CN113268078A - Target tracking and trapping method for self-adaptive environment of unmanned aerial vehicle group - Google Patents
Target tracking and trapping method for self-adaptive environment of unmanned aerial vehicle group Download PDFInfo
- Publication number
- CN113268078A CN113268078A CN202110423332.0A CN202110423332A CN113268078A CN 113268078 A CN113268078 A CN 113268078A CN 202110423332 A CN202110423332 A CN 202110423332A CN 113268078 A CN113268078 A CN 113268078A
- Authority
- CN
- China
- Prior art keywords
- unmanned aerial
- aerial vehicle
- enclosure
- environment
- trapping
- 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
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
- G05D1/104—Simultaneous control of position or course in three dimensions specially adapted for aircraft involving a plurality of aircrafts, e.g. formation flying
Abstract
The invention discloses a target tracking and trapping method of an unmanned aerial vehicle group self-adaptive environment, which comprises the following steps: (1) establishing a multi-agent collaborative planning model by using an MADDPG algorithm to realize the tracking and trapping of the unmanned aerial vehicle group to the target; (2) when the unmanned aerial vehicle cluster approaches to the threat area, the GA algorithm is utilized to automatically adjust and re-plan the position of the unmanned aerial vehicle so as to avoid entering the threat area, the survival rate of the unmanned aerial vehicle is improved, and meanwhile, the enclosure task is completed. The establishment of the layered trapping model is divided into two layers: a trapping layer and a multi-agent training layer. The unmanned aerial vehicle cluster interacts with the environment in real time, so that the current environment state can be obtained at any time. And the trapping layer judges whether to adjust the formation from the current state and calculates a trapping position distribution scheme. Aiming at the dynamic change of the environment and the task, the invention improves the success rate of task execution in the relatively complex environment with threat, and simultaneously autonomously changes the enclosure position to avoid the threat area, thereby reducing the risk of the unmanned aerial vehicle cluster.
Description
Technical Field
The invention relates to the technical field of unmanned aerial vehicle cluster task planning, in particular to a target tracking and trapping method of an unmanned aerial vehicle cluster self-adaptive environment.
Background
Unmanned aerial vehicle has zero casualties, continuous operation, low cost and outstanding mobility's unique advantage such as for unmanned aerial vehicle cluster battle becomes research focus in recent years. In the aspect of unmanned aerial vehicle cluster command control cooperative task decision, colony intelligent algorithms based on ant colony algorithm, wolf colony algorithm and the like are mostly adopted.
However, the group intelligent algorithm cannot meet the characteristics of autonomy and autonomy of unmanned aerial vehicle clusters, and intensive learning is widely concerned and applied in recent years. However, the existing unmanned aerial vehicle based on reinforcement learning is simpler in target trapping problem scene setting, threat factors are less considered, and the trapping mode is not flexible enough.
Disclosure of Invention
In order to solve the technical problems, the invention provides a target tracking and trapping method of an unmanned aerial vehicle group self-adaptive environment, aiming at the dynamic change of the environment and tasks, the success rate of task execution is improved in a relatively complex environment with threat, and meanwhile, the trapping position is autonomously changed to avoid a threat area, so that the risk of the unmanned aerial vehicle group is reduced.
The technical purpose of the invention is realized by the following technical scheme:
a target tracking and enclosing method of an unmanned aerial vehicle cluster self-adaptive environment comprises the following steps:
s1: constructing an enclosure layer and a multi-agent training layer;
s2: determining the current environment state of the unmanned aerial vehicle through the real-time interaction between the unmanned aerial vehicle cluster and the environment;
s3: judging whether the ideal enclosure position of the unmanned aerial vehicle cluster is in a threat area, if so, executing S4, otherwise, executing S5;
s4: calculating a new capture position from the newly calculated capture position through the capture layer, and then calculating to obtain a capture position distribution scheme consumed by the shortest voyage;
s5: through the multi-agent training layer, a multi-agent cooperation model is constructed, and tracking and trapping of the unmanned aerial vehicle group on the target are achieved.
