models of the three stages involved in the present bionic unmanned collective confrontation control method.
1. Stage of wandering to find enemy
The bionic unmanned cluster confrontation control method is characterized in that the cluster wandering enemy-seeking stage mainly corresponds to the initial stage of confrontation starting and the clearance of multiple confrontations. At this stage, the my intelligent agent still has a long distance from the enemy target, and the position of the enemy target can be roughly estimated only by timing information provided by a communication satellite and other systems, so that the main targets of the enemy can be divided into two types in order to more effectively contact the enemy target: the intelligent clusters remain dispersed within the clusters, located approximately close to the enemy target to increase the probability of contact.
According to the two action purposes, the control method obtains two corresponding action vectors uto_ene(i, t) and ufr_gar(i, t) are action vectors of the approaching of the agent i to the enemy cluster and the cluster dispersion at the time t respectively. The unit action vectors obtained by normalizing the two are respectively recorded as
And
the calculation method is as follows:
in the formula, Dg(i, t) is the set of i agents in the i-field of view of the agent at time t, GeFor the set of data adversary targets in the satellite's latest interaction, rand (G,5) is an operation of randomly picking up a maximum of 5 objects (without repetition) in set G, Xe(e, t) is a two-dimensional coordinate of the enemy target e at the time t; xg(g, t) is the two-dimensional coordinates of my agent g at time t.
The target priority of the approaching enemy is higher than the cluster spread, and the distance (R) represented by two motion vectorstoAnd Rfr) The proportion of the two factors is influenced, and the final strategic action vector of the agent i at the time t can be balanced through the characteristic. The calculation method is as follows:
Rto＝‖uto_ene(i,t)‖,Pfr＝‖uto_ene(i,t)‖
if my agent does not already exist in the field of view, the strategic action vector is directly equivalent to the action vector of the proximate enemy group, and the calculation method is as follows:
the moving speed of the intelligent agent at the stage of my party is adjusted in real time according to the distance between the intelligent agent and the target of the enemy, and the speed is slower when the intelligent agent is closer to the enemy. The calculation formula is as follows:
in the formula, v0(i) Is the maximum stable speed, R, of my agent iver(i) Is the radius of the field of view of agent i.
2. Stage of soft surrounding enemy disturbance
The soft surrounding enemy disturbing stage corresponds to a stage of forming a surrounding rudiment after meeting enemies, and in the stage, the main target of the intelligent agent of our party is to improve the target density of the enemy in the surrounding ring through small-scale disturbing blocking and disturbing the action of the enemy, so that the striking effect is optimized.
In order to meet the threat to enemy which the scrambling behavior itself must have and ensure the ammunition amount of the intelligent agent in the next stage, the intelligent agent needs to control the scrambling behavior in a smaller scale while ensuring the existence of the scrambling behavior. To meet this, the control method at this stage introduces two random parameters, namely a spoiling parameter pc(i, t) and a predation parameter pa(i, t) and their corresponding parameter indicators, i.e. predator indicators PcAnd enemy index Pa. In order to ensure the threat, the enemy hit index should be larger; in order to ensure small scale, the enemy disturbing index should be smaller.
In the countermeasure at this stage, the intelligent agent of our party continuously generates the parameter p of the offending enemy at randomc(i, t). If p isc(i, t) does not exceed the disturbance hit index PcThe intelligent agent is in a random walk state; otherwise, the intelligent body enters an enemy disturbing impact state, and the Lima quickly impacts the target center of the enemy party once. In the rapid impact process, the intelligent agent continuously and randomly generates an attack enemy parameter pa(i, t) once pa(i, t) exceeding the predator index PaThe agent hits any target within the attack range immediately.
In the formula, De(i, t) is a set of enemy targets in i-field of view of the agent at time t
Under the condition of an impact of an interference enemy, an intelligent agent needs to select a strategic action vector before interference attack, and keeps the vector in the process of one interference attack at the maximum stable speed v0(i) And (6) moving. The calculation method of the strategic action vector comprises the following steps:
in the formula (I), the compound is shown in the specification,
is a random unit vector. .
The intelligent agent needs to judge whether the quick impact is finished or not when the intelligent agent disturbs the impact. The termination condition is calculated as follows:
wherein R (e, i, t) is the distance between the intelligent agent i and the enemy target e at the moment t, and RverIs the radius of the field of view of agent i.
