CN112969240B

CN112969240B - Heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision

Info

Publication number: CN112969240B
Application number: CN202110138142.4A
Authority: CN
Inventors: 张周; 谢佳; 王彤彤; 闫野; 李梦烁; 王一竹; 马丕明
Original assignee: Tianjin (binhai) Intelligence Military-Civil Integration Innovation Center; National Defense Technology Innovation Institute PLA Academy of Military Science
Current assignee: Tianjin (binhai) Intelligence Military-Civil Integration Innovation Center; National Defense Technology Innovation Institute PLA Academy of Military Science
Priority date: 2021-02-01
Filing date: 2021-02-01
Publication date: 2022-09-20
Anticipated expiration: 2041-02-01
Also published as: CN112969240A

Abstract

Aiming at the defects of high complexity and low frequency band utilization rate of the existing heterogeneous wireless distributed network in the access process, the invention discloses a heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision. The method is used for carrying out distributed channel intelligent detection and access decision in a heterogeneous wireless cooperative network consisting of a plurality of signal source-signal sink pairs and a plurality of signal amplification forwarding type relay nodes, a plurality of signal sources independently send data packets in a distributed mode to compete for channels, the successful signal source-signal sink pairs compete for channel states through pure threshold decision, the optimal access mode is dynamically selected, balance between relay channel detection overhead and channel access income is searched, and system average throughput performance of the whole heterogeneous wireless cooperative network is optimal. Through simulation result verification, the method can realize effective networking of the heterogeneous wireless distributed cooperative network, and improve the average throughput of the system and the utilization rate of network spectrum resources.

Description

Heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision

Technical Field

The invention relates to the technical field of wireless communication, in particular to a heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision.

Background

In recent years, wireless communication networks have been developed unprecedentedly, and the wireless communication network technology has made higher demands on network performance (including system reliability and throughput) in practical scenarios while meeting various application requirements such as mobile networks and internet of things. Due to the scarcity of frequency spectrum resources and time-varying channel conditions, in order to further improve the performance of the wireless network, an opportunistic scheduling method is proposed and used for realizing the joint optimization of the multi-layer wireless network. The opportunistic scheduling method enables the MAC layer of the user to carry out opportunistic scheduling channel access among a plurality of users under the condition of knowing the physical layer information, can effectively utilize multi-user diversity and time diversity, and greatly improves the overall performance of the system.

The current wireless cooperative network mainly includes a centralized cooperative network and a distributed cooperative network. The centralized network collects global Channel State Information (CSI) from all users by deploying a central controller, and schedules a user access channel with the best channel condition, significantly improving network performance by using multi-user diversity. The distributed opportunistic scheduling is still under investigation in the initial phase, each user experiences channel contention due to the distributed nature of multiple users and the time-varying nature of the wireless channel, and only local information is available, and the winner user must make channel access decisions based on this limited information. Aiming at the problem, an optimal stopping strategy under a distributed wireless network and an optimal strategy of a pure threshold structure by utilizing an optimal stopping theory are provided, the optimal access problem under different channel constraints is also researched, the concept of the distributed wireless cooperative network is provided, the cooperative transmission is utilized to realize relay diversity, and a winning signal source-signal sink pair can select a mode with better channel quality to access after detecting a relay channel, so that the network performance is obviously improved.

The existing centralized network depends on a central controller to process and optimize information of the whole network, when the central controller fails, the network cannot operate, and meanwhile, the central controller needs to know channel state information of the whole network, so that the signaling overhead is large, the network is only suitable for networks with small scale and small user number, and when the number of users increases, the signaling overhead is increased rapidly, so that the frequency spectrum utilization rate of the network is reduced.

The existing distributed network technology only detects a fixed number of relays before the cooperative channel is accessed, and when the number of deployed relays is large, serious signaling overhead can be caused, so that the system throughput is reduced; in order to reduce the complexity of the problem, the characteristics of a wireless channel are usually limited, and the applicability to a general wireless environment is limited on the assumption that the characteristics have reciprocity; most studies are only directed to homogeneous wireless networks, but in a practical distributed wireless network environment, the direct channel and the relay channel usually experience different channel fading, with heterogeneity. In general, the existing heterogeneous wireless distributed cooperative network has the disadvantages of high complexity and low frequency band utilization rate in the access process,

disclosure of Invention

Aiming at the defects of high complexity and low frequency band utilization rate of the existing heterogeneous wireless distributed cooperative network in the access process, the invention discloses an intelligent channel access method based on pure threshold decision under the condition of the heterogeneous wireless distributed cooperative network. The method is used for carrying out distributed channel intelligent detection and access decision in a heterogeneous wireless cooperative network consisting of a plurality of signal source-signal sink pairs and a plurality of signal amplification forwarding type relay nodes, a plurality of signal sources independently send data packets in a distributed mode to compete for channels, the successful signal source-signal sink pairs compete for channel states through pure threshold decision, the optimal access mode is dynamically selected, balance between relay channel detection overhead and channel access income is searched, and system average throughput performance of the whole heterogeneous wireless cooperative network is optimal. Through simulation result verification, the method can realize effective networking of the heterogeneous wireless distributed cooperative network, and improve the average throughput of the system and the utilization rate of network spectrum resources.

The invention discloses a heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision. And at the beginning of each competition time slot, the multiple information sources independently send RTS packets in a distributed mode to compete for the channel, if and only if only one information source sends the RTS packet, the information source competes for the channel successfully to obtain a channel access opportunity, and the information source is a winning information source. Considering the cooperative performance of the distributed cooperative networking, the winning information source makes an optimal decision according to the current limited local information so as to maximize the system average throughput of the network; and if the channel condition of the winning information source is too poor, the winning information source gives up the access opportunity and participates in the channel competition again, otherwise, the winning information source is directly accessed or accessed in an auxiliary manner by a relay.

The heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision-making comprises the steps of firstly calculating global parameters based on channel statistical parameters and classifying communication pairs, wherein the I-type communication pairs can carry out threshold comparison according to the current channel state, and direct connection/abandonment/further detection relay is selected; class ii communication pairs can only choose direct connections/abandon. After channel competition succeeds, a winning information sink obtains SNR of a direct connection channel from the winning information source, a corresponding access strategy is selected according to a classification interval of the SNR, if the communication pair belongs to a classification I, whether relay is to be detected further is judged through a threshold value, if the communication pair belongs to a classification I, threshold value comparison of the number of detected relays is carried out, the optimal number of detected relays is determined, and then access transmission is carried out in a mode of better channel rate according to the obtained CSI of the relay channel; then the information sink informs the access decision to all other information sources by sending a CTS packet, if the decision result is an access channel, the information source of the information sink selects a direct connection/relay channel for transmission according to the decision result, at the moment, other information sources stop sending competitive data packets, and all the information sources restart the next round of channel competition until the transmission of a winning information source-information sink pair is finished; and if the decision result is that the access is abandoned, all the information sources continue to send RTS packets to compete for the channel when the next competition time slot starts.