As a preferred scheme, the S4 process specifically includes the following steps:
s401: judging an overlapping area of an enclosure position around the enclosure target and a threat area;
s402: and selecting a set of non-overlapping regions theta ═ theta1,θ2Maximum value of θmaxThe average value of the angles of the non-overlapping areas of the unmanned planes is
s405: if it is notThen at the selected thetamaxThe location of the entrapment is calculated in part according to the following formula:
wherein: t isiIs the enclosure position of unmanned plane i, (c)x,cy) Is the position coordinate of the object to be enclosed, r is the enclosing radius, thetaaIs the average range of capture, theta, of each unmanned aerial vehicletIs the extent of overlap of the standard formation with the threat without taking into account the avoidance of the threat zone, θsIs the starting angle;
s406: calculating all allocation schemes according to the arrangement method;
s407: and calculating the route consumption of each scheme, and selecting the scheme with the minimum route consumption as the optimal distribution scheme.
As a preferred scheme, the S5 process specifically includes the following steps:
let a scene include N drones, each drone having a policy parameter θ ═ θ1,θ2,...θNUInstruction ofFor the strategy set of all drones, the strategy gradient of the unmanned plane i is obtained as follows:
wherein:is a function of Q value, aiIs the action of unmanned aerial vehicle i, oiIs the observation information of the unmanned aerial vehicle, including the position and speed information of the unmanned aerial vehicle relative to the target, x ═ o1,...oN) Representing observation information for N drones.
As a preferable scheme, in the S5 process, the reward sparseness problem is solved through a guiding reward function, which specifically includes the following steps:
let D denote an experience pool, which is used to store (x, x', a)1,...aN,r1,...,rN) Recording the experience of all the drones, x' being the new state after all the drones have performed the action, riIs a reward for interaction with the environment after the drone i performs an action.
As a preferred scheme, in the process of S5, the minimum policy gradient formula of the Critic network is:
wherein: s is a small batch of random samples and j is the index of the samples.
As a preferred scheme, the Actor network and the Critic network both adopt 4-layer fully-connected artificial neural networks, the number of neurons in each layer of the Actor network is [64,64,64,2], the input of the last layer is a 2-dimensional vector corresponding to the acceleration of the unmanned aerial vehicle on the x axis and the y axis; the number of neurons in the Critic network coal seam is [64,64,64,1], and the output of the last layer is the evaluation of the action.
As a preferred scheme, the Reward structure of each unmanned aerial vehicle is composed of three parts1+r2+r3,r1Reward, r, representing the distance of the drone from its location of enclosure2Penalty, r, representing collision of unmanned aerial vehicle with threat zone3Represents the penalty of collision of the drone with other drones.
In conclusion, the invention has the following beneficial effects:
the target tracking and enclosing method of the self-adaptive environment of the unmanned aerial vehicle cluster provided by the invention is combined with the MADDPG (Multi-agent Deep Deterministic Policy Gradient) and the GA Algorithm (Greedy Algorithm) to design the self-adaptive Algorithm of the MADDPG-GA, the fusion Algorithm can accelerate the learning efficiency, improve the rapidity of target tracking and enclosing of the unmanned aerial vehicle cluster in the complex environment, optimize the formation of the unmanned aerial vehicle cluster to reduce the probability of the unmanned aerial vehicle falling into the threat area, improve the enclosing success rate and reduce the cluster risk;
the unmanned aerial vehicle group trained by the unmanned aerial vehicle group self-adaptive environment target tracking and trapping method provided by the invention can rapidly trap targets in a complex environment, so that the defects of slow training and low learning rate of the traditional unmanned aerial vehicle based on reinforcement learning are overcome, and the defect that the traditional fixed trapping mode is easy to fall into a threat area and cannot complete trapping is avoided.
Drawings
FIG. 1 is a diagram of a layered containment model in an embodiment of the invention;
FIG. 2 is a schematic diagram of an unmanned aerial vehicle fleet for capturing targets proximate to a threat zone in an embodiment of the present invention;
fig. 3 is a schematic diagram of unmanned aerial vehicle enclosure location allocation in the embodiment of the present invention;
FIG. 4 is a block diagram of a MADDPG in an embodiment of the present invention;
fig. 5 is a schematic diagram of an Actor network in an embodiment of the present invention.
