CN111343023A - Distributed cooperative decision-making method for motion control of self-adaptive mobile network node - Google Patents
Distributed cooperative decision-making method for motion control of self-adaptive mobile network node Download PDFInfo
- Publication number
- CN111343023A CN111343023A CN202010130516.3A CN202010130516A CN111343023A CN 111343023 A CN111343023 A CN 111343023A CN 202010130516 A CN202010130516 A CN 202010130516A CN 111343023 A CN111343023 A CN 111343023A
- Authority
- CN
- China
- Prior art keywords
- node
- decision
- nodes
- estimation
- target
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/12—Discovery or management of network topologies
Abstract
The invention belongs to the field of signal processing, in particular to a distributed cooperative decision method for motion control of a self-adaptive mobile network node; in the case of multiple targets with mobility, the decision of the node changes from moment to moment according to the position of the node and the position of the target. The decision between the node and the neighbor node is fused and exchanged in a diffusion cooperation mode, and is combined with the decision of the node to perform distributed decision, so that more accurate decision is obtained. Compared with the existing self-adaptive mobile network with a plurality of targets, the application scene is more complex and the complex situation of dynamic change of the positions of the targets and the situation that the motion states of the nodes have different choices are considered; meanwhile, a distributed method is used for making motion state decision, so that the survival rate of the nodes is effectively increased while the predation success rate is ensured.
Description
Technical Field
The invention belongs to the field of signal processing, and particularly relates to a distributed cooperative decision in an adaptive mobile network, in particular to a distributed cooperative decision for the adaptive mobile network with various motion states.
Background
In recent years, due to the rapid development of modern digital signal processing and wireless communication technologies and the wide application of wireless sensor devices, data processing of networks becomes a future development direction, wherein a distributed method is widely researched due to the fact that information sharing among nodes is fully utilized, and the distributed method has low calculation complexity, good expandability and good anti-interference performance. Compared with a centralized method, the distributed method can more accurately transmit and process data, and can be widely applied to actual scenes such as parameter estimation, target tracking and the like.
In recent years, some researchers have studied Distributed Estimation methods, for example, in "dominant F S, SayedA H.Di dispersion LMS Strategies for Distributed Estimation [ J. IEEE Transactionson Signal processing g,2010,58(3): 1035-; the distributed method is widely applied to network signal processing due to simple structure and low calculation complexity; but only for the case where the topology of the network is static and does not change, parameter estimation is performed for one target.
Furthermore, the scholars optimize the combination coefficient of the distributed dispersive LMS method aiming at the condition that the network has a plurality of targets, and can achieve better estimation performance under the condition that the network has a plurality of targets. For example, a real-time updated method for calculating distributed combination coefficients applied to multi-target parameter estimation is proposed in "Chen J, Richard C, S eye A H. Difsion LMS Over Multi task Networks [ J ]. IEEETransactions on Signal processing 2015,63(11): 2733-; however, the method does not consider the situation of the change of the topological structure of the network and the situation of tracking the moving target, does not control the moving state of the network node, and does not relate to distributed decision making.
Still further, some researchers apply the distributed method to the Adaptive mobile network, for example, in "Tu S, SayedA h. mobile Adaptive Networks [ J ]. IEEE Journal of Selected topocs in signaling processing,2011,5(4):649 + 664" the distributed diffusible LMS method is applied to the Adaptive mobile network with real-time change in the topology of the network to track a single target; the application scene of the proposed method is single, only a single static target (predator) exists in the space, the maneuvering performance of the target is not considered, only a single target exists in the space, and the situation of multiple targets is not considered; the motion control method is simple and only considers that the node only has a tracked motion state, and does not relate to distributed decision making.
Still further, there are trainees who have expanded the motion situation of the adaptive mobile network, for example, in "TuS, Sa ed a h. tracking is latent of mobile adaptive networks [ C ].2010 Conference record of h e form Fourth adaptive communication on Signals, Systems and computers, performance group, CA, USA: ieee.2010: 698-one 702", two forms of targets including predators and predators, and network nodes are in a motion state of tracking or escaping according to the position of the target of motion; however, the application scenario is that only a single predator target and a predator target are in space, the situation of a plurality of predator targets and predator targets is not considered, and distributed decision making is not involved.
