CN116318478A - Energy efficiency-based normalized unmanned aerial vehicle frequency spectrum sensing method and system - Google Patents

Abstract

The invention relates to UAV spectrum sensing technology, which is a normalized UAV spectrum sensing method and system based on energy efficiency. The method includes: the cognitive UAV senses the frequency band, and reports the sensing data to the central UAV, The central UAV performs averaging and normalization processing on the sensing data and judges. If the judgment result is idle, then access the perceived frequency band for communication, otherwise wait until the next frame to start sensing the frequency band again; the UAV’s Classify the results of spectrum sensing, and calculate the average throughput and average energy consumption of each frame of the cognitive UAV network according to the classification results of spectrum sensing; define energy efficiency as the reciprocal of the energy consumed per bit per unit bandwidth, and calculate Energy efficiency, and adjust drone perception time based on energy efficiency. The invention can realize high-performance and low-consumption spectrum sensing for the frequency band under the condition of high dynamic change of the background noise.

Description

A normalized UAV spectrum sensing method and system based on energy efficiency

technical field

The invention relates to the technical field of unmanned aerial vehicle spectrum sensing, in particular to an energy efficiency-based normalized unmanned aerial vehicle spectrum sensing method and system.

Background technique

Due to the increasing number of wireless communication devices, the limited wireless spectrum resources are becoming increasingly scarce. Cognitive radio technology enables different users to use the same spectrum resources, that is, spectrum sharing. Spectrum sensing is to find out idle spectrum resources called "spectrum holes" in multiple dimensions such as time domain, frequency domain and air domain as accurately as possible in the complex and changeable electromagnetic environment for allocation and utilization.

During UAV missions, due to the constantly changing environment, the spectrum sensing algorithm needs to be robust against the influence of noise power and signal-to-noise ratio uncertainty. In addition, the introduction of spectrum sensing technology will lead to further increase of perceived energy consumption of UAVs in the case of energy constraints. Small rotor UAVs usually work in battery-powered mode, and limited battery energy has become the main limiting factor in UAV application scenarios.

There is an existing normalized cooperative spectrum sensing algorithm, which has the ability to adapt to the high dynamic change of background noise, but does not consider the energy efficiency problem in the multi-UAV cooperative scenario. By optimizing the settings of parameters such as sensing time and fusion threshold, spectrum sensing with low energy consumption and high detection performance can be realized in the case of highly dynamic background noise changes.

Contents of the invention

In order to solve the problems existing in the prior art, the present invention provides a normalized UAV spectrum sensing method and system based on energy efficiency, which is suitable for cooperative sensing scenarios of multiple UAVs, and takes energy efficiency and detection probability as optimization goals , to design a reasonable cooperative perception strategy, which reduces resource overhead while improving anti-noise performance.

The purpose of the present invention is achieved through the following technical solutions, a normalized UAV spectrum sensing method based on energy efficiency, comprising the following steps:

The cognitive UAV senses the frequency band, and reports the sensing data to the central UAV one by one, and the central UAV averages and normalizes the sensing data for judgment. Communication in the perceived frequency band, otherwise wait until the next frame to start sensing the frequency band again;

Classify the spectrum sensing results of UAVs, and calculate the average throughput and average energy consumption of each frame of the cognitive UAV network according to the classification results of spectrum sensing;

Define energy efficiency as the reciprocal of energy consumed per bit under unit bandwidth, calculate energy efficiency, and adjust UAV perception time according to energy efficiency.

Preferably, the process of frequency band sensing includes steps:

Design a cognitive UAV network. The designed cognitive UAV network includes N cognitive UAVs and a central UAV. The cognitive UAV operates on a circle centered on the main user transmitter. spectrum sensing;

Filter the received signals at the sensing nodes of each cognitive UAV to obtain signals in the sensing frequency range; sample the signals in the sensing frequency range to obtain continuous discrete samples of the signals in the sensing frequency range;

Performing periodogram power spectrum estimation on the discrete samples collected within the range of the sensing frequency band to obtain the spectral line intensity of the discrete samples;

Take half of the spectral line intensities for summing and segmentation to represent the power spectrum; upload the power spectrum sampled by N cognitive drones to the central drone to obtain the estimated periodic spectrum of all drones average value;

Divide all UAV periodic atlases evenly into several groups to obtain the spectral line intensity of each group; perform normalization according to the spectral line intensity of each group and the sum of estimated average values of all UAV periodic atlases to obtain normalized chemical spectrum;

Preset the false alarm probability of a single-segment normalized signal, obtain the judgment threshold under the preview setting, and obtain the detection probability of a single-segment normalized signal according to the judgment threshold;

The central UAV adopts a fusion strategy to make a final judgment, and obtains false alarm probability and detection probability.