Detailed Description
This specification and claims do not intend to distinguish between components that differ in name but not function. In the following description and in the claims, the terms "include" and "comprise" are used in an open-ended fashion, and thus should be interpreted to mean "include, but not limited to. "substantially" means within an acceptable error range, and a person skilled in the art can solve the technical problem within a certain error range to substantially achieve the technical effect.
The terms in upper, lower, left, right and the like in the description and the claims are combined with the drawings to facilitate further explanation, so that the application is more convenient to understand and is not limited to the application.
The present invention will be described in further detail with reference to the accompanying drawings.
The invention designs a target tracking and trapping method of an unmanned aerial vehicle group self-adaptive environment, which mainly comprises the following steps: (1) establishing a multi-agent collaborative planning model by using an MADDPG algorithm to realize the tracking and trapping of the unmanned aerial vehicle group to the target; (2) when the unmanned aerial vehicle cluster approaches to the threat area, the GA algorithm is utilized to automatically adjust and re-plan the position of the unmanned aerial vehicle so as to avoid entering the threat area, the survival rate of the unmanned aerial vehicle is improved, and meanwhile, the enclosure task is completed. Establishing a layered trapping model as shown in figure 1, which is divided into two layers: a trapping layer and a multi-agent training layer. The unmanned aerial vehicle cluster interacts with the environment in real time, so that the current environment state can be obtained at any time. And the layer of enclosure judges whether to adjust the formation from the current state, and when the ideal enclosure position falls into a threat area, the layer calculates a new enclosure position by using a GA algorithm, and then calculates an enclosure position distribution scheme consumed by the shortest flight. After the pursuit position is determined, each unmanned aerial vehicle is trained in turn on a multi-agent system training layer, and the trained unmanned aerial vehicle cluster can execute a capture strategy to interact with the environment. The feasibility of the scheme is verified by simulation experiments.
1. Trapping layer
Determining the encirclement position according to the target position by using a greedy algorithm as shown in FIG. 2 comprises the following specific steps: if the surrounding positions around the target are overlapped with the environment threat area, selecting a non-overlapped area set theta as { theta ═ theta1,θ2Maximum value of θmaxThe average value of the angles of the non-overlapping areas of the unmanned planes is
If it isThen orderIf it is notThen at the selected thetamaxThe location of the entrapment is calculated in part according to the following formula:
in the formula: t isiIs the enclosure position of unmanned plane i, (c)x,cy) Is the position coordinate of the object to be enclosed, r is the enclosing radius, thetaaIs the average range of capture, theta, of each unmanned aerial vehicletIs the extent of overlap of the standard formation with the threat without taking into account the avoidance of the threat zone, θsIs the starting angle.
After the encirclement position is determined, the encirclement position is distributed according to the minimum total route consumption and the constraint relation between the unmanned aerial vehicle and the encirclement position as shown in figure 3, and the specific steps are as follows: calculating all allocation schemes according to the arrangement method; and further calculating the route consumption of each scheme, and selecting the scheme with the minimum route consumption as the optimal distribution scheme.
2. Multi-agent training layer
As shown in FIG. 4, the multi-agent system training layer based on the MADDPG framework has the characteristics of centralized training and distributed application. A plurality of strategies are learned for each intelligent agent (unmanned aerial vehicle), and the overall effect of all the strategies is utilized for optimization during improvement, so that the stability and robustness of the algorithm are improved. Each agent has an Actor and a Critic network for training and learning, and the specific principle is as follows: considering that there are N agents in a scene, the policy parameter of each drone is θ ═ θ1,θ2,...θNUInstruction ofFor the set of policies for all drones, it can be derived that the policy gradient for drone i is:
is a function of Q value, aiIs the action of unmanned aerial vehicle i, oiIs the observation information of the unmanned aerial vehicle, including the position and speed information of the unmanned aerial vehicle relative to the target, x ═ o1,...oN) Representing observation information of N drones for each agentAre independent, so the Reward structure of each agent is arbitrary.