Still further, some scholars have proposed a distributed Decision method in the adaptive mobile network, for example, a Decision method applied to the distributed network is proposed in "Kh awatmi S, Zoubir a M, safe a h. decentralized Decision-Making Over Multi-Task Netwo rks [ J ]. Signal Processing,2019,160: 229-; the node carries out decision making by utilizing information sharing of the node and surrounding neighbor nodes, but the decision making process is independent and does not influence each other at each node, decision making results of other nodes are not considered, and the decision making method can cause decision making errors due to insufficient information; and a mechanism of nodes in the self-adaptive mobile network, which can not realize the target avoidance. In practice, the motion state of the node of the adaptive mobile network for tracking or avoiding the target may change, for example, when the predator is too close to the node, the node changes from the motion state of target tracking to the motion state of avoiding; therefore, it is necessary and reasonable to adopt a distributed collaborative decision-making based method to realize the adaptive and continuous motion control of the network nodes.
Disclosure of Invention
The invention aims to provide a distributed cooperative decision method for controlling the motion of nodes of an adaptive mobile network, which is used for solving the problem of performing distributed cooperative decision under the condition that the nodes in the adaptive mobile network have different motion states.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the distributed cooperative decision method for the motion control of the self-adaptive maneuvering network node is characterized by comprising the following steps of:
step 1: setting N nodes in the network, and automatically clustering according to different estimation targets;
for node k at the current time i, defining a matrix E:
wherein [ ·]l,kThe elements of the ith row and kth column of the matrix in brackets are indicated,is the domain of node k;
all of (E)]l,k Node 1 is in the same cluster as node k, labeled asAnd all in the neighborhood of node k do not belong toThe nodes of (a) form a set, labelThe symbol | represents the number of elements in the set;
wherein, wm(i)、wn(i) Respectively representing the estimation values of a certain target at the time i in the node m and the node n, wherein zeta > 0 is a preset threshold value;
if [ F ]]m,nIf 1, the node m and the node n are determined to be the same, belong to the same group, and will be taken according to the value of the matrix FGrouping to obtain G groups:
Wherein the content of the first and second substances,meaning rounding to the nearest integer, sn(i) Representing the estimation type of the estimation target of the node n at the time i;
and 4, step 4: carrying out information sharing of estimation values between adjacent nodes, arranging the estimation values of all groups together with the estimation values of the nodes to obtain a matrix Yk(i)、Zk(i):
Wherein, the matrix Yk(i) A position estimate vector containing all predators within node k, having a dimension ofZk(i) A position estimate vector comprising all predators within node k, having a dimension ofRepresenting the number of different estimates of predators and predators, respectively, that are present at node k; t is t1,...,t n1 st to nth groups indicating that the estimated type is predator among the G groups, e1,...,e m1 st to m-th groups representing estimated types as predators among the G groups;
and 5: determining a motion state decision vector g based on the position of the node and the included estimatek(i):
If there is no estimated value of predators in node k, node k is in an avoidance state:
wherein the content of the first and second substances,distance between node k and nearest predator, symbol [. ]]:,lVector, x, representing the l-th column of the matrixk(i) Denotes the position of node k, ReIs the threat radius of the predator;
if there is no estimated predator in node k, node k is in a tracking state:
gk(i)=[0,1]T;
if two types of estimates are contained in node k:
step 6: carrying out information interaction on the decision vector of the node and the decision of the neighbor node to obtain a decision intermediate value iota:
wherein, [ g ]l(i)]1Deciding vector g for node ll(i) Epsilon is a preset threshold;
and 7: and carrying out distributed cooperative decision updating on the decision vector of the node containing the two types of estimation values, wherein if the node k contains the two types of estimation values:
and 8: and performing motion control according to the decision vector:
wherein the content of the first and second substances,to track the tracking speed of the target: is a constant, representing the speed,
to hide inEscape speed outside the threat radius of the avoidance objective: is a constant, representing velocity;
the invention has the beneficial effects that:
the distributed cooperative decision method for the motion control of the self-adaptive mobile network node provided by the invention has the following advantages:
1. the method provided by the invention can perform distributed, simultaneous, self-adaptive and continuous motion control on the self-adaptive mobile network node under the condition of tracking a single or a plurality of targets;
2. the method provided by the invention can implement distributed, simultaneous, adaptive and continuous cooperative decision in the motion process on the adaptive mobile network node under the condition of tracking a single or a plurality of targets;
3. the distributed cooperative decision method provided by the invention can effectively reduce decision errors possibly caused by insufficient information of a single node;
4. according to the invention, the decision result is shared among the neighbor nodes through a diffusion mechanism, and the local decision result and the decision results of the neighbor nodes around are combined to obtain a cooperative decision result.