The system of the present invention adopts the following technical solutions to realize: a normalized UAV spectrum sensing system based on energy efficiency, comprising:

Cognitive UAV network. The cognitive UAV network includes N cognitive UAVs and a central UAV. The cognitive UAV senses the frequency band and reports the sensing data to the central UAV one by one. Human-machine, and the central drone averages and normalizes the sensing data and judges. If the judgment result is idle, then access the perceived frequency band for communication, otherwise wait until the next frame to start sensing the frequency band again;

The perception result classification module is used to classify the spectrum sensing results of the UAV, and calculate the average throughput and average energy consumption of each frame of the cognitive UAV network according to the classification results of the spectrum sensing;

The energy efficiency feedback module is used to define the energy efficiency as the reciprocal of the energy consumed per bit under the unit bandwidth, calculate the energy efficiency, and adjust the UAV perception time according to the energy efficiency.

Preferably, the cognitive UAV network filters the received signals at the sensing nodes of each cognitive UAV to obtain signals in the sensing frequency band range; samples the signals in the sensing frequency band range to obtain continuous sensing frequency bands discrete samples of the signal in range;

Performing periodogram power spectrum estimation on the discrete samples collected within the range of the sensing frequency band to obtain the spectral line intensity of the discrete samples;

Take half of the spectral line intensities for summing and segmentation to represent the power spectrum; upload the power spectrum sampled by N cognitive drones to the central drone to obtain the estimated periodic spectrum of all drones average value;

Divide all UAV periodic atlases evenly into several groups to obtain the spectral line intensity of each group; perform normalization according to the spectral line intensity of each group and the sum of estimated average values of all UAV periodic atlases to obtain normalized chemical spectrum;

Preset the false alarm probability of a single-segment normalized signal, obtain the judgment threshold under the preview setting, and obtain the detection probability of a single-segment normalized signal according to the judgment threshold;

The central UAV adopts a fusion strategy to make a final judgment, and obtains false alarm probability and detection probability.

The present invention realizes spectrum sensing with low energy consumption and high detection performance in the case of high dynamic changes in background noise; compared with the prior art, it achieves the following technical effects:

1. The present invention takes energy efficiency and detection probability as the optimization target, designs a reasonable collaborative sensing strategy, studies the influence of the UAV fusion threshold k and sensing time τ on the optimization target, and then selects the best parameters to ensure the detection performance. Under the premise of achieving the goal of green energy saving, while improving the anti-noise performance, it reduces resource overhead.

2. The present invention can realize high-performance and low-consumption spectrum sensing for frequency bands under the condition of high dynamic changes in background noise; normalize the sampling sequence to resist the impact of high dynamic noise on spectrum sensing performance, and at the same time It can reduce the dimension of the input to a certain extent and reduce the amount of calculation; it adopts the collaborative spectrum sensing method of soft combining, which can overcome the influence of multipath fading and shadow effect on the sensing results compared with single-user spectrum sensing.

3. The present invention takes energy efficiency and detection probability as the optimization target, designs a reasonable collaborative sensing strategy, studies the influence of the perception time of the UAV and the number of segments of the normalization algorithm on the optimization target, and then selects the best parameters so that in Under the premise of ensuring the detection performance, the goal of green energy saving can be achieved.

4. The present invention can reduce unnecessary resource overhead while ensuring the accuracy of perception. Since most drones are powered by batteries and the background noise in the process of performing tasks is highly dynamic, the present invention is suitable for drone scenarios. The following spectrum sensing has practical significance.