The heterogeneous wireless distributed network comprises K source-sink pairs (namely, a source S) ₁ ,...,S _i ,...,S _K And a signal sink D ₁ ,...,D _i ,...,D _K ) L amplify-and-forward relay nodes (i.e. relays R) ₁ ,R ₂ ,...R _L ). The selectable access modes of different source-sink pairs include three types: accessing a direct connection channel when the relay node is not detected; accessing a relay auxiliary channel; the direct connection channel access is realized when a certain number of relay nodes are detected, and the mode detects the certain number of relay nodes, but the channel condition of the direct connection channel is found to be better, so that the direct connection channel is selected for access; when all the access channel conditions are poor, the information source selects to give up the access opportunity and perform channel competition again, so that the transmission time of other communication pairs with better channel conditions is increased, and the average throughput of the whole distributed network system is increasedIs large. RTS (request-to-send), which indicates a request-to-send packet, is a data packet in the channel aware access protocol, and is used for a sending node user to detect the channel occupancy and estimate the channel quality. Csi (channel state information), which represents channel state information and data information reflecting real-time conditions of a radio channel. CTS (clear-to-send), which indicates clear-to-send packet, is a data packet in the channel aware access protocol for the receiving node to respond to the sending node.

Specific definitions and assumptions of parameters used in the present invention include:

k signal source-signal sink pairs, wherein the signal source index number is S ₁ ,...,S _i ,...,S _K Sink index number D ₁ ,...,D _i ,...,D _K L relay nodes and R index number ₁ ,R ₂ ,...R _L (ii) a All nodes compete for the channel in a distributed manner on a micro-slot basis, assuming time synchronization. When each minislot starts, all sources have the same probability p ₀ Independently sending a channel competition RTS packet, if and only if only one source sends the RTS packet in the same micro-slot, the source is a winning source, the process is called a successful channel competition, and the process from the beginning of the channel competition to the appearance of the winning source is defined as an observation. For each observation, the time to be spent before the winning source appears is random, and since each channel contention is independent, the total number of channel contention experienced for a single observation follows a parameter Kp ₀ (1-p ₀ ) ^K-1 The mean time of a single observation is:

wherein tau is _RTS Indicating the time of transmission of RTS packets, τ _CTS Indicates the time of transmission of the CTS packet, delta indicates the duration of the idle slot, and the time of collision is represented by tau _RTS And (4) showing.

Considering a random channel fading model with statistical properties, from the ith source S _i To its sink D _i The SNR of the direct-connection channel of (i), i.e. the ith source-sink pair channel, is represented by γ _i SNR of a first hop channel from an ith source to a jth relay and a second hop channel from the jth relay to an ith sink are respectively expressed as

And

all direct-connected channels and relay channels are subject to Rayleigh fading models, and the SNR variable gamma of the channels _i 、

And

are all subject to an exponential random distribution, which is expected to be respectively

And

the channel noise follows a gaussian distribution of normalized variance.

The channel coherence time is recorded as τ _d Probing a set of relay nodes

The signaling interaction time of all the relay nodes is recorded as

I | represents a modulo operation,

representing a set of relay nodes

The number of the relay nodes is the winning signal source-information sink pair if the detection relay node set is selected

And for the probing relay node set

After detection, the winning information source is accessed into the channel for data transmission in the time of

Achievable rate of the direct connection channel is R _d (γ _i )＝log ₂ (1+γ _i ) The receiving end signal under the relay auxiliary channel access comprises a direct connection channel signal, a relay two-hop channel signal and an information source S _i In sounding relay node set

The maximum channel SNR for the source-sink pair to assist channel transmission through the best single relay in the set of probed relay nodes is then obtained as

The corresponding channel achievable rate is

Modeling a channel observation process after single channel competition is successful into two sub-observation processes, wherein for the k-th observation, the obtained observation information is phi for the 2k-1 and 2 k-th sub-observation processes, and specifically for the 2k-1 sub-observation process _k ＝s(k),γ _s(k) (k),t _s (k) Where s (k) denotes the index number of the winning source in the k-th observation, γ _s(k) (k) SNR, t representing a direct connection channel between a winning source s (k) and its sink _s (k) Representing the channel contention time of the kth observation; then, a winning information sink d (k) corresponding to a winning information source s (k) selects a further detection relay channel, then a 2 k-time sub-observation is carried out, and observation information obtained by the 2 k-time sub-observation is represented as

Wherein

Is the set of relay nodes, γ, probed in the 2 k-th sub-observation _s(k),j (k) And gamma _j,d(k) (k) First and second hop channel SNRs from source s (k) to relay node j and relay node j to sink d (k), respectively.

Determining the time of accessing the channel by a winning information source, carrying out observation path modeling on the sub-observation process of channel competition at the previous time, and observing the path model when the observation times | Pi | is odd

When the observation time | Pi | is even number, observing the path model

Wherein a is _k Is a binary number with a value of 0 or 1, a _k 1 indicates that after the kth channel contention, some source wins the channel and its sink gets the CSI of the direct link, a _k 0 indicates that the channel competition fails, i.e. channel collision or idle occurs,

indicating that the winning sink further decides whether and how to probe the relaying channel,

indicating that the relay channel is not probed, if the probing relay channel is selected

Representing a set of relay nodes to be probed. For the observation path pi, the cumulatively obtained observation information is denoted as B _π The gain function is denoted as Y (pi), and when the number of observations | pi | is odd,

Y(π)＝τ _d R _d (γ _s(k) ) When the observation time | pi | is an even number,

the time cost is the time spent by all sub-observation processes plus the data transmission duration, expressed as

Wherein

Is a hypothesis indicator if it is set]If the inner hypothesis is true, the value is 1, otherwise, the value is 0; accordingly, the instantaneous throughput is determined by

And (4) showing.

Based on the heterogeneous wireless distributed cooperative network model, the method provided by the invention aims to find the optimal intelligent channel access and decision method, namely the optimal strategy N ^* To average the system throughput of the network

Most preferably, wherein

Indicating expectation, sup indicates the minimum upper bound.