Drawings
Fig. 1 is a flowchart of a distributed cooperative decision method when an adaptive mobile network node moves.
FIG. 2 is a diagram of the location of predators, and the network at an initial time in an embodiment.
FIG. 3 is a diagram of the location of predators and predators during movement at a network node in an embodiment.
FIG. 4 is a diagram showing simulation results in the example.
Detailed Description
The invention will be further described with reference to the accompanying drawings in which:
the embodiment provides a distributed cooperative decision method for motion control of a self-adaptive mobile network node, and the flow of the method is shown in fig. 1; the method comprises the following specific steps:
step 1: setting N nodes in the network, wherein each node selects one target from a plurality of targets to estimate the position of the target, and automatically clustering according to different estimated targets;
a matrix E of dimension N × is defined as:
wherein [ ·]l,kThe elements of the ith row and kth column of the matrix in brackets are indicated,the method is a set composed of all neighbor nodes which directly communicate with the node k in the field of the node k;
and, define all of the groups [ E ]]l,kAll nodes l 1 belong to the same cluster as node kMeaning that node l is a neighbor node of node k and node l and node k have the same estimation target at the current moment;
wherein [ ·]m,nThe elements of the m-th row and n-th column of the matrix in brackets are indicated,all of the neighbors in the neighborhood of node k do not belong toRepresents the number of elements in the set, |, a column vector w of dimension M × 1 (M represents the spatial dimension)m(i) A column vector w representing an estimated value of a target at time i and node mn(i) Showing the estimated value of a certain target in the time i and the node n, wherein zeta > 0 is a preset threshold value and a smaller threshold value used for judging whether the targets of the node m and the node n are the same or not;
if [ F ]]m,nIf 1, the node m and the node n are regarded as the same target, i.e. belong to the same group, and the value of the matrix F will be taken as the basisGrouping is carried out, and the grouping is divided into G groups:the targets estimated within each group are different from each other;
and step 3: within node k, estimates and estimate types for common targets within each group are computed:
wherein the content of the first and second substances,denotes taking the nearest integer, sn(i) Indicating the estimation type of the estimation target of node n at time i, respectively representThe estimation type and corresponding estimation value of the group;
the invention assumes that the node knows the estimation type of the estimation target: predators(s) that the node wishes to avoid pursuingk(i) 0) or predators(s) whose node wishes to prey onk(i) 1), a uniform weighted combination method is applied in each group to obtain the estimation type and estimation value of the common target in each group;
and 4, step 4: carrying out information sharing of estimation values among the adjacent nodes, and arranging the estimation values of all groups together with the estimation values of the nodes in the node k;
wherein the dimensions are respectivelyMatrix Y ofk(i)、Zk(i) Respectively storing all predators and predators in the node kIs determined by the position estimation vector of (a),representing the number of different estimates of predators and predators, respectively, that are present at node k; t is t1,...,t n1 st to nth groups indicating that the estimated type is predator among the G groups, e1,...,em1 st to m-th groups representing estimated types as predators among the G groups;
all will beIs estimated value ofTogether with if sk(i) 0 matrix Y of estimates of node k itselfk(i) All will be similarIs estimated value ofTogether with if sk(i) The matrix Z is formed by the estimated values of 1 node k itselfk(i) (ii) a A plurality of estimates of predator and predator targets are obtained within node k;
and 5: within each node, a preliminary motion state decision vector g is determined based on the node's location and the included estimatesk(i);
For node k:
if there is no estimated value of predators in node k, node k is in an avoidance state:
wherein the content of the first and second substances,distance between node k and nearest predator, symbol [. ]]:,lTo representVector of the l-th column of the matrix, xk(i) Denotes the position of node k, ReIs the threat radius of the predator;
if there is no estimated predator in node k, node k is in a tracking state:
gk(i)=[0,1]T;
if two types of estimates are contained in node k:
preliminary decision vector gk(i) Represents the preliminary motion state of the node k, 0 represents the tracking state, and 1 represents the avoidance state;