Description of drawings

Fig. 1 is a network model diagram of a cognitive unmanned aerial vehicle provided by an embodiment of the present invention, wherein (a) is a 3D diagram, and (b) is a 2D diagram;

Fig. 2 is a comparison diagram of the obtained energy efficiency, perception time and number of segments under the "1-out-of-N" fusion strategy provided by the embodiment of the present invention, as well as the comparison with the energy detection algorithm ;

Fig. 3 is the relationship between the obtained energy efficiency, the perception time and the number of segments, and the comparison with the energy detection algorithm under the "2-out-of-N" fusion strategy provided by the embodiment of the present invention ;

Fig. 4 is a comparison diagram of the energy efficiency of the sensing method provided by the embodiment of the present invention and the energy detection algorithm adopted with the probability of a single decision-making false alarm;

Fig. 5 is a diagram of the relationship between the perception time and the number of segments under the optimized condition obtained by the embodiment of the present invention;

Fig. 6 is a graph showing the relationship between energy efficiency and the number of segments under the optimal condition obtained in the embodiment of the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

Example 1

This embodiment discloses a normalized UAV spectrum sensing method based on energy efficiency. The cognitive UAV network model is shown in Figure 1, including multiple cognitive UAVs and a central UAV, which includes Follow the steps below:

S1. The cognitive drone senses the frequency band, and reports the sensing data to the central drone one by one, and the central drone averages and normalizes the sensing data for judgment. If the judgment result is idle, then the Enter the sensing frequency band for communication, such as entering the data transmission stage of the cognitive user, otherwise wait until the next frame to start sensing the frequency band again.

In the perception phase, all cognitive drones perform perception operations at the same time; in the reporting phase, all cognitive drones report the perception data to the central drone in turn, and finally decide the cognitive user according to the judgment result of the central drone. Whether it can access the perceived frequency band for communication.

In this embodiment, the process of frequency band sensing includes the following steps:

(1) Design the cognitive UAV network. The designed cognitive UAV network is mainly composed of N cognitive UAVs and a central UAV FC formation network, as shown in Figure 1. It is known that the UAV performs spectrum sensing on a circle centered on the primary user transmitter PU. Assuming the channel is a line-of-sight (LoS) channel, the channel gain h _ij (r _ij ) is expressed as:

where r _ij represents the distance between node i and node j; PL _ij (r _ij ) represents the path loss between node i and node j, and its expression is:

Among them, c represents the speed of light, f represents the carrier frequency, and L _LoS is the average additional loss of the link. The signal received by the UAV can be written as:

Where w(t) is Gaussian white noise, s(t) is the signal sent by the primary user, h _PU (t) is the channel gain between the signal sent by the primary user and the UAV, H ₀ means that the primary user does not occupy the Frequency band, H ₁ indicates that the primary user is occupying this frequency band for communication.

In this step, it is assumed that all cognitive drones fly at the same altitude and speed, and the distance between cognitive drones is negligible relative to the distance from the drone to the main user. Furthermore, since the flight time (s) of the UAV is much higher than the channel coherence time (ms), the system performance achieved by the instantaneous channel cannot be studied. Therefore, this embodiment mainly focuses on the average statistics of the channel, and only considers the large-scale fading in the channel.

(2) Filter the received signal at the sensing nodes of each cognitive UAV to obtain the signal x _n (t) within the sensing frequency range, where n=1, 2,..., N represents the serial number of the UAV; Sampling the signals in the perceptual frequency range to obtain discrete samples x _n (m) of signals in the continuous perceptual frequency range, where m=0, 1, . . . , M−1.

(3) Periodogram power spectrum estimation is carried out to the discrete samples collected within the range of the sensing frequency band:

Get the spectral line intensity of the discrete sample. Among them, the periodic spectrum estimation of the sampled signal is to take time average to obtain a flatter power spectrum.

(4) Since the sampling signal, that is, the discrete sample x _n (m) is a real signal, its power spectrum is symmetrical, so it is only necessary to take half of the spectral line intensity for summing and segmentation, that is, only M/2 P _n (k) can represent the power spectrum. For the signal received by the UAV, its distribution can be expressed as:

Therefore, there are:

当主用户存在时，P_avg(k)的均值和方差分别为/>

When the primary user does not exist, the mean and variance of P _avg (k) are

When the primary user exists, the mean and variance of P _avg (k) are />

Upload the power spectrum sampled by N cognitive UAVs to the central UAV, and obtain the average value of N (that is, all) UAV period spectrum estimates:

(5) Divide all UAV periodic atlases evenly into several groups to obtain the spectral line intensity of each group; carry out normalization process according to the spectral line intensity of each group and the sum of estimated average values of all UAV periodic atlases, Get a normalized spectrum.