The judgment can ensure that the intelligent agent escapes from the attack range of the enemy object as far as possible, and the intelligent agent can have the same enemy cluster in the later view as soon as possible to maintain the soft surrounding state. Once entering a soft surrounding enemy disturbing stage, the intelligent agent starts timing and concerns the number of enemy targets in the visual field so as to judge whether entering a hard surrounding enemy destroying stage.
3. Stage of hard enclosure to destroy enemy
The stage of the hard fighter plane corresponds to the final fighter plane action of the bionic unmanned cluster. Once the number of enemy targets in the visual field reaches the hard-surrounding index NcOr the soft surrounding time exceeds the soft surrounding time threshold TcThe agent enters this stage. At this stage, the main targets of the intelligent agents of our party are gathered together to improve the survival probability of the whole cluster, and the targets of the enemy are rapidly attacked and killed.
At this stage, my agent should keep maximum steady speed action, and its strategic action vector is calculated as follows:
Rto＝||uto_ene(i，t)||，Rfr＝||uto_ene(i，t)||
in the aspect of striking strength, the intelligent agent realizes the survival difficulty in the stage and can strike multiple objects quickly at the same time so as to better reflect the advantages of the intensive distribution of enemies on the striking of the intelligent agent.
Once the density of the enemy is low, the intelligent body enters the stage of wandering and foraging the enemy again; but because the battlefield is nearby, the stage of soft surrounding enemy can be entered quickly to keep the continuity of battle.
Claims (3)
1. Unmanned intelligent cluster confrontation control method based on bionic eagle swarm hunting behaviors, which is characterized in that: the bionic unmanned intelligent cluster countermeasure control method enables each intelligent individual in the cluster to periodically broadcast the latest global battlefield information with lower broadcasting frequency (20 minutes as a reference unit) to the intelligent battle cluster through the communication satellite system, and respectively make real-time countermeasure decisions according to a unified countermeasure behavior mathematical model; the bionic unmanned intelligent cluster countermeasure control method comprises three countermeasure stages of wandering enemy seeking, soft surrounding enemy disturbing and strong surrounding enemy fighting.
2. The intelligent combat cluster of claim 1, wherein: except basic real-time local environment perception capability and maneuvering performance advantages of relative enemies, subordinate individuals of the intelligent battle cluster have no necessary limitation on other performance aspects; the basic algorithm logic of the eagle swarm bionic control method is embedded in the subordinate individuals of the intelligent combat swarm.
3. The basic algorithm logic of the eagle swarm bionic control method according to claim 2 is characterized in that: the basic algorithm logic is organically composed of three confrontation stages of walking enemy seeking, soft surrounding enemy disturbing and strong surrounding enemy, and an unmanned intelligent individual under the control of the algorithm is autonomously switched to a proper confrontation stage according to the situation of surrounding enemy and my power; in the stage of walking to find enemies, the unmanned intelligent cluster formation is kept loose and approaches to enemies through approximate global information timed by the satellite; in the stage of soft surrounding to the enemy, the unmanned intelligent cluster harasses the enemy by fast impact with low probability while approaching the enemy, so as to induce the enemy matrix to the direction favorable for the final surrounding of the enemy; the hard surrounding enemy-fighting mode enables the intelligent cluster to be gathered quickly, and the enemy force can be surrounded quickly and effectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110883706.7ACN113627081B (en) | 2021-08-03 | 2021-08-03 | Unmanned intelligent cluster countermeasure control method based on eagle crowd hunting behavior bionical |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110883706.7ACN113627081B (en) | 2021-08-03 | 2021-08-03 | Unmanned intelligent cluster countermeasure control method based on eagle crowd hunting behavior bionical |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113627081Atrue CN113627081A (en) | 2021-11-09 |
| CN113627081B CN113627081B (en) | 2024-02-20 |
Family
