CN113316156B - Intelligent coexistence method on unlicensed frequency band - Google Patents

Intelligent coexistence method on unlicensed frequency band Download PDF

Info

Publication number
CN113316156B
CN113316156B CN202110579013.9A CN202110579013A CN113316156B CN 113316156 B CN113316156 B CN 113316156B CN 202110579013 A CN202110579013 A CN 202110579013A CN 113316156 B CN113316156 B CN 113316156B
Authority
CN
China
Prior art keywords
base station
small base
laa
action
wifi
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110579013.9A
Other languages
Chinese (zh)
Other versions
CN113316156A (en
Inventor
裴二荣
陶凯
徐成义
宋珈锐
黄一格
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing University of Post and Telecommunications
Original Assignee
Chongqing University of Post and Telecommunications
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Chongqing University of Post and Telecommunications filed Critical Chongqing University of Post and Telecommunications
Priority to CN202110579013.9A priority Critical patent/CN113316156B/en
Publication of CN113316156A publication Critical patent/CN113316156A/en
Application granted granted Critical
Publication of CN113316156B publication Critical patent/CN113316156B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/14—Spectrum sharing arrangements between different networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/06—Testing, supervising or monitoring using simulated traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/0268—Traffic management, e.g. flow control or congestion control using specific QoS parameters for wireless networks, e.g. QoS class identifier [QCI] or guaranteed bit rate [GBR]
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention relates to an intelligent coexistence method on an unlicensed frequency band, belonging to the technical field of wireless communication. The invention comprises the following steps: s1: designing an LAA small base station action set, a reward function and a state set; s2: initializing environment, LAA small base station obtaining initial state value st(ii) a S3: according to the epsilon-greedy strategy, the LAA small base station randomly selects the action with the probability epsilon, and selects the action a corresponding to the maximum Q value with the probability (1-epsilon)t(ii) a S4: performing action atThen, LAA small base station obtains environment reward value rtAnd an experience sample obtained by one-time interaction is used<st，at，rt，st+1>Storing in a memory pool, and entering the next state st+1(ii) a S5: feeding the experience samples to an estimation value network by the LAA small base station in a batch mode to update the weight theta of the neural network and adjust the estimation mode of the Q value; s6: repeating the steps S5-S7 until the optimal access strategy pi is obtained*(s)。