In the heterogeneous wireless distributed cooperative network, the method for accessing the heterogeneous wireless distributed network intelligent channel based on the pure threshold decision comprises the following specific steps:

and S1, obtaining the optimal average throughput of the system, the communication pair classification interval, the primary decision threshold and the relay number decision threshold through offline iterative computation according to the channel statistical characteristic parameters of the heterogeneous wireless network. The offline iterative computation refers to iterative computation which does not occupy network resources. In the step S1, the specific calculation procedure is,

probing relay node of ith signal source-signal sink pairPoint collection

The revenue function of (a) is defined by the expression:

in which the channel coherence time is recorded as tau _d The achievable rate of the direct connection channel is R _d (γ)＝log ₂ (1+ gamma), gamma is the signal-to-noise ratio variable of the channel, U ₀ (λ) represents the maximum average benefit of giving up access, λ is the average throughput of the system, and is a continuous random variable, and the defined expression of the benefit function is developed to obtain:

wherein

Expressing probability, expressing the definition expression of the gain function as an analytic expression according to the channel statistical characteristics and the probability distribution of the wireless heterogeneous network to obtain

Wherein, the set of relay nodes to be detected is represented as

i ₁ ,i ₂ ,...,i _J The index sequence number of the j detection relay of the i information source-information sink pair in the relay node set to be detected is represented, and the j detection relay has a sequencing relation i ₁ ＜...＜i _j ＜...＜i _J To a set of

The starting point of the summation may be any of themOne relay, therefore denoted as

That is, the sum of the relay nodes taking any point in the relay node set as the starting point is expressed as

The mean function of gamma is beta when the jth relay node is detected _j (γ)＝c _i,j α (λ, γ) -1), the corresponding threshold function

Indicating the SNR threshold at which access is given up,

E ₁ (x) Is an exponential integral function, specifically expressed as

Indicates that the SNR of the channel is gamma at the ith source-sink pair _i Probing a set of relay nodes

Temporal relay supplemental channel SNR variant

The cumulative probability distribution function of (a), specifically expressed as,

wherein

λ ^* For the optimal system average throughput, when λ is λ ═ λ, for the optimal system average throughput achievable under the current network channel conditions ^* In time, a revenue function for detecting the relay under the average throughput of the optimal system is defined

The analytical expression is as follows:

wherein,

for the

source S

_i 1,2, K and relay R _j J 1, 2.. said, L, the optimal set of relay nodes is defined to implement the maximum revenue function

Represented as a set of relay nodes

Wherein σ _i,j Indicates that mu in all relay channels when j relay nodes are detected _i,j The index number of the relay node with the smallest value,

thus is provided with

Representing the set of all relay nodes in the network.

For the source node S _i I 1, 2. -, K and

relay node R

_j 1, 2.. L, defining a threshold function based on the number of relays

And source-based threshold function

For

i

1,2, K and

j

1,2, 1 _i，j (λ ^* ，γ)＝V _i ^* (γ，j+1)-V _i ^* (γ, j), let W _i,j (λ ^* When γ) is 0, the decision threshold for solving the relay SNR value γ is { υ _i,j } _{i＝1,2,...,K,j＝1,2,...,L-1} 。

Primary decision threshold

Is defined as the revenue function V since for each fixed j _i ^* (γ, j) monotonically increases with γ, and thus, is defined

Is an equation

The unique solution of (a); likewise, the difference function τ _d (R _d (γ)-λ ^* )-V _i ^* (γ, j) monotonically increasing with γ, define

Is an equation

The unique solution of (a); for each fixed i, a primary decision threshold is defined

Wherein

For the purpose of the lower primary decision threshold,

is the upper primary decision threshold.

For the index number of the sounding relay, two sounding relay index values are defined as

j _i,l And j _i,u Respectively representing and preliminary decision thresholds

And

the number of corresponding detection relay nodes;

optimal system average throughput lambda ^* Satisfies the equation:

wherein the first term on the left τ _d R _d (γ _i )-λτ _d Representing the gain of the direct connection channel, the second term U on the left ₀ (λ)＝E[max{τ _d R _d (γ _i )-λτ _d ,U ₀ (λ),V _i ^* (γ _i )}]-λτ _o It is clear that when λ ═ λ ^* While, U ₀ (λ ^* ) 0, the third term V on the left _i ^* (γ _i ) Is a threshold function based on the information source i, and represents that when the SNR of the direct connection channel is gamma _i Detecting the maximum expected benefit of the relay channel;

for λ > 0, when

While, U ₀ The expression of (lambda) is given as,

for λ > 0, when

While, U ₀ The expression of (lambda) is given as,

wherein eta _i,0 (λ) to satisfy equation R _d (x) Lambda is the only solution.

Due to the revenue function V _i ^* (γ) increases monotonically with γ, so for λ > 0, when V _i ^* (2 ^λ When the content is less than or equal to-1) and less than or equal to 0,

when V is _i ^* (2 ^λ -1) > 0, and (c),

defining the classification interval of the information source-information sink pair according to the gain function, when V is _i ^* (2 ^λ -1) > 0 and

time, corresponding source-sink pair

Is a class I communication pair interval; otherwise

Is a class ii communication pair interval.

The average throughput lambda of the optimal system is calculated ^* And the off-line iterative computation for classifying the intra-network communication pairs comprises the following specific steps:

s11, inputting channel parameter tau _RTS 、τ _CTS 、τ _d 、

And

s12, inputting an initialization parameter, wherein k is the iteration number, and the initial value is 0 and lambda _k As a result of the kth iteration, Δ _k Is the precision of the kth iteration, with initial values of λ ₀ ＝0、Δ ₀ The convergence threshold of the iterative algorithm is 1, and epsilon is set according to the precision requirement, and the typical value is 10 ^-3 And alpha is the step length of iterative update, and the value of alpha satisfies

S13, judging the iteration precision delta of the kth time _k Whether a precision convergence threshold epsilon is reached, if delta _k If the iteration precision of the kth time does not reach the precision convergence threshold, continuing to iterate, and entering a step S14, otherwise, ending the iteration, and entering a step S17;

s14, calculating the result lambda of the (k + 1) th iteration _k+1 ，λ _k+1 ＝λ _k +αΔ _k ；

S15, calculating the communication pair classification section updated by the iteration number k as k +1

S16, calculating the iteration precision delta after the iteration times k are updated _k ，

Returning to step S13;

s17, finishing the iteration, wherein the iteration result is the optimal system average throughput with lambda ^* ＝λ _k Calculating

And

s18, outputting the average throughput lambda of the optimal system ^* And communication pair classification interval