In the invention, a guiding Reward function is designed to solve the Reward Sparse problem (Sparse Reward). D represents the experience pool used to store (x, x', a)1,...aN,r1,...,rN) The experience of all agents is recorded, x' is the new state after all agents have performed the action, riIs a reward for the agent i to interact with the environment after performing an action. Of Critic networksBy a loss functionIs updated, wherein
The Actor network is formed by the minimum policy gradient formula:
wherein: s is a small batch of random samples and j is the index of the samples.
Actor and Critic network architecture
The invention relates to an Actor and Critic network structure of an intelligent agent, namely an unmanned aerial vehicle, which adopts a 4-layer full-connection artificial neural network, wherein the number of neurons of each layer of the Actor is (64, 64,64, 2)]As in fig. 5, the input to the last layer is a 2-dimensional vector, corresponding to the accelerations of the drone in the x and y axes. The number of neurons in each layer of the Critic network is [64,64,64,1]]The output of the last layer is an evaluation of the action. The Reward structure of each unmanned aerial vehicle consists of three parts r1+r2+r3,r1The reward that the unmanned aerial vehicle is far away from the target position of the surrounding capture is represented, and the closer the distance is, the greater the reward is; r is2The penalty of collision between the unmanned aerial vehicle and the threat zone is represented and is related to a penalty coefficient xi, r3Represents the penalty of collision of the drone with other drones. The Reward structure can be flexibly set according to the maneuverability, the trapping task characteristic and the like of the unmanned aerial vehicle.
The present embodiment is only for explaining the present invention, and it is not limited to the present invention, and those skilled in the art can make modifications of the present embodiment without inventive contribution as needed after reading the present specification, but all of them are protected by patent law within the scope of the claims of the present invention.
Claims (8)
1. A target tracking and enclosing method of an unmanned aerial vehicle group self-adaptive environment is characterized by comprising the following steps:
s1: constructing an enclosure layer and a multi-agent training layer;
s2: determining the current environment state of the unmanned aerial vehicle through the real-time interaction between the unmanned aerial vehicle cluster and the environment;
s3: judging whether the ideal enclosure position of the unmanned aerial vehicle cluster is in a threat area, if so, executing S4, otherwise, executing S5;
s4: calculating a new capture position from the newly calculated capture position through the capture layer, and then calculating to obtain a capture position distribution scheme consumed by the shortest voyage;
s5: through the multi-agent training layer, a multi-agent cooperation model is constructed, and tracking and trapping of the unmanned aerial vehicle group on the target are achieved.
2. The method for target tracking and enclosure in a drone swarm adaptive environment according to claim 1, wherein the S4 process specifically includes the following steps:
s401: judging an overlapping area of an enclosure position around the enclosure target and a threat area;
s402: and selecting a set of non-overlapping regions theta ═ theta1,θ2Maximum value of θmaxThe average value of the angles of the non-overlapping areas of the unmanned planes is
s405: if it is notThen at the selected thetamaxThe location of the entrapment is calculated in part according to the following formula:
wherein: t isiIs the enclosure position of unmanned plane i, (c)x,cy) Is the position coordinate of the object to be enclosed, r is the enclosing radius, thetaaIs the average range of capture, theta, of each unmanned aerial vehicletIs the extent of overlap of the standard formation with the threat without taking into account the avoidance of the threat zone, θsIs the starting angle;
s406: calculating all allocation schemes according to the arrangement method;
s407: and calculating the route consumption of each scheme, and selecting the scheme with the minimum route consumption as the optimal distribution scheme.
3. The method for target tracking and enclosure in a drone swarm adaptive environment according to claim 2, wherein the S5 process specifically includes the following steps:
let a scene include N drones, each drone having a policy parameter θ ═ θ1,θ2,...θNUInstruction ofFor the strategy set of all drones, the strategy gradient of the unmanned plane i is obtained as follows:
wherein:is a function of Q value, aiIs the action of unmanned aerial vehicle i, oiIs the observation information of the unmanned aerial vehicle, including the position and speed information of the unmanned aerial vehicle relative to the target, x ═ o1,...oN) Representing observation information for N drones.