Description

Intelligent coexistence method on unlicensed frequency band
Technical Field
The invention belongs to the technical field of wireless communication, and relates to an intelligent coexistence method on an unlicensed frequency band.
Background
In recent years, the popularity of smart mobile devices and their various applications has led to an explosive increase in mobile traffic demand, which places a heavy burden on current cellular networks. On the other hand, the capacity of the network is limited by the scarce and expensive licensed spectrum, and on the contrary, the unlicensed band has abundant spectrum resources and the utilization rate is not high. To increase cellular capacity, deployment of LTE networks in unlicensed bands is considered a promising technology to support explosively increasing mobile traffic.
However, there are already well-developed wireless technologies in the existing unlicensed frequency band, so when selecting the unlicensed frequency spectrum for the construction of the communication network, the problem of coexistence with the wireless technologies in the existing unlicensed frequency band must be considered. In spectrum resource allocation, considering that LTE is a spectrum allocation scheme using centralized scheduling, and WiFi is a contention-based spectrum allocation scheme, there is a great difference in spectrum resource allocation between them. If the LTE service is directly operated on the unlicensed frequency band, the WiFi performance will be reduced to an extremely low level, which affects the experience of the WiFi user. Therefore, under the condition of ensuring the service quality of the WiFi network, the WiFi network and the WiFi network coexist harmoniously to form a research hotspot.
The deep reinforcement learning integrates the characteristics of reinforcement learning modeless and the capability of deep learning for processing big data, and makes good progress in the fields of intelligent decision making, unmanned driving, edge unloading and the like. Inspired by deep reinforcement learning, the invention aims to introduce a reinforcement learning strategy into an authorization-free spectrum sharing scheme, so that an LAA small base station monitors a dynamic environment in real time, potential important data and information are mined, and the LAA small base station can learn a spectrum resource access scheme in a self-organizing manner, thereby realizing fair and efficient coexistence of a cellular network and a WiFi network.
Abundant bandwidth resources exist near the 5GHz unlicensed frequency band, and the development of the 5G technology is facilitated by expanding the mobile communication technology from the licensed frequency band to the unlicensed frequency band. By using the LAA technology as a basis, a network with higher transmission rate, low time delay and low power consumption is constructed, and the connection requirement of next generation mobile communication mass equipment can be further met.
Disclosure of Invention
In view of this, the present invention provides an intelligent coexistence method in an unlicensed frequency band, which maximizes the throughput of the LTE network and improves the utilization rate of the unlicensed frequency band while ensuring the quality of service of the WiFi network.
In order to achieve the purpose, the invention provides the following technical scheme:
1. an intelligent coexistence method on an unlicensed frequency band comprises the following steps:
s1: designing an LAA small base station action set, a reward function and a state set;
s2: initializing environment, and obtaining initial state value s by LAA small base stationt；
S3: according to the epsilon-greedy strategy, the LAA small base station randomly selects the action with the probability epsilon, and selects the action a corresponding to the maximum Q value with the probability (1-epsilon)t；
S4: performing action atThen, LAA small base station acquiresEnvironmental reward value rtAnd an experience sample < s obtained by one-time interactiont,at,rt,st+1Is stored in the memory pool and then enters the next state st+1；
S5: feeding the experience samples to an estimation value network by the LAA small base station in a batch mode to update the weight theta of the neural network and adjust the estimation mode of the Q value;
s6: repeating the steps S3-S5 until the optimal access strategy pi is obtained*(s)。
2. Further, in step S1, we consider the WiFi network as a random environment, model the access problem of the LAA small cell as a markov decision process, and introduce a DRL to solve the problem.
For the coexistence system, our objective is to maximize throughput of the LAA small base station while fully protecting WiFi network performance, so as to improve spectrum utilization. We consider the expected WiFi packet delivery rate n '/n to indicate whether WiFi network performance is protected, where n' indicates each TFThe number of successfully transmitted WiFi packets in the group, n represents each TFThe number of WiFi data packets to be transmitted. Therefore, this coexistence problem is described as the mathematical formula:
max TSBS/TF
Figure BDA0003085339000000021
TSBS/TFexpressed as the throughput of the normalized LAA cell site,
Figure BDA0003085339000000022
indicating that the expected WiFi packet arrival rate is not below the guaranteed rate
Figure BDA0003085339000000023
Therefore, in order to correctly guide the LAA cell to adjust the access policy as desired, the reward function is set to:
Figure BDA0003085339000000024
secondly, the actions of the LAA small base station are actually the combination of the access condition, the access length and the dormancy duration, which are respectively Td、TL、TqIndicates that action atIs represented by [ Td,TL,Tq]。TLAnd TqThere is a respective maximum minimum in ms, TdThe threshold value of the dummyacket is expressed in slots.
The status s is the basis for the decision and should contain enough information to indicate the operation of the WiFi network. The LAA small base station monitors channel activity by energy detection and other modes to collect WiFi activity, obtains the quantity of WiFi data packets successfully sent in each LAA frame, the quantity of conflicts and the quantity of idle time slots, and respectively uses ns，ncAnd TIIs shown, in addition to this, should also be included in the state stAction taken atAnd the resulting prize rtBecause they contain rules that implicitly evaluate actions. I.e. state st+1Expressed as:
st+1＝〈TI,ns,nc,at,rt〉
3. further, in step S2, the environment status is initialized, and the LAA small cell acquires the current status value StAnd transmitting the data into an estimation value network, and predicting the Q value of each action.
4. Further, in step S3, the LAA small base station selects an action a corresponding to the maximum Q value with a probability (1-epsilon) according to the epsilon-greedy policytRandomly selecting an action a from the set of actions with a probability epsilont。
5. Further, in step S4, the LAA small cell operates atAfter acting on the environment, an in-state s is obtainedtLower execution action atIs awarded a value
Figure BDA0003085339000000031
The environmental state will also be from stIs updated to st+1。LAAThe small base station will experience the sample of each interaction
Figure BDA0003085339000000032
Stored in a memory pool D ═ e1,e2,...etIn (j) };
6. further, in step S5, the LAA cell site feeds the empirical samples to the estimation value network in a batched manner, and applies a gradient descent algorithm to minimize the loss function L (θ), thereby obtaining the weight θ of the estimation value networktIs updated to thetat+1。
Wherein the loss function is defined as:
Figure BDA0003085339000000033
Figure BDA0003085339000000034
7. further, in step S6, repeating steps S3-S5, the LAA cell site continuously interacting with the environment, and repeatedly training the neural network and updating the weight θ thereof by using the experience sample until the optimal weight θ is obtained*Finally, the optimal access strategy pi is obtained*(s)。
The invention has the effective effects that: under the scene that LTE and WiFi networks share the unlicensed frequency band, a new frame-based MAC protocol is designed for a coexistence system, and a spectrum sharing framework for the protocol is provided. The method mainly considers that the LTE service running on the unlicensed frequency band in the existing coexistence scheme seriously influences the WiFi performance, models the LAA small base station access problem as a Markov decision process, and provides a coexistence scheme based on DRL. According to the coexistence scheme, the LAA small base station monitors a dynamic environment in real time, and learns the spectrum resource access strategy in a self-organizing manner through the monitored information, so that the throughput of the LAA small base station is maximized on the premise of ensuring the service quality of the WiFi network, and the fair and efficient coexistence of the cellular network and the WiFi network is realized.
Drawings
In order to make the object, technical scheme and beneficial effect of the invention more clear, the invention provides the following drawings for explanation:
fig. 1 is a diagram illustrating a coexistence model of LTE and WiFi networks according to an embodiment of the present invention;
FIG. 2 is a frame structure diagram of a sharing framework according to an embodiment of the present invention;
fig. 3 is an access flow chart of an LAA small cell according to an embodiment of the present invention;
FIG. 4 is a DRL algorithm model diagram according to an embodiment of the present invention.
Detailed Description
Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
The invention provides a coexistence method of LAA-LTE and WiFi networks based on DRL (distributed resource locator) aiming at the coexistence problem of LTE based on an LBT (local binary transmission) mechanism on a WiFi unlicensed frequency band (5 GHz). Compared with the traditional coexistence algorithm, the coexistence algorithm based on the DRL avoids a large amount of information exchange between the two systems, so that the LAA small base station analyzes the flow change of the WiFi network according to the information acquired by the monitoring channel, flexibly adjusts the access strategy, and maximizes the self throughput on the premise of ensuring the performance of the WiFi network.
