DYNAMIC SPECTRUM SHARING METHOD BASED ON USER ONLINE LEARNING AND LOW-OVERHEAD COOPERATION

Abstract

The present invention provides a dynamic spectrum sharing method based on user online learning and low-overhead cooperation, and relates to the technical field of electromagnetic spectrums. The method comprises: first, a secondary user obtaining spectrum sharing scene information and requirement parameters; then initializing a state parameter and a spectrum sensing parameter of a channel; then, at the beginning of each time slot, the secondary user determining whether a direct transmission condition is met or not according to the transmission condition of the previous time slot, selecting to directly continue transmission on the channel of the previous time slot or obtain idle channel according to the channel state and spectrum sensing for transmission, and updating the state parameter and spectrum sensing parameter of the channel; and finally, a base station of the secondary user performing conflict coordination according to the channel occupancy condition every a fixed number of time slots. According to the present invention, low-overhead and efficient spectrum sharing is realized by the secondary user learning the channel state and the transmission condition, the spectrum utilization rate in a high user density and dynamic spectrum environment can be effectively improved, and efficient spectrum sharing is realized.

Claims

1. A dynamic spectrum sharing method based on user online learning and low-overhead cooperation, comprising: S1: obtaining, by a secondary user (SU), scenario information and requirement parameters of spectrum sharing, wherein obtaining, by the SU, the scenario information of spectrum sharing comprises: obtaining, by the SU, a sharable channel set M={1, . . . , M}, a SU set N={1, . . . , N} participating in spectrum sharing, a quantity L of candidate channels of each SU, a time τ required to sense each channel, and an interval T.sub.1 at which the SU reports a channel usage to a secondary user base station (SUBS); and the requirement parameters comprise a spectrum usage conflict probability threshold η of the SU and a channel conflict adjustment parameter φ; S2: initializing, for each SU, state parameters and spectrum sensing parameters of each channel m∈M, wherein the state parameters of the channel comprise a first channel state parameter α.sub.m and a second channel state parameter β.sub.m, and the spectrum sensing parameters comprise a quantity κ.sub.m of slots skippable for channel sensing and a quantity l.sub.m of remaining slots skippable for channel sensing; S3: at a beginning of each slot, determining, by the SU, whether a direct transmission condition is met based on a transmission status of a previous slot; and continuing transmission directly on a channel of the previous slot when the direct transmission condition is met; or obtaining an idle channel for transmission based on a channel state and spectrum sensing when the direct transmission condition is not met, wherein obtaining the idle channel for transmission based on the channel state and spectrum sensing when the direct transmission condition is not met comprises: S31: calculating, by the SU, a current priority of each channel according to a formula $P_{m,t}=\stackrel{\_}{x_{m,t}}+c\sqrt{\frac{\ln t}{h_{m,t}}},$ wherein P.sub.m, t indicates a priority parameter, t indicates a number of a current slot, x.sub.m,t indicates an average throughput obtained when the SU selects a channel m in previous t−1 slots, h.sub.m,t indicates a quantity of times the channel m is selected by the SU, and c is a constant; S32: arranging all channels in descending order of P.sub.m, t and taking first L channels as candidate channels of the SU; S33: sequentially performing, by the SU, spectrum sensing on the L candidate channels one by one, until the idle channel is obtained; obtaining a channel occupancy estimation parameter θ of a primary user (PU) through random sampling in a Γ distribution Γ(α.sub.m,β.sub.m) based on current values of the first channel state parameter α.sub.m and the second channel state parameter β.sub.m; and calculating the quantity κ.sub.m of slots skippable for channel sensing, wherein κ.sub.m is calculated by ${\kappa }_m=.Math.\frac{1}{2T_0}\left(\frac{1}{\min \left(\theta ,{\alpha }_m/{\beta }_m\right)}\right).Math.,$ wherein • indicates an upward rounding function, min(⋅) indicates a function for finding a smaller value, and T.sub.0 indicates a slot length; S34: setting l.sub.m=κ.sub.m and performing, by the SU, transmission on the obtained idle channel; and S35: skipping, by the SU, transmission, when no idle channel is sensed among the L candidate channels, and attempting transmission in a next slot; S4: updating the state parameters and the spectrum sensing parameters of the channel; and S5: reporting, by each SU, a channel occupancy status to the SUBS every T.sub.1 slots, and performing, by the SUBS, conflict coordination based on the channel occupancy status.

2. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 1, wherein in the step S2, initializing the state parameters of each channel m∈M comprises: setting the first channel state parameter α.sub.m=1; setting the second channel state parameter β.sub.m=T.sub.0, wherein T.sub.0 indicates the slot length; initializing the spectrum sensing parameters of each channel m∈M comprises: setting the quantity of slots skippable for channel sensing κ.sub.m=0; and setting the quantity of remaining slots skippable for channel sensing l.sub.m=0.

3. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 2, wherein in the step S3, determining, by the SU, whether the direct transmission condition is met based on the transmission status of the previous slot comprises: determining that the direct transmission condition is met when the SU performs transmission successfully on a channel in the previous slot and the quantity l.sub.m of remaining slots skippable for channel sensing is greater than or equal to 1; otherwise, determining that the direct transmission condition is not met.

4. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 3, wherein in the step S3, continuing transmission directly on the channel of the previous slot when the direct transmission condition is met comprises: setting the parameter l.sub.m=l.sub.m−1, and directly selecting, by the SU, the channel of the previous slot for transmission.

5. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 4, wherein in the step S4, updating the state parameters and the spectrum sensing parameters of the channel when continuing transmission directly on the channel of the previous slot when the direct transmission condition is met in the step S3 comprises: when the transmission succeeds and l.sub.m=0, setting β.sub.m=β.sub.m+κ.sub.mT.sub.0, when the transmission succeeds and l.sub.m≠0, skipping updating the state parameters of the channel; or when the transmission fails, setting α.sub.m=α.sub.m+1, β.sub.m=β.sub.m+(κ.sub.m−1)T.sub.0+0.5T.sub.0, and l.sub.m=0, wherein κ.sub.m is calculated by the formula ${\kappa }_m=.Math.\frac{1}{2T_0}\left(\frac{1}{\min \left(\theta ,{\alpha }_m/{\beta }_m\right)}\right).Math.$ after obtaining the channel occupancy estimation parameter θ of the PU through random sampling in the Γ distribution Γ(α.sub.m,β.sub.m) based on the current values of the first channel state parameter α.sub.m and the second channel state parameter β.sub.m.

6. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 4, wherein in the step S4, updating the state parameters and the spectrum sensing parameters of the channel when obtaining the idle channel for transmission based on the channel state and spectrum sensing when the direct transmission condition is not met in the step S3 comprises: when the transmission succeeds, setting β.sub.m=β.sub.m+κ.sub.mT.sub.0, and l.sub.m=l.sub.m−1; or when the transmission fails, setting l.sub.m=κ.sub.m=0, α.sub.m=α.sub.m+1, and β.sub.m=β.sub.m+(κ.sub.m−1)T.sub.0+0.5T.sub.0.

7. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 1, wherein in the step S5, performing, by the SUBS, the conflict coordination based on the channel occupancy status comprises: adjusting, by the SUBS, a return value of the channel to the SU based on the channel occupancy status; and when a probability of a channel being occupied by two or more SUs at the same time in T.sub.1 slots is greater than the conflict probability threshold η, decreasing average return values by φ for the SUs selecting the channel except an SU with a largest average return value, and feeding back a new return value to each SU.

8. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 1, wherein in the step S1, the quantity of the candidate channels of each SU is determined based on a channel occupancy status of the PU and an expected probability of obtaining an available channel per access attempt, and the time required to sense each channel is determined based on a protection requirement of the PU and a sensing capability of the SU.

9. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 2, wherein in the step S1, the quantity of the candidate channels of each SU is determined based on a channel occupancy status of the PU and an expected probability of obtaining an available channel per access attempt, and the time required to sense each channel is determined based on a protection requirement of the PU and a sensing capability of the SU.

10. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 3, wherein in the step S1, the quantity of the candidate channels of each SU is determined based on a channel occupancy status of the PU and an expected probability of obtaining an available channel per access attempt, and the time required to sense each channel is determined based on a protection requirement of the PU and a sensing capability of the SU.

11. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 4, wherein in the step S1, the quantity of the candidate channels of each SU is determined based on a channel occupancy status of the PU and an expected probability of obtaining an available channel per access attempt, and the time required to sense each channel is determined based on a protection requirement of the PU and a sensing capability of the SU.

12. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 5, wherein in the step S1, the quantity of the candidate channels of each SU is determined based on a channel occupancy status of the PU and an expected probability of obtaining an available channel per access attempt, and the time required to sense each channel is determined based on a protection requirement of the PU and a sensing capability of the SU.

13. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 6, wherein in the step S1, the quantity of the candidate channels of each SU is determined based on a channel occupancy status of the PU and an expected probability of obtaining an available channel per access attempt, and the time required to sense each channel is determined based on a protection requirement of the PU and a sensing capability of the SU.

14. The dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to claim 7, wherein in the step S1, the quantity of the candidate channels of each SU is determined based on a channel occupancy status of the PU and an expected probability of obtaining an available channel per access attempt, and the time required to sense each channel is determined based on a protection requirement of the PU and a sensing capability of the SU.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0044] To describe the technical solutions of the embodiments of the present disclosure or the prior art more clearly, the accompanying drawings required for the embodiments are briefly described below. Apparently, the accompanying drawings described below show merely some embodiments of the present disclosure, and persons of ordinary skill in the art may still derive other accompanying drawings from these accompanying drawings without creative efforts.

[0045] FIG. 1 is a flowchart illustrating a dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to the present disclosure;

[0046] FIG. 2 is a schematic diagram of a spectrum sharing scenario;

[0047] FIG. 3 is a diagram of a slot/frame structure of spectrum access by an SU; and

[0048] FIG. 4 is a diagram of simulation comparison between the effect of the prior art and the effect of the dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to the present disclosure.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] The technical solutions in the embodiments of the present disclosure will be described below clearly and completely with reference to the accompanying drawings in the embodiments of the present disclosure. Apparently, the described embodiments are merely some rather than all of the embodiments of the present disclosure. All other examples obtained by persons of ordinary skill in the art based on the examples of the present disclosure without creative efforts shall fall within the protection scope of the present disclosure.

[0050] The purpose of the present disclosure is to provide a dynamic spectrum sharing method based on user online learning and low-overhead cooperation, to mainly resolve the problem how a plurality of SUs select channels based on channel competition and adjust spectrum sensing strategies based on activities of PUs, thereby achieving efficient spectrum sharing.

[0051] To make the above-mentioned objective, features, and advantages of the present disclosure clearer and more comprehensible, the present disclosure will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0052] As shown in FIG. 1, the dynamic spectrum sharing method based on user online learning and low-overhead cooperation according to the present disclosure includes:

[0053] S1: An SU obtains scenario information and requirement parameters of spectrum sharing.

[0054] In this embodiment of the present disclosure, the SU obtains the scenario information of spectrum sharing may include: obtaining, by the SU, a sharable channel set M={1, . . . , M}, an SU set N={1, . . . , N} participating in spectrum sharing, a quantity L of candidate channels of each SU, a time τ required to sense each channel, and an interval T.sub.1 at which the SU reports a channel usage to an SUBS; and the requirement parameters include a spectrum usage conflict probability threshold η of the SU and a channel conflict adjustment parameter φ. The quantity L of candidate channels of each SU is determined based on a channel occupancy status of the PU and an expected probability of obtaining an available channel by each access attempt, and the time τ required to sense each channel is determined based on a protection requirement of the PU and a sensing capability of the SU. It is noted that in the channel set M={1, . . . , M}, M indicates the largest numbered channel, and M also indicates the quantity of channels in the channel set. Similarly, in the SU set N={1, . . . , N}, N indicates the largest numbered SU, and N also indicates the quantity of SUs in the SU set.