S2, multiple sources compete for channels. Starting from a minislot of duration delta, all sources are given a probability p ₀ After sending RTS packets to contend for the channel independently, the following three situations occur:

if no information source sends RTS packet in the micro time slot, the channel is idle, and all the information sources compete for the channel in the next time slot;

if two or more information sources send RTS data packets at the same time, packet collision occurs, and all the information sources need to continuously compete in the next time slot;

if only one information source s (k) sends an RTS packet, wherein k represents that the current channel competition is the k-th successful channel competition, and s (k) is an information source index number, the information source obtains a channel access opportunity and is called as a winning information source; an information sink d (k) corresponding to a winning information source obtains a signal to noise ratio (SNR) value gamma of a direct connection channel by receiving an RTS data packet and utilizing a training sequence carried in the RTS data packet, all relay nodes obtain the SNR of a first hop relay channel from the information source to the relay nodes, and d (k) is an information sink index number;

s3, the signal sink d (k) determines the classification section to which the signal sink d (k) belongs according to the calculation result of the step S1, if so, the signal sink d (k) determines that the signal sink d (k) belongs to the classification section

For class II communication pair interval, go to step S8, otherwise

For the class i communication pair section, the process proceeds to step S4;

s4, comparing the primary decision threshold value, if the gain gamma of the direct connection channel is less than or equal to the lower primary decision threshold value

The destination d (k) abandons the access opportunity and returns to step S2; if the gain gamma of the direct connection channel is more than or equal to the upper limit primary judgment threshold value

The information destination d (k) selects a direct connection channel to perform channel access, and informs the decision result to the information source S (k), the information source S (k) performs corresponding channel access, and the step S2 is returned after the single transmission is finished; otherwise, further detecting the relay channel, and entering step S5; λ is the system average throughput;

s5, comparing the threshold value of the detection relay number, and taking the value of the index number j of the relay node from j _i,l To j _i,u ，j _i,l And j _i,u Respectively representing and preliminary decision threshold

And

the number of corresponding detection relay nodes is determined, and if the direct connection channel gain gamma belongs to a certain relay SNR value decision threshold interval (upsilon) _s(k),j ,υ _s(k),j-1 ]Wherein upsilon is _s(k),j L-1 is the relay SNR value decision threshold calculated in step S1, the signal sink further detects the relay node set

The sink d (k) sends CTS packets to the relay node to be probed and to the source,

after receiving the CTS packet, the relay node in the network replies an RTS packet to the information sink, wherein the RTS packet carries the information of the first hop relay channel, and the information sink is connectedOver-demodulating the training sequence in the RTS packet to obtain a set of probing relay nodes

SNR value of the post-optimal single-relay supplemental channel

Proceeding to step S6;

s6, if

λ ^* If the average throughput of the system is optimal, the information destination d (k) selects an access mode with higher channel rate from the direct channel access and the relay auxiliary channel access, and sends a CTS packet to the information source S (k), and the information source S (k) is enabled to perform corresponding channel access by informing the information source S (k) of an access decision result through the CTS packet, and then the step S7 is performed; otherwise, abandoning the access opportunity, and returning to the step S2;

s7, if

Then the source s (k) is subsequent

Maximum achievable rate R in direct-connected channel over time _d (gamma) transmitting information, and returning to the step S2 after the transmission is finished; otherwise, the source s (k) is at the maximum achievable rate under the relay supplemental channel

The transmission is carried out in two stages, in the first stage the source s (k) broadcasts data to its sink and all relay nodes, and in the second stage the index is

Forwards the received signal to the sink, gamma _s(k),l For the SNR value of the single relaying supplemental channel with index l, then, the sink will receive two signals from the direct channel and from the relaying supplemental channel,with a transmission time of

Returning to step S2 after the transmission is finished;

s8, if

Then the destination d (k) selects direct channel access, then the destination d (k) sends CTS to the source s (k) and all other sources, let the source s (k) follow τ _d Within time, at a maximum achievable rate R _d (gamma) data transmission on direct connection channel, all other sources at tau _d Wait in time, passing tau _d After the time, the single transmission ends, and the step returns to step S2; otherwise, the destination d (k) gives up the access opportunity, and broadcasts CTS packet to all source nodes, and all source nodes restart channel competition in the next time slot, and returns to step S2.

The invention has the beneficial effects that:

compared with a centralized cooperative network, the distributed network has higher flexibility, does not need a centralized controller and has low signaling overhead; the use of the distributed relay increases the communication range of the network and improves the overall quality of a communication link; compared with the existing distributed cooperative algorithm, the method comprises the following steps: channel competition and access of an access layer are optimized under the condition that physical layer information is known, cross-layer distributed network intelligent detection and access are achieved, and multi-user diversity and relay diversity are fully utilized; the number of the detection relay nodes is variable and can be adjusted in a self-adaptive manner along with the channel state, so that unnecessary signaling overhead of relay detection is saved; according to the actual situation of a wireless communication network, modeling is carried out on heterogeneous wireless networking, different information source-information sink pairs adopt different access strategies, and corresponding judgment thresholds depend on respective channel statistical characteristics, so that the method has high flexibility and applicability; all the decision threshold values and the global parameters can be calculated off-line under the condition of statistical information based on a wireless network channel, an iterative algorithm with linear complexity is provided, the optimal solution can be rapidly converged, the calculation time can be effectively reduced for a large-scale distributed network, and errors are not easy to occur; compared with a multi-relay transmission mode, the mode does not need time synchronization among the multiple relays, achieves low complexity, and has full diversity characteristics and higher relay efficiency. The method is simple, easy to realize engineering, and has strong robustness and applicability.

Drawings

Fig. 1 is a system model of a heterogeneous wireless distributed collaborative network.

Fig. 2 is a flowchart of a method for accessing an intelligent channel of a heterogeneous wireless distributed network based on pure threshold decision.

Fig. 3 is a flowchart of an algorithm for offline iterative computation.

Fig. 4 is a graph of the effect of coherence time on the average throughput of a system.

Fig. 5 is a graph of the impact of the number of relays on the average throughput of the system.

Fig. 6 is a graph of the revenue functions of different access methods of communication pair 1 as a function of gamma.

Fig. 7 is a graph of the revenue function for communication pair 3, and fig. 7(a) is a graph of the revenue function for different access methods; fig. 7(b) is a difference function curve of the sounding relay gain function.

Fig. 8 is a graph of the revenue function for communication versus 5: FIG. 8(a) is a graph of a revenue function for different access methods; fig. 8(b) is a difference function of the sounding relay gain function.

FIG. 9 is [ tau ] _d Curve of the difference in return function for communication pair 3 and communication pair 5 at 4ms as a function of γ: fig. 9(a) is a plot of communication versus 3; fig. 9(b) is a graph of communication versus 5.

FIG. 10 is a drawing showing

And comparing the performance of each access method with the performance of each access method when the access method is changed.