step 6: sharing the information of the decision of the node and the decision of the neighbor node to obtain a decision intermediate value iota:
wherein, [ g ]l(i)]1Deciding vector g for node ll(i) Epsilon is a preset threshold value and is used for judging whether the device is in an escape state or not;
and 7: and carrying out distributed cooperative decision making to determine a final decision in the node containing two forms of target estimation values:
when the node determines that the motion state is escape according to the state of the neighbor node and the initial decision of the node is inconsistent, the node does not adopt decision determination of the node in order to ensure the safety of the node, and adopts a result obtained by using distributed cooperative decision;
and 8: and performing motion control according to the final decision vector:
wherein the content of the first and second substances,to track the tracking speed of the target: is a constant, representing the speed,
to escape speed outside the threat radius of the evasive target: is a constant, representing velocity;
The feasibility and the superiority of the method of the invention are illustrated by applying the method of the invention to a basic model of the Adaptive mobile network proposed in the document "TuS, Sayed A H. Mobile Adaptive Networks [ J ]. IEEE Journal of Selected Topicsin Signal processing,2011,5(4): 649-.
Simulation conditions are as follows:
an adaptive mobile network containing 80 nodes is used, the topological structure of the adaptive mobile network is changed along with the time, and each node selects 7 nodes closest to the node as neighbor nodes within the range of 5 radiuses of the node at each moment. There are 2 targets in the space, including 3 predators who wish to move to the nodal target, and 10 predators who are in constant linear motion. At each moment the predator estimates the node in space using the LMS method and chooses the node that is tracked closest to it.
Predators apply a state space model:
xi+1=Φxi+Γξi,
wherein the content of the first and second substances,in 2-dimensional space, the position and the velocity of the x axis and the position and the velocity of the y axis respectively, T is the time interval of sampling between each iteration and is taken as 0.1, and the noise ξiIs white gaussian noise with a variance of 0.02.
Setting parameters: threshold ζ of grouping is 0.1, threat radius R of predator e5, threshold epsilon of decision 0.2, tracking speedEscape velocity of state of motion 2 isThe node is considered to be predated when the distance between the node and the nearest predator is less than 1, and the predator is considered to be successfully predated by the node when the distance between the node and the nearest predator is greater than 3 nodes within the range of the safe distance of 2 of the predator. Average survival rate of nodes RsAverage predation success rate R of sum nodepIs defined as
Wherein L is the number of independent experiments,the number of nodes still surviving in the l independent experiment,the number of predators who had been successfully ingested in the i independent experiments.
The movement of the network and the object over time is shown in fig. 3. When a predator approaches to the node, part of the nodes begin to escape and share the decision of the predator to the neighbor nodes around the predator, and the nearby nodes are determined to be in the escaping state through distributed cooperative decision making. Some nodes in this process are successfully ingested because of their proximity to the predator. Some nodes are in the state of tracking the predators, and after one predator is successfully predated, the nodes go to track the next predator, and after a period of time, the predators are successfully predated by the nodes.
The mean of the node survival rates among 100 simulation experiments is shown in fig. 4. Three algorithms of different forms of distributed decision making, information sharing without estimation values between adjacent nodes and information sharing without decision making between adjacent nodes are respectively carried out. The graph and the following table show that the survival rate of the nodes is effectively increased by applying the distributed cooperative decision method for network node motion control provided by the invention under the condition of ensuring that the predation success rate is not changed.
Distributed decision making | Decision-free information sharing | Estimate-free information sharing | |
Average predation rate | 99.8% | 99.7% | 99.8% |
While the invention has been described with reference to specific embodiments, any feature disclosed in this specification may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise; all of the disclosed features, or all of the method or process steps, may be combined in any combination, except mutually exclusive features and/or steps.