[0055] FIG. 2 is a schematic diagram of a spectrum sharing scenario. There are a plurality of PUs, a plurality of SUs and one SUBS. The sharable channel set M={1, . . . , M} is preferentially used by the PUs, and the SUs can access and use it without intolerable interference to the PUs. In actual sharing, the SUs need to know the status of the sharable channel. However, due to the large quantity and variability of the SUs, each SU does not need to know the status of the other SUs, but determines a channel competition status with the other SUs according to a determination whether the transmission succeeds or not.

[0056] FIG. 3 is a diagram of a slot/frame structure of spectrum access by each SU. In each slot, the SU senses at most L channels, and L is determined based on the channel occupancy status of the PU and the expected probability of obtaining the available channel by each access attempt. For example, in a case that the probability of channel occupancy by the PU is about 70% and the expected probability of obtaining the channel for each attempt is 80%, only about 5 channels need to be sensed to obtain an idle channel, i.e., L=5. The time required to sense each channel τ depends on the sensing capability of the SU and the protection requirement of the PU (for example, a detection probability), provided that the detection probability is not less than the probability required to protect the PU. Therefore, it can be seen from FIG. 3 that the duration of the sensing phase is L.sub.τ, that is, the product of L and τ, and the duration of the transmission phase is T.sub.0-L.sub.τ.

[0057] The parameter T.sub.1 is used to specify a period of coordination of channel occupancy competition by the SUBS, and its specific value may be estimated based on the quantity of users and the quantity of shareable channels in the area. The parameters η and φ are used to specify the degree of adjustment to channel conflict, and their specific parameter values may be about 10% and 20%.

[0058] S2: State parameters and spectrum sensing parameters of each channel m∈M are initialized for each SU, where the state parameters of the channel include a first channel state parameter α.sub.m and a second channel state parameter β.sub.m, and the spectrum sensing parameters include a quantity κ.sub.m of slots skippable for channel sensing and a quantity l.sub.m of remaining slots skippable for channel sensing.

[0059] It is noted that the SUs all belong to the SU set N={1, . . . , N}. The first channel state parameter α.sub.m indicates a quantity of selections for a particular channel, and the second channel state parameter β.sub.m indicates a time of successful transmission on the current channel.

[0060] Specifically, the state parameters of the channel m∈M are initialized by setting the first channel state parameter α.sub.m=1 and the second channel state parameter β.sub.m=T.sub.0. T.sub.0 indicates a slot length. The spectrum sensing parameters of the channel m∈M are initialized by setting the quantity of slots skippable for the channel sensing κ.sub.m=0 and the quantity of remaining slots skippable for the channel sensing l.sub.m=0.

[0061] S3: At the beginning of each slot, the SU determines, based on a transmission status of a previous slot, whether a direct transmission condition is met, and continues transmission directly on a channel of the previous slot if the direct transmission condition is met, or obtains an idle channel for transmission based on a channel state and spectrum sensing if the direct transmission condition is not met.

[0062] Specifically, in this embodiment of the present disclosure, the method for determining, by the SU based on the transmission status of the previous slot, whether the direct transmission condition is met, includes: determining that the direct transmission condition is met if the SU performs transmission successfully on a channel in the previous slot and the quantity l.sub.m of remaining slots skippable for the channel sensing is greater than or equal to 1; otherwise, determining that the direct transmission condition is not met.

[0063] If the direct transmission condition is met, set l.sub.m=l.sub.m−1, and the SU directly selects the channel of the previous slot for transmission.

[0064] If the direct transmission condition is not met, the method for obtaining the idle channel for transmission based on historical channel selection and transmission status includes steps S31-S35.