Detailed Description

For a better understanding of the present disclosure, an example is given here.

The invention discloses a heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision. And the multiple information sources distributively and independently send RTS packets to compete for the channel at the beginning of each competition time slot, and if and only if only one information source sends the RTS packet, the information source competes for the channel successfully to obtain the channel access opportunity, and the information source is the winning information source. Considering the cooperative performance of the distributed cooperative networking, the winning information source makes an optimal decision according to the current limited local information so as to maximize the system average throughput of the network; and if the channel condition of the winning information source is too poor, the winning information source gives up the access opportunity and participates in the channel competition again, otherwise, the winning information source is directly accessed or accessed in an auxiliary manner by a relay.

The heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision-making comprises the steps of firstly calculating global parameters based on channel statistical parameters and classifying communication pairs, wherein the I-type communication pairs can carry out threshold comparison according to the current channel state, and direct connection/abandonment/further detection relay is selected; class ii communication pairs can only choose direct connections/abandon. After channel competition succeeds, a winning information sink obtains SNR of a direct connection channel from the winning information source, a corresponding access strategy is selected according to a classification interval of the SNR, if the communication pair belongs to a classification I, whether relay is to be detected further is judged through a threshold value, if the communication pair belongs to a classification I, threshold value comparison of the number of detected relays is carried out, the optimal number of detected relays is determined, and then access transmission is carried out in a mode of better channel rate according to the obtained CSI of the relay channel; then the information sink informs all other information sources of the access decision by sending a CTS packet, if the decision result is an access channel, the information source of the information sink selects a direct connection/relay channel for transmission according to the decision result, at the moment, other information sources stop sending competition data packets, and all the information sources restart the next round of channel competition until the transmission of the winning information source-information sink pair is finished; and if the decision result is that the access is abandoned, all the information sources continue to send RTS packets to compete for the channel when the next competition time slot starts.

FIG. 1 is a system model, in which a heterogeneous wireless distributed collaboration network includes K source-sink pairs (i.e., source S) ₁ ,...,S _i ,...,S _K And a signal sink D ₁ ,...,D _i ,...,D _K ) L amplify-and-forward relay nodes (i.e. relays R) ₁ ,R ₂ ,...R _L ). The selectable access modes of different source-sink pairs include three types: accessing a direct connection channel when the relay node is not detected; accessing a relay auxiliary channel; the direct connection channel access is realized when a certain number of relay nodes are detected, and the mode detects the certain number of relay nodes, but the channel condition of the direct connection channel is found to be better, so that the direct connection channel is selected for access; when all the access channel conditions are poor, the information source selects to give up the access opportunity and perform channel competition again, so that the transmission time of other communication pairs with better channel conditions is increased, and the system average throughput of the whole distributed network is increased.

Most preferably, wherein

Indicating expectation, sup indicates the minimum upper bound.

In the heterogeneous wireless distributed cooperative network, the implementation flow of the method of the present invention is shown in fig. 2, and the specific steps include:

and S1, obtaining the optimal average throughput of the system, the communication pair classification interval, the primary decision threshold and the relay number decision threshold through offline iterative computation according to the channel statistical characteristic parameters of the heterogeneous wireless network. The offline iterative computation refers to iterative computation which does not occupy network resources. In step S1, the specific calculation process is that the probing relay node set of the ith source-sink pair

The revenue function of (a) is defined by the expression:

wherein

Expressing probability, expressing the definition expression of the gain function as an analytic expression according to the channel statistical characteristics and the probability distribution of the wireless heterogeneous network to obtain,

wherein, the set of relay nodes to be detected is represented as

The starting point of the summation may be any one of the relays and is therefore denoted as

The mean function of gamma is beta when the j relay node is detected _j (γ)＝c _i,j 1, corresponding threshold function

Indicating the SNR threshold at which access is dropped,

E ₁ (x) Is an exponential integral function, specifically expressed as

Temporal relay supplemental channel SNR variant

wherein

The analytical expression is as follows:

wherein,

for the

source S

Represented as a set of relay nodes

thus is provided with

Representing the set of all relay nodes in the network.

For the source node S _i I 1, 2. -, K and

relay node R

_j 1, 2.. L, defining a threshold function based on the number of relays

And source-based threshold function

For

i

1,2, K and

j

1,2, 1Number W _i,j (λ ^* ,γ)＝V _i ^* (γ,j+1)-V _i ^* (γ, j), let W _i,j (λ ^* When γ) is 0, the decision threshold for solving the relay SNR value γ is { υ _i,j } _{i＝1,2,...,K,j＝1,2,...,L-1} 。

Primary decision threshold

Is an equation

The unique solution of (a); similarly, the difference function τ _d (R _d (γ)-λ ^* )-V _i ^* (γ, j) monotonically increasing with γ, define

Is an equation

Wherein

For the purpose of the lower primary decision threshold,

is an upper primary decision threshold;

j _i,l And j _i,u Respectively representing and preliminary decision threshold

And

the number of corresponding detection relay nodes;

optimal system average throughput lambda ^* Satisfies the equation:

for λ > 0, when

While, U ₀ The expression of (lambda) is as follows,

for λ > 0, when

While, U ₀ The expression of (lambda) is as follows,

wherein eta is _i,0 (λ) to satisfy equation R _d (x) Lambda is the only solution.

when V is _i ^* (2 ^λ -1) > 0, and (c) the reaction solution,

time, corresponding source-sink pair

Is a class I communication pair interval; otherwise

Is a class ii communication pair interval. The average throughput lambda of the optimal system is calculated ^* And performing offline iterative computation for classifying intra-network communication pairs, wherein the flow chart of the algorithm is shown in fig. 3.

if no source sends RTS packet in the micro time slot, the channel is idle, and all sources compete for the channel in the next time slot; if two or more information sources send RTS data packets at the same time, packet collision occurs, and all the information sources need to continuously compete in the next time slot; if only one information source s (k) sends an RTS packet, wherein k represents that the current channel competition is the k-th successful channel competition, and s (k) is an information source index number, the information source obtains a channel access opportunity and is called as a winning information source; an information sink d (k) corresponding to a winning information source obtains a signal to noise ratio (SNR) value gamma of a direct connection channel by receiving an RTS data packet and utilizing a training sequence carried in the RTS data packet, all relay nodes obtain the SNR of a first hop relay channel from the information source to the relay nodes, and d (k) is an information sink index number;

For class II communication pair interval, go to step S8, otherwise

For the class i communication pair section, the process proceeds to step S4;

The information destination d (k) selects the direct connection channel to perform channel access, and informs the decision result to the information source S (k), the information source S (k) performs corresponding channel access, and returns to the step S2 after the single transmission is finished; otherwise, further detecting the relay channel, and entering step S5; λ is the system average throughput;