4. The target tracking and trapping method for the unmanned aerial vehicle fleet adaptive environment as claimed in claim 3, wherein in the step of S5, the problem of sparse rewards is solved through a guiding reward function, specifically comprising the steps of:
let D denote an experience pool, which is used to store (x, x', a)1,...aN,r1,...,rN) Recording the experience of all the drones, x' being the new state after all the drones have performed the action, riIs a reward for interaction with the environment after the drone i performs an action.
6. The method for target tracking and enclosure in a fleet of unmanned aerial vehicles adaptive environment of claim 3, wherein in the step of S5, the minimum policy gradient formula of Critic network is:
wherein: s is a small batch of random samples and j is the index of the samples.
7. The target tracking and capturing method for the unmanned aerial vehicle cluster self-adaptive environment according to any one of claims 4 to 6, wherein the Actor network and the Critic network both adopt 4 layers of fully connected artificial neural networks, the number of neurons in each layer of the Actor network is [64,64,64,2], the input of the last layer is a 2-dimensional vector corresponding to the acceleration of the unmanned aerial vehicle in the x and y axes; the number of neurons in the Critic network coal seam is [64,64,64,1], and the output of the last layer is the evaluation of the action.
8. The method for target tracking and encirclement in a fleet of unmanned aerial vehicles adaptive environment as claimed in claim 7, wherein the Reward structure of each unmanned aerial vehicle is composed of three parts r1+r2+r3,r1Reward, r, representing the distance of the drone from its location of enclosure2Penalty, r, representing collision of unmanned aerial vehicle with threat zone3Represents the penalty of collision of the drone with other drones.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110423332.0A CN113268078B (en) | 2021-04-20 | 2021-04-20 | Target tracking and capturing method for self-adaptive environment of unmanned aerial vehicle group |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110423332.0A CN113268078B (en) | 2021-04-20 | 2021-04-20 | Target tracking and capturing method for self-adaptive environment of unmanned aerial vehicle group |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113268078A true CN113268078A (en) | 2021-08-17 |
CN113268078B CN113268078B (en) | 2022-11-18 |
Family
ID=77228960
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110423332.0A Active CN113268078B (en) | 2021-04-20 | 2021-04-20 | Target tracking and capturing method for self-adaptive environment of unmanned aerial vehicle group |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113268078B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113625775A (en) * | 2021-09-10 | 2021-11-09 | 南京航空航天大学 | Multi-Unmanned Aerial Vehicle (UAV) enclosure method combining state prediction and distributed generation group (DDPG) |
CN113741525A (en) * | 2021-09-10 | 2021-12-03 | 南京航空航天大学 | Strategy set based MADDPG multi-unmanned aerial vehicle cooperative attack and defense countermeasure method |
CN114756052A (en) * | 2022-03-31 | 2022-07-15 | 电子科技大学 | Multi-target cooperative tracking method based on unmanned aerial vehicle group |
CN115019185A (en) * | 2022-08-03 | 2022-09-06 | 华中科技大学 | Brain-like continuous learning cooperative trapping method, system and medium |
CN115097861A (en) * | 2022-05-15 | 2022-09-23 | 西北工业大学 | Multi-Unmanned Aerial Vehicle (UAV) capture strategy method based on CEL-MADDPG |
CN116750211B (en) * | 2023-07-13 | 2024-01-23 | 四川大学 | Tracking defense method based on tracking target distribution and track planning |
CN115097861B (en) * | 2022-05-15 | 2024-04-26 | 西北工业大学 | Multi-unmanned aerial vehicle trapping strategy method based on CEL-MADDPG |