Claims (1)
1. The distributed cooperative decision method for the motion control of the self-adaptive maneuvering network node is characterized by comprising the following steps of:
step 1: setting N nodes in the network, and automatically clustering according to different estimation targets;
for node k at the current time i, defining a matrix E:
wherein [ ·]l,kThe elements of the ith row and kth column of the matrix in brackets are indicated,is the domain of node k;
all of (E)]l,kNode 1 is in the same cluster as node k, labeled asAnd all in the neighborhood of node k do not belong toThe nodes of (a) form a set, labelThe symbol | represents the number of elements in the set;
wherein, wm(i)、wn(i) Respectively representing the estimation values of a certain target at the time i in the node m and the node n, wherein zeta > 0 is a preset threshold value;
if [ F ]]m,nIf 1, the node m and the node n are determined to be the same, belong to the same group, and will be taken according to the value of the matrix FGrouping to obtain G groups:
Wherein the content of the first and second substances,meaning rounding to the nearest integer, sn(i) Representing the estimation type of the estimation target of the node n at the time i;
and 4, step 4: carrying out information sharing of estimation values between adjacent nodes, arranging the estimation values of all groups together with the estimation values of the nodes to obtain a matrix Yk(i)、Zk(i):
Wherein, the matrix Yk(i) A position estimate vector containing all predators within node k, having a dimension ofZk(i) A position estimate vector comprising all predators within node k, having a dimension of Representing the number of different estimates of predators and predators, respectively, that are present at node k; t is t1,...,tn1 st to nth groups indicating that the estimated type is predator among the G groups, e1,...,emRepresenting groups 1 to m indicating that the estimated type is predator among the G groups;
and 5: determining a motion state decision vector g based on the position of the node and the included estimatek(i):
If there is no estimated value of predators in node k, node k is in an avoidance state:
wherein the content of the first and second substances, distance between node k and nearest predator, symbol [. ]]:,lVector, x, representing the l-th column of the matrixk(i) Denotes the position of node k, ReIs the threat radius of the predator;
if there is no estimated predator in node k, node k is in a tracking state:
gk(i)=[0,1]T;
if two types of estimation values are contained in node kDetermines whether in tracking or evasive state:
step 6: sharing the information of the decision of the node and the decision of the neighbor node to obtain a decision intermediate value iota:
wherein, [ g ]l(i)]1Deciding vector g for node ll(i) Epsilon is a preset threshold;
and 7: and carrying out distributed cooperative decision updating on the decision vector of the node containing the two types of estimation values, wherein if the node k contains the two types of estimation values:
and 8: and performing motion control according to the decision vector:
wherein the content of the first and second substances,to track the tracking speed of the target: is a constant and represents the speed,
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010130516.3A CN111343023B (en) | 2020-02-28 | 2020-02-28 | Distributed cooperative decision-making method for motion control of self-adaptive mobile network node |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010130516.3A CN111343023B (en) | 2020-02-28 | 2020-02-28 | Distributed cooperative decision-making method for motion control of self-adaptive mobile network node |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111343023A true CN111343023A (en) | 2020-06-26 |
CN111343023B CN111343023B (en) | 2021-10-26 |
Family
ID=71181892
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010130516.3A Active CN111343023B (en) | 2020-02-28 | 2020-02-28 | Distributed cooperative decision-making method for motion control of self-adaptive mobile network node |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111343023B (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022095246A1 (en) * | 2020-11-06 | 2022-05-12 | 山东大学 | Edge smart grid cooperative decision-making method based on differential privacy mechanism |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104023394A (en) * | 2014-06-17 | 2014-09-03 | 河海大学 | WSN positioning method based on self-adaptation inertia weight |
US20140278946A1 (en) * | 2013-03-15 | 2014-09-18 | Codegeo, Inc. | Advertisement and actual benefit distribution in a geocaching/treasure hunting environment |
CN106157154A (en) * | 2016-07-21 | 2016-11-23 | 重庆大学 | For the complex network community discovery method under the adaptive Evolutionary Vespertilio algorithm of media network data |
US20170076046A1 (en) * | 2015-09-10 | 2017-03-16 | Roche Molecular Systems, Inc. | Informatics platform for integrated clinical care |