Considering the presence of one LAA small base station and multiple WiFi access points in the coexistence scenario, the network model is shown in fig. 1.
Firstly, a frame-based MAC protocol is designed for the coexistence system, and a spectrum sharing framework for the protocol is proposed, the frame structure of the proposed protocol is shown in fig. 2, and the access flow chart of the LAA small cell is shown in fig. 3.
For the coexistence system, the aim is to maximize the throughput of the LAA small base station under the condition of fully protecting the WiFi network performance so as to improve the spectrum utilization rate. We consider the expected WiFi packet delivery rate n '/n to indicate whether WiFi network performance is protected, where n' indicates each TFThe number of successfully transmitted WiFi packets in the group, n represents each TFThe number of WiFi data packets to be transmitted. Thus, this coexistence problem is described as the mathematical formula:
max TSBS/TF
Figure BDA0003085339000000051
TSBS/TFexpressed as the throughput of the normalized LAA cell site,
Figure BDA0003085339000000052
indicating that the expected WiFi packet arrival rate is not below the guaranteed rate
Figure BDA0003085339000000053
Let us consider a WiFi network as a stochastic environment, model the access problem of LAA small base stations as a markov decision process, and introduce a DRL to solve the problem. And the LAA small base station is regarded as an intelligent agent, and the intelligent agent monitors the change of the WiFi flow in the channel in real time to adjust the access strategy of the intelligent agent so as to realize efficient and harmonious coexistence.
In a coexisting network, LAA small cell users coexist harmoniously with WiFi users. Based on the working principle of the deep Q learning algorithm, an action set A and a reward function are set
Figure BDA0003085339000000054
And a state s.
According to the access flow of the LAA small cell, the action set can be expressed as all combinations of the access condition, the access length and the sleep duration, which are respectively represented by Td、TL、TqIndicates that action atIs represented by [ Td,TL,Tq]，at∈A。
The status s is the basis for the decision and should contain enough information to indicate the operation of the WiFi network. The LAA small base station monitors channel activity by energy detection and other modes to collect WiFi activity, obtains the quantity of WiFi data packets successfully sent in each LAA frame, the quantity of conflicts and the quantity of idle time slots, and respectively uses ns，ncAnd TIIs shown, in addition to this, should also be included in the state stAction a taken at the timetAnd the resulting prize rtBecause they contain rules that implicitly evaluate actions. I.e. state st+1Expressed as:
st+1＝<TI,ns,nc,at,rt>
reward function
Figure BDA0003085339000000055
The method aims to better guide the LAA small base station to find the optimal access strategy. It is expected that the LAA small base station meets the guarantee rate of the WiFi data packet arrival rate
Figure BDA0003085339000000056
In the above case, as many channels are accessed as possible, and then the reward function is defined as:
Figure BDA0003085339000000057
initializing the environment state, monitoring the environment by the LAA small base station, and acquiring the current state value stAnd inputting it into a nonce network, the nonce network being based on stThe Q value for each action is estimated.
Then according to the epsilon-greedy strategy, the LAA small base station randomly selects an action a from the action set according to the probability epsilontThe action a corresponding to the maximum Q estimation value is selected with probability (1-epsilon)t。
Will act atAfter acting on the environment, the LAA small base station gets a message about action atIs awarded
Figure BDA0003085339000000061
The state of the environment will also be from stIs updated to st+1. Each interaction between the LAA small base station and the environment obtains an experience sample
Figure BDA0003085339000000062
The LAA small base station stores the transfer sample in a memory poolD＝{e1,e2,...etIn (c) }.
After obtaining a certain amount of experience samples, the LAA small base station feeds the experience samples to the estimation value network in a batch mode, a gradient descent algorithm is applied to minimize a loss function L (theta), and the weight theta of the estimation value network is obtainedtIs updated to thetat+1。
Wherein the loss function is defined as:
Figure BDA0003085339000000063
Figure BDA0003085339000000064
the LAA small base station continuously interacts with the environment, the neural network is repeatedly trained by using experience samples, and the weight theta of the neural network is updated until the optimal weight theta is obtained*Finally, the optimal access strategy pi is obtained*(s)。