ID=78382337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110883706.7AActiveCN113627081B (en) | 2021-08-03 | 2021-08-03 | Unmanned intelligent cluster countermeasure control method based on eagle crowd hunting behavior bionical |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113627081B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015162330A1 (en)* | 2014-04-25 | 2015-10-29 | Consejo Superior De Investigaciones Científicas (Csic) | Biomimetic and zoosemiotic aerial vehicle guided by an automatic pilot device |
| CN110109477A (en)* | 2019-05-10 | 2019-08-09 | 北京航空航天大学 | Unmanned plane cluster multi objective control optimization method based on dove colony intelligence backward learning |
| CN112269396A (en)* | 2020-10-14 | 2021-01-26 | 北京航空航天大学 | Unmanned aerial vehicle cluster cooperative confrontation control method for eagle pigeon-imitated intelligent game |
| CN112783209A (en)* | 2020-12-31 | 2021-05-11 | 北京航空航天大学 | Unmanned aerial vehicle cluster confrontation control method based on pigeon intelligent competition learning |
- 2021
- 2021-08-03CNCN202110883706.7Apatent/CN113627081B/enactiveActive
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015162330A1 (en)* | 2014-04-25 | 2015-10-29 | Consejo Superior De Investigaciones Científicas (Csic) | Biomimetic and zoosemiotic aerial vehicle guided by an automatic pilot device |
| CN110109477A (en)* | 2019-05-10 | 2019-08-09 | 北京航空航天大学 | Unmanned plane cluster multi objective control optimization method based on dove colony intelligence backward learning |
| CN112269396A (en)* | 2020-10-14 | 2021-01-26 | 北京航空航天大学 | Unmanned aerial vehicle cluster cooperative confrontation control method for eagle pigeon-imitated intelligent game |
| CN112783209A (en)* | 2020-12-31 | 2021-05-11 | 北京航空航天大学 | Unmanned aerial vehicle cluster confrontation control method based on pigeon intelligent competition learning |
Non-Patent Citations (3)
| Title |
|---|
| 姚宗信;: "基于多智能体的无人作战平台多机协同对抗多目标任务决策方法", 航空科学技术, no. 03* |
| 王虎;邓大松;: "集群式无人机能力分析及其防御对策研究", 飞航导弹, no. 04* |
| 郑国锋;于劲松;刘浩: "基于PXI总线的光学雷达自动测试系统研制", 计算机测量与控制, vol. 18, no. 05* |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113627081B (en) | 2024-02-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112269396B (en) | Unmanned aerial vehicle cluster cooperative confrontation control method for eagle pigeon-imitated intelligent game | |
| Zheng et al. | Behavior pattern (innate action) of individuals in fish schools generating efficient collective evasion from predation | |
| Duan et al. | A predator-prey particle swarm optimization approach to multiple UCAV air combat modeled by dynamic game theory | |
| CN108680063A (en) | A kind of decision-making technique for the dynamic confrontation of extensive unmanned plane cluster | |
| CN113741525B (en) | Multi-UAV collaborative attack and defense countermeasures based on strategy set MADDPG | |
| Demšar et al. | Simulating predator attacks on schools: Evolving composite tactics | |
| Fu et al. | Alpha C2–an intelligent air defense commander independent of human decision-making | |
| Hagelbäck | Hybrid pathfinding in StarCraft | |
| CN109460056A (en) | Unmanned plane cluster fight game decision-making method based on quantum krill group's mechanism of Evolution | |
| CN113435598B (en) | Knowledge-driven intelligent decision-making method for wargames | |
| Xiang et al. | Research on UAV swarm confrontation task based on MADDPG algorithm | |
| CN112783209A (en) | Unmanned aerial vehicle cluster confrontation control method based on pigeon intelligent competition learning | |
| CN109709981B (en) | Competitive antagonism method for unmanned aerial vehicle group | |
| CN111624996A (en) | Multi-unmanned-boat incomplete information trapping method based on game theory | |
| CN115031586B (en) | Intelligent target control method for antagonism based on array defensive close-range shooting behavior | |
| Zha et al. | Dynamic multi-team antagonistic games model with incomplete information and its application to multi-UAV | |
| CN117408376A (en) | A method and system for predicting the position of wargame operators based on battlefield division and attraction map | |
| CN116258003B (en) | A distributed computing method for collaborative situational awareness of intelligent agent clusters | |
| CN116661496A (en) | Multi-patrol-missile collaborative track planning method based on intelligent algorithm | |
| Wang et al. | A decision-making method for swarm agents in attack-defense confrontation | |
| Chen et al. | A MADDPG-based multi-agent antagonistic algorithm for sea battlefield confrontation | |
| CN113627081A (en) | Unmanned intelligent cluster confrontation control method based on bionic eagle hunting behavior | |
| CN115061495B (en) | Unmanned aerial vehicle group countermeasure autonomous control method based on eagle pigeon game | |
| CN116202373B (en) | A method for autonomously jamming and intercepting missiles by unmanned aerial vehicles | |
| CN116542129A (en) | Pilot following structure population countering method and system based on wolf population organization characteristics |
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 |