CN107659989A (en) * | 2017-10-24 | 2018-02-02 | 东南大学 | The dormancy of wireless sensor network node distributed measurement and method for tracking target |
CN109902873A (en) * | 2019-02-28 | 2019-06-18 | 长安大学 | A method of the cloud manufacturing resource allocation based on modified whale algorithm |
-
2020
- 2020-02-28 CN CN202010130516.3A patent/CN111343023B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140278946A1 (en) * | 2013-03-15 | 2014-09-18 | Codegeo, Inc. | Advertisement and actual benefit distribution in a geocaching/treasure hunting environment |
CN104023394A (en) * | 2014-06-17 | 2014-09-03 | 河海大学 | WSN positioning method based on self-adaptation inertia weight |
US20170076046A1 (en) * | 2015-09-10 | 2017-03-16 | Roche Molecular Systems, Inc. | Informatics platform for integrated clinical care |
CN106157154A (en) * | 2016-07-21 | 2016-11-23 | 重庆大学 | For the complex network community discovery method under the adaptive Evolutionary Vespertilio algorithm of media network data |
CN107659989A (en) * | 2017-10-24 | 2018-02-02 | 东南大学 | The dormancy of wireless sensor network node distributed measurement and method for tracking target |
CN109902873A (en) * | 2019-02-28 | 2019-06-18 | 长安大学 | A method of the cloud manufacturing resource allocation based on modified whale algorithm |
Non-Patent Citations (3)
Title |
---|
JIE CHEN等: ""ADAPTIVE CLUSTERING FOR MULTITASK DIFFUSION NETWORKS"", 《2015 23RD EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO)》 * |
SHENG-YUAN TU等: ""Mobile Adaptive Networks"", 《IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING》 * |
袁莞迈: ""移动传感器网络中自适应蜂拥控制策略研究"", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2022095246A1 (en) * | 2020-11-06 | 2022-05-12 | 山东大学 | Edge smart grid cooperative decision-making method based on differential privacy mechanism |
Also Published As
Publication number | Publication date |
---|---|
CN111343023B (en) | 2021-10-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhu et al. | Deep reinforcement learning for mobile edge caching: Review, new features, and open issues | |
CN112181971A (en) | Edge-based federated learning model cleaning and equipment clustering method, system, equipment and readable storage medium | |
CN111416774B (en) | Network congestion control method and device, computer equipment and storage medium | |
CN113613301B (en) | Air-ground integrated network intelligent switching method based on DQN | |
CN111343023B (en) | Distributed cooperative decision-making method for motion control of self-adaptive mobile network node | |
CN110958625B (en) | Real-time multi-modal language analysis system and method based on mobile edge intelligence | |
CN113378466A (en) | DQN-based radar countermeasure intelligent decision-making method | |
CN110601777A (en) | Method for estimating satellite-ground downlink co-channel interference under low-orbit mobile satellite constellation | |
CN116974751A (en) | Task scheduling method based on multi-agent auxiliary edge cloud server | |
Wang et al. | Design of intelligent connected cruise control with vehicle-to-vehicle communication delays | |
CN114779827A (en) | Virtual potential field collaborative obstacle avoidance topological control method based on heterogeneous unmanned aerial vehicle formation | |
CN115499921A (en) | Three-dimensional trajectory design and resource scheduling optimization method for complex unmanned aerial vehicle network | |
CN113342010B (en) | Self-triggering communication underwater glider cooperative control system | |
CN116704823B (en) | Unmanned aerial vehicle intelligent track planning and general sense resource allocation method based on reinforcement learning | |
Evmorfos et al. | Deep q learning with fourier feature mapping for mobile relay beamforming networks | |
CN113727308A (en) | Edge calculation unloading optimization method based on vehicle position prediction | |
CN116600316A (en) | Air-ground integrated Internet of things joint resource allocation method based on deep double Q networks and federal learning | |
CN111930435B (en) | Task unloading decision method based on PD-BPSO technology | |
CN114779828B (en) | Unmanned aerial vehicle cluster topological control and intelligent anti-collision method based on heterogeneous formation datum points | |
CN115811788B (en) | D2D network distributed resource allocation method combining deep reinforcement learning and unsupervised learning | |
Pak et al. | A distributed target localization algorithm for mobile adaptive networks | |
CN114298244A (en) | Decision control method, device and system for intelligent agent group interaction | |
KR102158465B1 (en) | Method for estimating position of vehicle and the vehicle | |
CN117250998B (en) | Multi-agent system tracking control method, agent and program product | |
CN114167896B (en) | Unmanned aerial vehicle cluster control method for artificial potential energy field based on state prediction |
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 |