Claims (1)

1. An intelligent coexistence method in an unlicensed frequency band is characterized in that: the method comprises the following steps:
s1: designing an LAA small base station action set, a reward function and a state set; firstly, designing an intelligent MAC protocol for an LTE system on the basis of the existing coexistence mechanism, and providing a spectrum sharing framework based on the protocol; whether WiFi network performance is protected or not is indicated using an expected WiFi packet delivery rate of n '/n, where n' represents each TFThe number of successfully transmitted WiFi packets in the group, n represents each TFThe number of WiFi data packets to be transmitted; thus, this coexistence issue is described as:
max TSBS/TF
Figure FDA0003629760530000011
wherein T isSBS/TFRepresenting the throughput of the normalized LAA cell site,
Figure FDA0003629760530000012
indicating that the expected WiFi packet arrival rate is not below the guaranteed rate
Figure FDA0003629760530000013
Regarding the coexistence system as a random environment, modeling the spectrum access problem of the LAA small base station as a Markov decision process, and mapping key elements in the MDP from the problem to be solved: action set A, reward function
Figure FDA0003629760530000014
And a state s;
wherein the reward function is defined as:
Figure FDA0003629760530000015
according to the proposed MAC protocol, the actions of the LAA cell are actually a combination of access conditions, access length and sleep duration, and are respectively denoted by Td、TL、TqIndicates that action atIs represented by [ Td,TL,Tq]；TLAnd TqThere are respective maximum and minimum values, whose basic unit is millisecond, TdIndicating the interception period of the LAA small base station, and the unit is a time slot;
state s is the basis for the decision, which contains information to indicate the operation of the WiFi network; the LAA small base station monitors the channel activity through an energy detection technology to collect WiFi activity, obtains the quantity of WiFi data packets successfully sent in each LAA frame, the quantity of conflicts and the quantity of idle time slots, and respectively uses ns，ncAnd TITo represent; in addition to this, the state stAction a taken at the timetAnd the resulting prize rt(ii) a Then state st+1Expressed as:
st+1＝<TI,ns,nc,at,rt>
s2: initializing an environment state, and acquiring current environment information s by the LAA small base station through monitoring the environment statetAnd transmits it as input to the current value network according to stEstimating a Q value of each action;
s3: according to the epsilon-greedy strategy, the LAA small base station selects the action a corresponding to the maximum Q value with the probability (1-epsilon)tRandomly selecting an action a with a probability epsilont；
S4: LAA small cell executes action atInteracting with the environment to obtain a corresponding action atIs given a prize value R (a)t,st) The state of the environment is from stIs updated to st+1(ii) a Each interaction between the LAA small base station and the environment obtains an experience sample et＝<st,at,R(at,st),st+1>The LAA cell stores the experience sample in a memory pool D ═ e1,e2,...etIn (j) };
s5: after obtaining the experience samples, the LAA small base station feeds the experience samples to the estimation value network in a batch mode, and minimizes a loss function L (theta) by applying a gradient descent algorithm, and weights theta of the estimation value networktIs updated to thetat+1；
Wherein the loss function is defined as:
Figure FDA0003629760530000021
Figure FDA0003629760530000022
s6: repeating the steps S3-S5, continuously interacting the LAA small base station with the environment, repeatedly training the neural network by using the experience samples and updating the weight theta until the optimal weight theta is obtained*Finally, the optimal access strategy pi is obtained*(s) straightThe target state is reached.
CN202110579013.9A 2021-05-26 2021-05-26 Intelligent coexistence method on unlicensed frequency band Active CN113316156B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110579013.9A CN113316156B (en) 2021-05-26 2021-05-26 Intelligent coexistence method on unlicensed frequency band

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110579013.9A CN113316156B (en) 2021-05-26 2021-05-26 Intelligent coexistence method on unlicensed frequency band

Publications (2)

Publication Number Publication Date
CN113316156A CN113316156A (en) 2021-08-27
CN113316156B true CN113316156B (en) 2022-07-12

Family

ID=77375175

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110579013.9A Active CN113316156B (en) 2021-05-26 2021-05-26 Intelligent coexistence method on unlicensed frequency band

Country Status (1)

Country Link
CN (1) CN113316156B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113810910B (en) * 2021-09-18 2022-05-20 大连理工大学 Deep reinforcement learning-based dynamic spectrum sharing method between 4G and 5G networks

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107426772A (en) * 2017-07-04 2017-12-01 北京邮电大学 A kind of dynamic contention window method of adjustment, device and equipment based on Q study
CN109039374A (en) * 2018-07-02 2018-12-18 北京交通大学 The method that LTE-LAA system and WiFi system coexist is realized in unauthorized frequency range