And

the number of corresponding detection relay nodes is determined, and if the direct connection channel gain gamma belongs to a certain relay SNR value decision threshold interval (upsilon) _s(k),j ,v _s(k),j-1 ]Wherein v is _s(k),j If j is 1,2,3.. L-1 is the relay SNR value decision threshold calculated in step S1, the sink further detects the relay node set

The sink d (k) sends CTS packets to the relay node and the source to be probed,

after receiving the CTS packet, the relay node in the network replies an RTS packet to the signal sink, wherein the RTS packet carries the information of the first hop relay channel, and the signal sink obtains a set of detection relay nodes by demodulating a training sequence in the RTS packet

SNR value of the post-optimal single-relay supplemental channel

Proceeding to step S6;

s6, if

λ ^* For the optimal average throughput of the system, the information destination d (k) selects an access mode with higher channel rate from the direct channel access and the relay auxiliary channel access, and sends a CTS packet to the information source s (k), the information source s (k) is informed of the access decision result through the CTS packet to perform corresponding channel access,proceeding to step S7; otherwise, abandoning the access opportunity, and returning to the step S2;

s7, if

Then the source s (k) is subsequent

Forwards the received signal to the sink, gamma _s(k),l For the SNR value of the single relay auxiliary channel with index l, then the sink will receive two signals from the direct connection channel and from the relay auxiliary channel with transmission time of

Returning to step S2 after the transmission is finished;

s8, if

Then the destination d (k) selects direct channel access, then the destination d (k) sends CTS to the source s (k) and all other sources, let the source s (k) follow τ _d Within time, at a maximum achievable rate R _d (gamma) data transmission on direct connection channel, all other sources at tau _d Wait in time, passing tau _d After the time, the single transmission ends, and the step returns to step S2; otherwise, the information destination d (k) gives up the access opportunity, broadcasts CTS packet to all information source nodes, and all information sources restart channel competition in the next time slotAnd returning to step S2.

The effectiveness and the process realizability of the method provided by the invention are verified through computer simulation. 5 information source-information sink pairs and a plurality of relay points are arranged in the wireless network. Both the direct channel from each source to its sink and the relayed first hop channel from each source to each relay and the relayed second hop channel from each relay to its sink experience random rayleigh fading. The main configuration parameter of the network model is p ₀ ＝p ₁ ＝0.3，δ＝25us，τ _RTS ＝τ _CTS 50 us. First, the statistical characteristics of the wireless channel are set to the values in table 1.

TABLE 1 radio channel statistics table

When the number of relays L is 4, the performance of the proposed optimal scheduling strategy is simulated. FIG. 4 shows SNR expectations of direct connection channels

Increasing from 1 to 7dB, the channel coherence time tau _d The system average throughput of the network increases from 1ms to 4 ms. It is clear that with the coherence time τ _d The average throughput of the system is obviously increased under the same channel condition. The influence of the number of different relay nodes on the average throughput performance of the network in the same channel environment is simulated, as shown in fig. 5. With the increase of the relay number L from 1 to 7, the throughput performance of the system is enhanced, which shows that the increase of the relay nodes does not cause the increase of the network load, but increases the relay diversity, so that the system performance is effectively improved. The optimal access strategy of each source-sink pair is further analyzed. Under the parameter configuration of table 1, the benefit of relay-assisted transmission, the benefit of direct link transmission and the expected benefit of giving up access are studied, wherein the direct channel is connected

Is 5dB, coherence time tau _d Is 1 ms. As shown in table 1, for different communication pairs i, the first hop channel and the second hop channel of the relay channel have different statistical characteristics, so that the access policy corresponding to each communication pair is different, and in the present simulation embodiment, the communication pairs are classified into class ii:

class I:

fig. 6 is a graph of the revenue functions of different access methods of communication pair 1 as a function of gamma, with particular curve intersections labeled. According to fig. 6, for communication pair 1,

thus communication pair 1 belongs to the set

For class I communication pairs, likewise, communication pair 2 belongs to the set

In the collection

In the method, the sounding relay can not bring extra benefits, and only can select a direct connection channel to access or give up. The revenue function curves for communication pair 3 and communication pair 5 are presented and compared in fig. 7 and 8, respectively. As shown in fig. 7, the intersection point

And

a decision is made as to whether or not a sounding relay is required, and, obviously, for communication pair 3,

thus communication pair 3 is a type of communication pair, belonging to the set

When the SNR of the direct connection channel is satisfied

Communication pair 3 will select the detection relay. In this case, the optimal number of relays to detect needs to be calculated, so we study the difference function W _3,j (λ ^* γ),

j

1,2, 5, it can be seen in fig. 7(b) that there is a unique intersection point υ _3,1 So that W is _3,j (λ ^* Y) is 0, in which case j is 1. Accordingly, as shown in FIG. 7(a), when the SNR γ ≦ ν for the direct-connected channel _3,1 Time, relay probe revenue function V ₃ ^* (γ,2) max, so it is optimal to probe both relays; otherwise, when the direct connection channel SNR gamma is more than upsilon _3,1 Time, relay probe revenue function V ₃ ^* And (gamma, 1) is maximum, and detection of one relay is optimal. In fig. 8 it can be seen that the simulation results for communication pair 5 are similar to communication pair 3. In order to verify the influence of channel heterogeneity on the proposed optimal access method, the system performance under different parameter configurations is simulated. As shown in table 2, the direct connection channel and the relay channel both consider 'homogeneous' and 'heterogeneous', and are specifically divided into four cases: 1) direct connection 'isomorphism' and relay 'isomorphism', the average value of the SNR of the direct connection channel is the same for different communication pairs i, and the average SNR of the relay channel is also the same for different relays j. 2) Direct connection 'isomorphism' and relay 'isomerism', average SNR of direct connection channels of all communication pairs is the same, and average SNR of relay channels of different relay nodes is different. 3) Direct connection 'heterogeneous' and relay 'isomorphism', the average SNR of direct connection channels of different communication pairs is different, and the average SNR of relay channels of all relay nodes is the same. 4) Direct connection 'heterogeneous' and relay 'heterogeneous', the average SNR of direct connection channels of different communication pairs is different, and the average SNR of different relay node channels is also different.