Citations (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090294584A1 (en) * | 2008-06-02 | 2009-12-03 | Gilbert Lovell | Stabilized UAV recovery system |
CN104942807A (en) * | 2015-04-16 | 2015-09-30 | 上海大学 | Method for capturing targets by aid of multiple robots on basis of extensive cooperative games |
US20160018224A1 (en) * | 2013-09-27 | 2016-01-21 | Regents Of The University Of Minnesota | Symbiotic Unmanned Aerial Vehicle and Unmanned Surface Vehicle System |
US20160241767A1 (en) * | 2015-02-13 | 2016-08-18 | Lg Electronics Inc. | Mobile terminal and method for controlling the same |
CN106647808A (en) * | 2017-01-05 | 2017-05-10 | 南宁市健佳网络科技有限公司 | Method for searching AUVs and allocating and controlling capturing tasks based on fuzzy control algorithm |
US20180155023A1 (en) * | 2016-12-05 | 2018-06-07 | Samsung Electronics Co., Ltd | Flight control method and electronic device for supporting the same |
CN108229465A (en) * | 2016-12-15 | 2018-06-29 | 谷歌公司 | For enhancing the system and method for the object visibility of air-borne imagery |
US20180268331A1 (en) * | 2014-06-11 | 2018-09-20 | Hartford Fire Insurance Company | UAV Routing and Data Extraction |
CN109079792A (en) * | 2018-09-05 | 2018-12-25 | 顺德职业技术学院 | A kind of target based on multirobot surrounds and seize method and system |
CN109451431A (en) * | 2018-10-23 | 2019-03-08 | 中国电子科技集团公司第二十九研究所 | It is a kind of for the civilian low slow navigation of small unmanned plane and the area-denial method of link |
CN109669477A (en) * | 2019-01-29 | 2019-04-23 | 华南理工大学 | A kind of cooperative control system and control method towards unmanned plane cluster |
CN110069076A (en) * | 2019-04-23 | 2019-07-30 | 北京航空航天大学 | A kind of unmanned plane cluster air battle method for surrounding and seize behavior based on violent wolf |
CN110608743A (en) * | 2019-10-18 | 2019-12-24 | 南京航空航天大学 | Multi-unmanned aerial vehicle collaborative route planning method based on multi-population chaotic grayling algorithm |
CN110716582A (en) * | 2019-10-16 | 2020-01-21 | 东南大学 | Multi-agent consistency tracking protocol design method suitable for intermittent DoS attack on communication |
CN110989626A (en) * | 2019-12-27 | 2020-04-10 | 四川大学 | Unmanned aerial vehicle path planning method based on control parameterization |
CN111399534A (en) * | 2020-02-25 | 2020-07-10 | 清华大学 | Method and system for capturing aerial medium-high speed moving targets by multiple unmanned aerial vehicles |
CN111624996A (en) * | 2020-05-12 | 2020-09-04 | 哈尔滨工程大学 | Multi-unmanned-boat incomplete information trapping method based on game theory |
CN111722643A (en) * | 2020-06-12 | 2020-09-29 | 北京航空航天大学 | Unmanned aerial vehicle cluster dynamic task allocation method imitating wolf colony cooperative hunting mechanism |
CN111880563A (en) * | 2020-07-17 | 2020-11-03 | 西北工业大学 | Multi-unmanned aerial vehicle task decision method based on MADDPG |
CN112216341A (en) * | 2020-09-16 | 2021-01-12 | 中国人民解放军国防科技大学 | Group behavior logic optimization method and computer readable storage medium |
CN112327872A (en) * | 2020-11-20 | 2021-02-05 | 哈尔滨工程大学 | Double unmanned ship cooperative track tracking method for oil spill containment |
-
2021
- 2021-04-20 CN CN202110423332.0A patent/CN113268078B/en active Active
Patent Citations (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090294584A1 (en) * | 2008-06-02 | 2009-12-03 | Gilbert Lovell | Stabilized UAV recovery system |
US20160018224A1 (en) * | 2013-09-27 | 2016-01-21 | Regents Of The University Of Minnesota | Symbiotic Unmanned Aerial Vehicle and Unmanned Surface Vehicle System |
US20180268331A1 (en) * | 2014-06-11 | 2018-09-20 | Hartford Fire Insurance Company | UAV Routing and Data Extraction |
US20160241767A1 (en) * | 2015-02-13 | 2016-08-18 | Lg Electronics Inc. | Mobile terminal and method for controlling the same |