[0065] S31: The SU calculates a current priority of each channel according to a formula

[00004]$P_{m,t}=\stackrel{\_}{x_{m,t}}+c\sqrt{\frac{\ln t}{h_{m,t}}},$

where P.sub.m, t indicates a priority parameter, t indicates the number of a current slot, x.sub.m,t indicates an average throughput obtained when the SU selects a channel m in previous t−1 slots, h.sub.m,t indicates a quantity of times the channel m is selected by the SU, and c is a constant. In a preferred embodiment of the present disclosure, c=√{square root over (2)}. For example, when the current slot is the twentieth slot, t=20.

[0066] S32: All channels are sorted in descending order of P.sub.m, t and first L channels are taken as candidate channels of the SU.

[0067] S33: The SU performs spectrum sensing on the L candidate channels sequentially one by one, until an idle channel is obtained; a channel occupancy estimation parameter θ of a PU is obtained through random sampling in a Γ distribution Γ(α.sub.m, β.sub.m) based on current values of the first channel state parameter α.sub.m and the second channel state parameter β.sub.m; and the quantity κ.sub.m of slots skippable for sensing by the SU is calculated, where κ.sub.m is calculated by

[00005]${\kappa }_m=.Math.\frac{1}{2T_0}\left(\frac{1}{\min \left(\theta ,{\alpha }_m/{\beta }_m\right)}\right).Math.,$

where • indicates an upward rounding function, and min(⋅) indicates a function for finding a smaller value.

[0068] S34: Set l.sub.m=κ.sub.m and the SU performs transmission on the obtained idle channel.

[0069] S35: The SU skips transmission if no idle channel is sensed among the L candidate channels, and attempts transmission in a next slot.

[0070] The reason for adopting the foregoing method lies in that the SU's spectrum detection before spectrum access in each slot, will cause unnecessary time waste because the PU may not use the channel for several consecutive slots. Therefore, each time an idle channel is detected, it is necessary to estimate the quantity of slots in which the PU's channel occupancy status does not need to be detected but communication is directly performed.

[0071] Taking the PU services following an exponential distribution as an example, the length of slots skippable for spectrum sensing is estimated, Bayesian analysis is used to estimate the remaining idle time of the channel, and in a case that θ indicates a channel idle time distribution parameter, i.e., the channel occupancy estimation parameter of the PU, the channel idle time satisfies f(t)=θe.sup.−θt, and the average idle time is 1/θ.

[0072] Based on the probability theory, it can be known that the posterior probability is proportional to the likelihood probability and the prior probability. If the posterior distribution and the prior distribution belong to the same class, the prior distribution and the posterior distribution are referred to as conjugate distributions, and the prior distribution is referred to as the conjugate prior of the likelihood function. According to the common conjugate prior distribution, the conjugate prior distribution of exponential distribution follows the Γ distribution with parameters α and β, and the prior probability of θ is

[00006]$p\left(\theta \right)=\gamma \left(\theta .Math.\alpha ,\beta \right)=\frac{{\beta }^{\alpha }{\theta }^{\alpha -1}{e}^{-\mathrm{\beta \theta }}}{\Gamma \left(\alpha \right)}.$

[0073] x indicates the duration in which data does not experience conflict during transmission, and the function of x with respect to θ is the exponential distribution. Therefore, the posterior probability of θ is proportional to the likelihood probability and the prior probability, that is,

[00007]$p\left(\theta .Math.x\right)\propto \frac{{\beta }^{\alpha }{\theta }^{\alpha -1}{e}^{-\mathrm{\beta \theta }}}{\Gamma \left(\alpha \right)}.Math.\theta e^{-\theta x},$$\propto \frac{{\beta }^{\alpha }{\theta }^{\alpha }{e}^{-\theta \left(\beta +x\right)}}{\Gamma \left(\alpha \right)}.$