Table 2 channel isomorphic/heterogeneous parameter configuration

All communication pairs access the channel in different ways according to our proposed intelligent channel access strategy. Table 3 lists the relay detection thresholds in the case that the channel statistical characteristics of the direct connection and the relay are 'homogeneous' and 'heterogeneous', respectively, and as can be seen from the table, in particular, for

communication pairs

1 and 2, the channel states of the direct connection and the relay channel are not good, and the relay detection threshold does not exist, so that an access mode in which the relay is not detected is always selected, that is, only the direct connection access is selected or abandoned. For communication pairs 3, 4 and 5, there is only one threshold v _i,1 The number of probing relays is determined, i.e. when selecting further probing relay nodes, the number of relay nodes that are best probed can only be 1 or 2.

Table 3 channel isomorphism/heterogeneity simulation results

In order to further study the influence of network heterogeneity on the proposed optimal access method, the channel coherence time τ was simulated _d The revenue function curve for each communication pair at 4 ms. FIG. 9 shows the equation when _d 4ms, the difference of return function W for communication pair 3 and communication pair 5 _i,j (λ ^* γ), i-3, 5, j-1, 2, 5, the curve of γ, when τ is shown in fig. 5 and 6 _d When the time is 1ms, both the communication pair 3 and the communication pair 5 have only one threshold, and the number of the detection relays can only be 1 or 2; when is tau _d At 4ms, the communication pair 3 still has only one threshold, but the value of the threshold changes, denoted as υ _3,1 There are three thresholds for communication pair 5, labeled as v, respectively _5,1 ，υ _5,2 And upsilon _5,3 That is, the communication pair 5 can selectively probe 1-4 relay nodes, and the corresponding optimal access method changes. In summary, under the same channel condition, the increase of the channel coherence time brings higher tolerance to the sounding delay, and the number of sounding relays is increased accordingly.

In order to verify the performance optimality of the method, the method is compared with the other two conventional access methodsAnd (5) carrying out performance comparison. Two existing classical strategies are described as 1) a no-relay optimal stop strategy (NR-OS) in which a winning source only has CSI of a direct-connected channel and selects the direct-connected channel to access or abandon according to the current channel quality; 2) and (4) a full-relay optimal stop strategy (FR-OS), namely, after a winning information source obtains the CSI of the direct link, the CSI of all relay channels is obtained by detecting all relay nodes, and then the access is carried out in a mode of selecting a channel with a better channel rate from a direct connection channel and a relay auxiliary channel. FIG. 10 shows the averaging when directly connected channels

System throughput performance versus curve for different access methods when varying from 1dB to 7 dB. It can be seen that the optimal channel access strategy proposed by the present invention has absolute performance advantage compared with other two existing access strategies; the system throughput of the NR-OS strategy is worst, since the average SNR of the direct channel is much smaller than the average SNR of the relayed channel; the FR-OS strategy significantly improves system throughput compared to the NR-OS strategy by probing and using all relays, however, it still has a gap compared to the proposed optimal access strategy since the channel gain due to cooperative transmission cannot compensate the time cost of relay probing.

The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision is characterized in that specific definition and assumption of parameters used in the method comprise the following steps:

k signal source-signal sink pairs, wherein the signal source index number is S ₁ ,...,S _i ,...,S _K Sink index number D ₁ ,...,D _i ,...,D _K L relay nodes and R index number ₁ ,R ₂ ,...R _L (ii) a All nodes assume time synchronization and compete for channels in a distributed mode on the basis of micro time slots; when each minislot starts, all sources have the same probability p ₀ Independently sending a channel competition RTS packet, wherein if and only if only one information source sends the RTS packet in the same micro time slot, the information source is a winning information source, the process is called a successful channel competition, and the process from the beginning of the channel competition to the appearance of the winning information source is defined as an observation; for each observation, the time to be spent before the winning source appears is random, and since each channel contention is independent, the total number of channel contention experienced for a single observation follows a parameter Kp ₀ (1-p ₀ ) ^K-1 The mean time of a single observation is:

wherein tau is _RTS Indicating the time of transmission of RTS packets, τ _CTS Indicates the time of transmission of the CTS packet, delta indicates the duration of the idle slot, and the time of collision is represented by tau _RTS Represents;

And

And

And

the channel noise follows the Gaussian distribution of the normalized variance;

the channel coherence time is recorded as τ _d Sounding relay node set

The signaling interaction time of all the relay nodes is recorded as

I | represents a modulo operation,

representing a set of probing relay nodes

The number of relay nodes; winning source-sink pair if selected to probe the set of relay nodes

And for the probing relay node set

Achievable rate of the direct connection channel is R _d (γ _i )＝log ₂ (1+γ _i ) The receiving end signal under the relay auxiliary channel access comprises a direct connection channel signal, a relay two-hop channel signal and an information source S _i In probing relay node sets

The resulting maximum channel SNR for the source-sink pair for supplemental channel transmission through the best single relay in the probed set of relay nodes is then

The corresponding channel achievable rate is

Modeling a channel observation process after single channel competition success into two sub-observation processes, wherein for the kth observation, the obtained observation information is phi for the 2k-1 and 2k sub-observation processes, specifically for the 2k-1 sub-observation process _k ＝{s(k),γ _s(k) (k),t _s (k) Where s (k) denotes the index number of the winning source in the k-th observation, γ _s(k) (k) SNR, t, representing the direct channel between the winning source s (k) and its sink _s (k) Representing the channel contention time of the k-th observation; then, a winning information sink d (k) corresponding to a winning information source s (k) selects a further detection relay channel, then a 2 k-time sub-observation is carried out, and observation information obtained by the 2 k-time sub-observation is represented as

Wherein

Is the set of relay nodes, γ, probed in the 2 k-th sub-observation _s(k),j (k) And gamma _j,d(k) (k) First-hop and second-hop channel SNRs from a source s (k) to a relay node j and from the relay node j to a sink d (k), respectively;

determining the moment of accessing the channel by the winning source, for whichThe sub-observation process of the channel competition at the previous moment carries out observation path modeling, and when the observation times | Pi | is odd, the observation path model

When the observation time | Pi | is even number, observing the path model

Representing a set of relay nodes to be probed; for the observation path pi, the cumulatively obtained observation information is denoted as B _π The revenue function is expressed as Y (pi), and when the observation times | pi | is odd,

Wherein

Represents;

based on the heterogeneous wireless distributed network model, the method aims to find the optimal intelligent channel access and decision method, namely the optimal strategy N ^* To average the system throughput of the network