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2800283B1 (en) * 2013-04-30 2019-07-10 Alcatel Lucent Non-linear precoder with separate modulo decision
CN106332094B (en) * 2016-09-19 2019-09-10 重庆邮电大学 The Dynamic Duty Cycle coexistence method of unauthorized frequency range LTE-U and WiFi system based on Q algorithm
CN106535355B (en) * 2016-11-04 2019-12-06 北京邮电大学 channel access method and system
CN108924944B (en) * 2018-07-19 2021-09-14 重庆邮电大学 LTE and WiFi coexistence competition window value dynamic optimization method based on Q-learning algorithm
CN109302709B (en) * 2018-09-14 2022-04-05 重庆邮电大学 Mobile edge computing-oriented vehicle networking task unloading and resource allocation strategy
CN110035559B (en) * 2019-04-25 2023-03-10 重庆邮电大学 Intelligent competition window size selection method based on chaotic Q-learning algorithm
CN111246502B (en) * 2020-01-09 2022-04-29 重庆邮电大学 Energy threshold dynamic optimization method based on Q learning

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107426772A (en) * 2017-07-04 2017-12-01 北京邮电大学 A kind of dynamic contention window method of adjustment, device and equipment based on Q study
CN109039374A (en) * 2018-07-02 2018-12-18 北京交通大学 The method that LTE-LAA system and WiFi system coexist is realized in unauthorized frequency range

Also Published As

Publication number Publication date
CN113316156A (en) 2021-08-27

Similar Documents

Publication Publication Date Title
Sharma et al. Toward massive machine type communications in ultra-dense cellular IoT networks: Current issues and machine learning-assisted solutions
WO2021233053A1 (en) Computing offloading method and communication apparatus
Sun et al. Joint offloading and computation energy efficiency maximization in a mobile edge computing system
Szott et al. Wi-Fi meets ML: A survey on improving IEEE 802.11 performance with machine learning
Azari et al. On the latency-energy performance of NB-IoT systems in providing wide-area IoT connectivity
CN112954651B (en) Low-delay high-reliability V2V resource allocation method based on deep reinforcement learning
Azari et al. User traffic prediction for proactive resource management: Learning-powered approaches
Han et al. Reinforcement learning for efficient and fair coexistence between LTE-LAA and Wi-Fi
CN113490184B (en) Random access resource optimization method and device for intelligent factory
Li et al. Predictive pre-allocation for low-latency uplink access in industrial wireless networks
Tan et al. Deep reinforcement learning for the coexistence of LAA-LTE and WiFi systems
Elsayed et al. Deep reinforcement learning for reducing latency in mission critical services
CN110035559B (en) Intelligent competition window size selection method based on chaotic Q-learning algorithm
Zhao et al. Deep reinforcement learning aided intelligent access control in energy harvesting based WLAN
He et al. Reinforcement learning-based computing and transmission scheduling for LTE-U-enabled IoT
Xu et al. Resource allocation for UAV-aided energy harvesting-powered D2D communications: A reinforcement learning-based scheme
CN113316156B (en) Intelligent coexistence method on unlicensed frequency band
Ashtari et al. Knowledge-defined networking: Applications, challenges and future work
Balcı et al. Massive connectivity with machine learning for the Internet of Things
Wang et al. Energy-efficient and delay-guaranteed routing algorithm for software-defined wireless sensor networks: A cooperative deep reinforcement learning approach
Tian et al. Multi-objective surrogate modeling for real-time energy-efficient station grouping in IEEE 802.11 ah
Chen et al. An mdp model-based initial strategy prediction method for lorawan
Abuzainab et al. Robust Bayesian learning for wireless RF energy harvesting networks
Salama et al. Flcc: Efficient distributed federated learning on iomt over csma/ca
CN113316174A (en) Intelligent access method for unlicensed spectrum

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