CN104942807A (en) * | 2015-04-16 | 2015-09-30 | 上海大学 | Method for capturing targets by aid of multiple robots on basis of extensive cooperative games |
US20180155023A1 (en) * | 2016-12-05 | 2018-06-07 | Samsung Electronics Co., Ltd | Flight control method and electronic device for supporting the same |
CN108229465A (en) * | 2016-12-15 | 2018-06-29 | 谷歌公司 | For enhancing the system and method for the object visibility of air-borne imagery |
CN106647808A (en) * | 2017-01-05 | 2017-05-10 | 南宁市健佳网络科技有限公司 | Method for searching AUVs and allocating and controlling capturing tasks based on fuzzy control algorithm |
CN109079792A (en) * | 2018-09-05 | 2018-12-25 | 顺德职业技术学院 | A kind of target based on multirobot surrounds and seize method and system |
CN109451431A (en) * | 2018-10-23 | 2019-03-08 | 中国电子科技集团公司第二十九研究所 | It is a kind of for the civilian low slow navigation of small unmanned plane and the area-denial method of link |
CN109669477A (en) * | 2019-01-29 | 2019-04-23 | 华南理工大学 | A kind of cooperative control system and control method towards unmanned plane cluster |
CN110069076A (en) * | 2019-04-23 | 2019-07-30 | 北京航空航天大学 | A kind of unmanned plane cluster air battle method for surrounding and seize behavior based on violent wolf |
CN110716582A (en) * | 2019-10-16 | 2020-01-21 | 东南大学 | Multi-agent consistency tracking protocol design method suitable for intermittent DoS attack on communication |
CN110608743A (en) * | 2019-10-18 | 2019-12-24 | 南京航空航天大学 | Multi-unmanned aerial vehicle collaborative route planning method based on multi-population chaotic grayling algorithm |
CN110989626A (en) * | 2019-12-27 | 2020-04-10 | 四川大学 | Unmanned aerial vehicle path planning method based on control parameterization |
CN111399534A (en) * | 2020-02-25 | 2020-07-10 | 清华大学 | Method and system for capturing aerial medium-high speed moving targets by multiple unmanned aerial vehicles |
CN111624996A (en) * | 2020-05-12 | 2020-09-04 | 哈尔滨工程大学 | Multi-unmanned-boat incomplete information trapping method based on game theory |
CN111722643A (en) * | 2020-06-12 | 2020-09-29 | 北京航空航天大学 | Unmanned aerial vehicle cluster dynamic task allocation method imitating wolf colony cooperative hunting mechanism |
CN111880563A (en) * | 2020-07-17 | 2020-11-03 | 西北工业大学 | Multi-unmanned aerial vehicle task decision method based on MADDPG |
CN112216341A (en) * | 2020-09-16 | 2021-01-12 | 中国人民解放军国防科技大学 | Group behavior logic optimization method and computer readable storage medium |
CN112327872A (en) * | 2020-11-20 | 2021-02-05 | 哈尔滨工程大学 | Double unmanned ship cooperative track tracking method for oil spill containment |
Non-Patent Citations (8)
Title |
---|
PIACENTINI, C等: "Autonomous Target Search with Multiple Coordinated UAVs", 《JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH》 * |
Q NING 等: "Multi-UAVs trajectory and mission cooperative planning based on the Markov model", 《PHYSICAL COMMUNICATION》 * |
严富函: "考虑追捕者能力不确定性的多Agent追逃问题研究", 《中国博士学位论文全文数据库 基础科学辑》 * |
宋梅萍: "结合围捕问题的合作多智能体强化学习研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
杨川力等: "基于GAPSO-TS的多基地无人机航路-时间协同规划", 《火力与指挥控制》 * |
胡中华: "基于智能优化算法的无人机航迹规划若干关键技术研究", 《中国博士学位论文全文数据库 工程科技Ⅱ辑》 * |
赵金: "基于NEAT的多机器人协调围捕问题研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
雷川等: "基于混合变异策略差分进化算法的无人机航路规划", 《火力与指挥控制》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113625775A (en) * | 2021-09-10 | 2021-11-09 | 南京航空航天大学 | Multi-Unmanned Aerial Vehicle (UAV) enclosure method combining state prediction and distributed generation group (DDPG) |
CN113741525A (en) * | 2021-09-10 | 2021-12-03 | 南京航空航天大学 | Strategy set based MADDPG multi-unmanned aerial vehicle cooperative attack and defense countermeasure method |
CN113741525B (en) * | 2021-09-10 | 2024-02-06 | 南京航空航天大学 | Policy set-based MADDPG multi-unmanned aerial vehicle cooperative attack and defense countermeasure method |