[0074] That is, p(θ|x)=γ(θ|α+1,β+x). The goal of the system is to best estimate the idle time, to maximize the system spectrum utilization and minimize the system conflict. When there are K samples, the posterior probability of θ is

[00008]$p\left(\theta .Math.x_1,K,x_K\right)=\gamma \left(\theta .Math.\alpha +K,\beta +\underset{k=1}{\overset{K}{.Math.}}{x}_k\right).$

[0075] k indicates a serial number. Based on the probability theory, the posterior probability converges to near the maximum likelihood estimation as the quantity of samples increases. In addition, with an inspiration from the Optimistic Thompson sampling, larger values can be taken, which also makes the skipped values converge to the optimum faster. Therefore, for the parameters α.sub.m, β.sub.m and the slot length T.sub.0, the quantity of skipped slots is calculated as follows:

[00009]${\kappa }_m=.Math.\frac{1}{2T_0}\left(\frac{1}{\min \left(\theta ,{\alpha }_m/{\beta }_m\right)}\right).Math..$

[0076] The parameter α.sub.m indicates the quantity of selections for a particular channel. Since the SU can select any channel in the network, the SU may access the same channel for many times. The parameter α.sub.m is updated only when a different channel is selected next time, or when a conflict occurs in the current channel. The parameter β.sub.m indicates the time of successful transmission on the current channel, and the feedback value for β.sub.m is reduced accordingly when a conflict is detected.

[0077] S4: The state parameters and the spectrum sensing parameters of the channel are updated.

[0078] When the transmission in a slot is completed, the parameters need to be updated to prepare for the next slot. Specifically, the method for updating the state parameters and the spectrum sensing parameters of the channel when continuing transmission directly on the channel of the previous slot if the direct transmission condition is met in the step S3 may include:

[0079] if the transmission succeeds and l.sub.m=0, setting β.sub.m=β.sub.m+κ.sub.mT.sub.0;

[0080] if the transmission succeeds and l.sub.m≠0, skipping updating the state parameters of the channel; or

[0081] if the transmission fails, setting α.sub.m=α.sub.m+1, β.sub.m=β.sub.m+(κ.sub.m−1)T.sub.0+0.5T.sub.0, and l.sub.m=0.

[0082] Here, κ.sub.m is calculated by the formula

[00010]${\kappa }_m=.Math.\frac{1}{2T_0}\left(\frac{1}{\min \left(\theta ,{\alpha }_m/{\beta }_m\right)}\right).Math.$

after obtaining the channel occupancy estimation parameter θ of the PU through random sampling in the Γ distribution Γ(α.sub.m, β.sub.m) based on the current values of the first channel state parameter α.sub.m and the second channel state parameter β.sub.m.

[0083] The method for updating the state parameters and the spectrum sensing parameters of the channel when obtaining an idle channel for transmission based on the channel state and spectrum sensing if the direct transmission condition is not met in the step S3 may include:

[0084] if the transmission succeeds, setting β.sub.m=β.sub.m+κ.sub.m T.sub.0 and l.sub.m=l.sub.m−1; or

[0085] if the transmission fails, setting l.sub.m=κ.sub.m=0, α.sub.m=α.sub.m+1, and β.sub.m=β.sub.m+(κ.sub.m−1)T.sub.0+0.5T.sub.0.

[0086] S5: Each SU reports, every T.sub.1 slots, a channel occupancy status to the SUBS, and the SUBS performs conflict coordination based on the channel occupancy status.

[0087] Specifically, all the SUs report, every T.sub.1 slots, the channel occupancy status of the previous T.sub.1 slots to the SUBS, and the SUBS adjusts return values of the channel for the SUs based on the channel selection conflict; and if a probability of a channel being occupied by two or more SUs at the same time in the T.sub.1 slots is greater than the conflict probability threshold η, average return values are decreased by φ for the SUs selecting the channel except an SU with a largest average return value, and a new return value is returned to each SU.