Most preferably, wherein

Representing expectation, sup represents minimum upper bound;

the heterogeneous wireless distributed network intelligent channel access method based on pure threshold decision comprises the following specific steps:

s1, obtaining the optimal average throughput of the system, the communication pair classification interval, the primary decision threshold and the relay number decision threshold through offline iterative computation according to the channel statistical characteristic parameters of the heterogeneous wireless network;

s2, carrying out channel competition by a plurality of information sources; starting from a minislot of duration delta, all sources are given a probability p ₀ After sending RTS packets to contend for the channel independently, the following three situations occur:

if no source sends RTS packet in the micro time slot, the channel is idle, and all sources compete for the channel in the next time slot;

if only one information source s (k) sends an RTS packet, wherein k represents that the current channel competition is the k-th successful channel competition, and s (k) is an information source index number, the information source obtains a channel access opportunity and is called as a winning information source; an information sink d (k) corresponding to a winning information source obtains a signal-to-noise ratio gamma of a direct-connected channel by receiving RTS data packets and utilizing a training sequence carried in the RTS data packets, all relay nodes obtain SNR (signal to noise ratio) from the information source to a first hop relay channel of the relay nodes, and d (k) is an index number of the information sink;

For class II communication pair interval, go to step S8, otherwise

For the class i communication pair section, the process proceeds to step S4;

s5, performing threshold comparison of detection relay quantity, and taking value of index number j of relay nodeFrom j _i,l To j _i,u ，j _i,l And j _i,u Respectively representing and preliminary decision threshold

And

The sink d (k) sends CTS packets to the relay node and the source to be probed,

SNR value of the post-optimal single-relay supplemental channel

The flow advances to step S6;

s6, if

s7, if

Then the source s (k) is subsequent

Reachable rate R in direct-connected channel within time _d (gamma) transmitting information, and returning to the step S2 after the transmission is finished; otherwise, the source s (k) is at the maximum achievable rate under the relay supplemental channel

Returning to step S2 after the transmission is finished;

s8, if

Then the destination d (k) selects direct channel access, then the destination d (k) sends CTS to the source s (k) and all other sources, let the source s (k) follow τ _d Within time, at an achievable rate R _d (gamma) data transmission on direct connection channel, all other sources at tau _d Wait in time, passing tau _d After the time, the single transmission ends, and the step returns to step S2; otherwise, the destination d (k) gives up the access opportunity and broadcasts CTS packet to all sourcesThe node, all the information sources restart the channel competition in the next time slot, and the step S2 is returned to;

in the step S1, the specific calculation procedure is,

probing relay node set of ith source-sink pair

The revenue function of (a) is defined by the expression:

wherein

Wherein, the set of relay nodes to be detected is represented as

i ₁ ,i ₂ ,...,i _J Indicating a set of relay nodes to be probedThe index sequence number of the j detection relay of the ith source-sink pair has a sequencing relation i ₁ ＜...＜i _j ＜...＜i _J To a set of

Indicating the SNR threshold at which access is given up,

E ₁ (x) Is an exponential integral function, specifically expressed as

Temporal relay supplemental channel SNR variant

The cumulative probability distribution function of (a), in particular,

wherein

λ ^* For the optimal system average throughput, when λ is λ ═ λ, for the optimal system average throughput achievable under the current network channel conditions ^* In time, a revenue function for detecting relays under optimal system average throughput is defined

The analytical expression is as follows:

wherein is beta' _j (γ)＝c _i，j ·(α(λ ^* ，γ)-1)，

For the source S _i 1,2, K and relay R _j J 1, 2.. said, L, the optimal set of relay nodes is defined to implement the maximum revenue function

Represented as a set of relay nodes

thus is provided with

Represents the set of all relay nodes in the network;

for the source node S _i I 1, 2. -, K and relay node R _j 1, 2.. L, defining a threshold function based on the number of relays

And source-based threshold function

For i 1,2, K and j 1,2, 1 _i，j (λ ^* ，γ)＝V _i ^* (γ，j+1)-V _i ^* (γ, j), let W _i,j (λ ^* When γ) is 0, the decision threshold for solving the relay SNR value γ is { υ _i,j } _{i＝1,2,...,K,j＝1,2,...,L-1} ；

Primary decision threshold

Is an equation

Is an equation

Wherein

For the purpose of the lower primary decision threshold,

is an upper primary decision threshold;

j _i,l And j _i,u Respectively representing and preliminary decision thresholds

And

the number of corresponding detection relay nodes;

optimal system average throughput lambda ^* Satisfies the equation:

wherein, the first term on the left side τ _d R _d (γ _i )-λτ _d Representing the gain of the direct connection channel, the second term U on the left ₀ (λ)＝E[max{τ _d R _d (γ _i )-λτ _d ,U ₀ (λ),V _i ^* (γ _i )}]-λτ _o It is clear that when λ ═ λ ^* While, U ₀ (λ ^* ) 0, the third term V on the left _i ^* (γ _i ) Is a threshold function based on the information source i, and represents that when the SNR of the direct connection channel is gamma _i Detecting the maximum expected benefit of the relay channel;

for λ > 0, when

While, U ₀ The expression of (lambda) is as follows,

for λ > 0, when

While, U ₀ The expression of (lambda) is as follows,

wherein eta is _i,0 (λ) to satisfy equation R _d (x) A unique solution for λ;

when V is _i ^* (2 ^λ -1) > 0, and (c),

time, corresponding source-sink pair

Is a class I communication pair interval; otherwise

Is a class ii communication pair interval.

2. The pure threshold decision based intelligent channel access method for heterogeneous wireless distributed networks according to claim 1,

the heterogeneous wireless distributed network comprises K information source-information sink pairs and L amplification forwarding relay nodes, and the selectable access modes of the information source-information sink pairs comprise three types: accessing a direct connection channel when the relay node is not detected; accessing a relay auxiliary channel; the direct connection channel access is realized when a certain number of relay nodes are detected, and the mode detects the certain number of relay nodes, but the channel condition of the direct connection channel is found to be better, so that the direct connection channel is selected for access; RTS, which represents a request-to-send packet, is a data packet in a channel sensing access protocol and is used for detecting the occupation condition of a channel and estimating the channel quality by a sending node user; CSI, data information representing channel state information and reflecting real-time conditions of a wireless channel; the CTS, which indicates a clear-to-send packet, is a data packet of a channel aware access protocol, and is used by the receiving node to respond to the sending node.

3. The pure threshold decision based intelligent channel access method for heterogeneous wireless distributed networks according to claim 1,

calculating the average throughput lambda of the optimal system ^* And sorting pairs of intra-network communicationsThe off-line iterative computation comprises the following specific steps:

s11, inputting channel parameter tau _RTS 、τ _CTS 、τ _d 、

And

S13, judging the iteration precision delta of the k time _k Whether a precision convergence threshold epsilon is reached, if delta _k If not, continuing the iteration until the k-th iteration precision does not reach the precision convergence threshold, and entering step S14, otherwise, ending the iteration and entering step S17;

Returning to step S13;

And