CN114756052A (en) * | 2022-03-31 | 2022-07-15 | 电子科技大学 | Multi-target cooperative tracking method based on unmanned aerial vehicle group |
CN115097861A (en) * | 2022-05-15 | 2022-09-23 | 西北工业大学 | Multi-Unmanned Aerial Vehicle (UAV) capture strategy method based on CEL-MADDPG |
CN115097861B (en) * | 2022-05-15 | 2024-04-26 | 西北工业大学 | Multi-unmanned aerial vehicle trapping strategy method based on CEL-MADDPG |
CN115019185A (en) * | 2022-08-03 | 2022-09-06 | 华中科技大学 | Brain-like continuous learning cooperative trapping method, system and medium |
CN115019185B (en) * | 2022-08-03 | 2022-10-21 | 华中科技大学 | Brain-like continuous learning cooperative trapping method, system and medium |
CN116750211B (en) * | 2023-07-13 | 2024-01-23 | 四川大学 | Tracking defense method based on tracking target distribution and track planning |
Also Published As
Publication number | Publication date |
---|---|
CN113268078B (en) | 2022-11-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113268078B (en) | Target tracking and capturing method for self-adaptive environment of unmanned aerial vehicle group | |
CN110488859B (en) | Unmanned aerial vehicle route planning method based on improved Q-learning algorithm | |
Zhu et al. | Multi-robot flocking control based on deep reinforcement learning | |
CN112180967B (en) | Multi-unmanned aerial vehicle cooperative countermeasure decision-making method based on evaluation-execution architecture | |
CN108459616B (en) | Unmanned aerial vehicle group collaborative coverage route planning method based on artificial bee colony algorithm | |
Ma et al. | Multi-robot target encirclement control with collision avoidance via deep reinforcement learning | |
EP1934894A2 (en) | Hybrid control device | |
CN113159432A (en) | Multi-agent path planning method based on deep reinforcement learning | |
CN113096446B (en) | Multi-ship collision avoidance decision-making method under hybrid navigation scene, storage medium and processor | |
CN109784201A (en) | AUV dynamic obstacle avoidance method based on four-dimensional risk assessment | |
Zhang et al. | A self-heuristic ant-based method for path planning of unmanned aerial vehicle in complex 3-D space with dense U-type obstacles | |
Cao et al. | Hunting algorithm for multi-auv based on dynamic prediction of target trajectory in 3d underwater environment | |
Ma et al. | CCIBA*: An improved BA* based collaborative coverage path planning method for multiple unmanned surface mapping vehicles | |
CN113359437B (en) | Hierarchical model prediction control method for multi-agent formation based on evolutionary game | |
CN116755474A (en) | Electric power line inspection method and system for unmanned aerial vehicle | |
Wang et al. | Obstacle avoidance of UAV based on neural networks and interfered fluid dynamical system | |
Farhood | Neural network based control system for robots group operating in 2-d uncertain environment | |
CN114815891A (en) | PER-IDQN-based multi-unmanned aerial vehicle enclosure capture tactical method | |
Hu et al. | Multi-UAV coverage path planning: a distributed online cooperation method | |
CN110806758A (en) | Unmanned aerial vehicle cluster autonomous level self-adaptive adjustment method based on scene fuzzy cognitive map | |
Li et al. | Vg-swarm: A vision-based gene regulation network for uavs swarm behavior emergence | |
CN114548663A (en) | Scheduling method for charging unmanned aerial vehicle to charge task unmanned aerial vehicle in air | |
CN112651486A (en) | Method for improving convergence rate of MADDPG algorithm and application thereof | |
CN115574826B (en) | National park unmanned aerial vehicle patrol path optimization method based on reinforcement learning | |
CN116757249A (en) | Unmanned aerial vehicle cluster strategy intention recognition method based on distributed reinforcement learning |
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 |