[0088] According to the present disclosure, through a small amount of cooperative overhead of a plurality of SUs, coordination of competition can be achieved when different SUs access the channel, which reduces the spectrum resource wastes caused by channel competition while ensuring that an SU, who can obtain the maximum throughput, gets the right to use the channel. In addition, after obtaining the idle channel, the SU can estimate the remaining idle time of the channel by learning the channel state parameters online, thus avoiding unnecessary spectrum sensing by the SU at the beginning of each slot, thereby significantly improving the idle spectrum utilization. The channel state learning and utilization mechanism is easy to implement and has low overhead, which is convenient for application in practical systems.

[0089] Simulation analysis is performed on the method of the present disclosure to illustrate the beneficial effects.

[0090] The quantity M of channels in the system is 9, the quantity N of SUs is 4, and each channel has only one authorized PU. The busy duration and idle duration of the PU follow the exponential distributions with parameters λ (idle) and μ (busy) respectively. The SUs attempt to access the primary channel in an overlay manner. In the slot/frame structure, it is assumed that the length T.sub.0 of each slot is 50 ms and the time τ required for each SU to sense a channel is 3 ms. The transmission power of each SU is fixed at 23 dBm, the power of the additive white Gaussian noise is 1 mW, and the channel fading factor a is 3. In conflict avoidance, it is assumed that the parameter L=4 and the slot interval T.sub.1=100, the information of each SU is aggregated to the central node for conflict detection every 100 slots. The SU spectrum usage conflict probability threshold η is set to 10%, to be specific, if a channel conflict exceeds 10 slots, the conflict avoidance strategy is executed. In the simulation analysis, the optimization scheme proposed in this method is compared with three other schemes:

[0091] Scheme (1). ε-greedy strategy: the SU detects the channel in each slot, and then selects and accesses the channel according to the ε-greedy algorithm. A random number ε decreases gradually with the increase of the simulation times, and the system does not execute the skipping strategy and conflict avoidance in the algorithm.

[0092] Scheme (2). Non-conflict avoidance strategy: The SU uses the step S3 in this method for channel selection and slot skipping, but no conflict avoidance strategy is executed.

[0093] Scheme (3). Fixed grouping strategy: A fixed quantity of channels are allocated to each SU before transmission. In the simulation analysis, the quantity of channels allocated to each SU is set to 4 and the channels are randomly allocated provided that each channel is used as the candidate channel by at most two SUs. The SU uses the step S3 in this method for channel selection and slot skipping, but no conflict avoidance strategy is executed.

[0094] FIG. 4 compares the total throughput of the SU in different schemes. It can be seen that Scheme (1) and Scheme (2) obtain a lower system throughput due to the absence of conflict avoidance strategy. Scheme (3) adopts fixed grouping to reduce the channel occupancy conflict among the SUs and improve the throughput. However, the system throughput of Scheme (3) is limited because the throughput of a channel to different users is not considered and the grouping cannot maximize the support for SUs preferably selecting channels based on throughput. The optimized scheme in the present disclosure can converge to the optimum in a relatively short time of about 3000 slots, with a throughput increase of about 66%.

[0095] According to the comparison results of the simulation analysis, it can be found that the dynamic spectrum sharing method of the present disclosure reduces the spectrum resource wastes caused by channel competition while ensuring that an SU who can obtain the maximum throughput gets the right to use the channel, and avoid unnecessary spectrum sensing by the SU at the beginning of each slot, thereby significantly improving the idle spectrum utilization.

[0096] Each example of the present specification is described in a progressive manner, each example focuses on the difference from other examples, and the same and similar parts between the examples may refer to each other.

[0097] In this specification, several examples are used for illustration of the principles and implementations of the present disclosure. The description of the foregoing examples is used to help to illustrate the method of the present disclosure and the core principles thereof. In addition, those of ordinary skill in the art can make various modifications in terms of specific implementations and scope of application in accordance with the teachings of the present disclosure. In conclusion, the content of the present description shall not be construed as limitations to the present disclosure.