将所有无人机周期图谱均匀分成U组，每一组的谱线强度记为P_seg(u)，则：The sum of all spectral line intensities (i.e. the average value of all UAV periodic atlas estimates) is defined as P _all , namely

Divide all UAV periodic spectra evenly into U groups, and record the spectral line intensity of each group as P _seg (u), then:

u＝0,1,…,N-1，/>

则为归一化谱。The ratio of P _seg (u) to P _all can be expressed as

u=0,1,...,N-1, />

is the normalized spectrum.

Y的均值和方差分别为/>

在主用户没有使用该频段的情况下，由中心无人机对接收到的每一个认知无人机检测到的该频段信号取平均后进行归一化处理得到的值，其分布与噪声功率无关，因此归一化的频谱感知算法在噪声电平高度变化的情况下更具有鲁棒性。In this embodiment, it is assumed that random variables X(u)=P _seg (u) and Y=P _all . According to the central limit theorem, when the number of sampling points is large enough, the random variables X(u) and Y are close to Gaussian distribution. Therefore, the mean and variance of X(u) are

The mean and variance of Y are respectively />

In the case that the main user does not use this frequency band, the central UAV averages the received signal of this frequency band detected by each cognitive UAV and then normalizes the value obtained, and its distribution is related to the noise power Therefore, the normalized spectrum sensing algorithm is more robust in the presence of highly variable noise levels.

(6) Preset the false alarm probability of the single-segment normalized signal, and calculate the judgment threshold under the preview setting according to the Neyman-Pearson criterion, and obtain the detection probability of the single-segment normalized signal according to the judgment threshold.

The false alarm probability and detection probability of a single-segment normalized signal are:

则可计算出门限γ为：The false alarm probability of a given single-segment signal is

Then the threshold γ can be calculated as:

Φ(·)是误差函数，Φ^-1(·)是误差函数的反函数。in

Φ(·) is an error function, and Φ ^-1 (·) is an inverse function of the error function.

(7) The central UAV makes a final judgment according to the "k out of N (k-out-of-N)" fusion criterion, and obtains the false alarm probability and detection probability.

In this embodiment, the k-out-of-N fusion strategy is adopted for the judgment results obtained in each section, and the false alarm probability _QF and the detection probability _QD can be obtained as follows:

Where k is the fusion threshold, u=1, 2,..., U, l=k, 1,..., U-1.

S2. Classify the spectrum sensing results of the UAV, and calculate the average throughput and average energy consumption of each frame of the cognitive UAV network according to the classification results of the spectrum sensing.

In this embodiment, the spectrum sensing results of the UAV are divided into four types: the UAV accurately detects that the sensing frequency band is idle; the UAV accurately detects that the sensing frequency band is occupied; The frequency band is idle; the UAV has a false alarm and wrongly detects that the sensing frequency band is occupied.

When the UAV accurately detects that the sensing frequency band is occupied or the UAV falsely detects that the sensing frequency band is occupied, the cognitive users do not send signals, and the throughput is 0 at this time. When the UAV accurately detects that the sensing frequency band is idle and the UAV has missed detection, due to the influence of the interference signal of the main user, the throughput of the former is much greater than that of the latter. When calculating energy efficiency, the throughput of the latter can be calculated as The amount is ignored.

According to the four kinds of spectrum sensing results, the average throughput and average energy consumption of each frame of the cognitive UAV network are calculated as:

E＝NτP _s +Nt _r P _r +NT(P _hov +P _fly )+(T-τ-Nt _r )P _U [π ₀ (1-Q _F )+π ₁ (1-Q _D )]

Among them, τ represents the perception time, t _r represents the reporting time of a single UAV, and T represents the time of each frame; π ₀ represents the probability that the main user of the channel is not occupied, and Q _F and Q _D are respectively the center UAV according to " Take k(k-out-of-N) out of N" to determine the false alarm probability and detection probability, and h _cs (t) represents the large-scale channel gain between the UAV and the ground receiver. P _u is the transmission power of the UAV, P _s is the perception power of the UAV, and P _r is the transmission power of the UAV. _Phov and P _fly represent the flying power and hovering power of the UAV, respectively.

S3, energy efficiency is defined as the reciprocal of the energy consumed per bit under the unit bandwidth, then can draw the display expression of energy efficiency η _EE :

The energy efficiency is calculated according to the above expression; the UAV perception time is adjusted according to the energy efficiency, so that both perception and energy efficiency are optimized.

In order to verify the excellent performance of the normalized sensing method of the present invention under the background of high dynamic noise, the normalized sensing method of the present invention is compared with the existing energy detection algorithm, as follows:

(1) Set the parameters as shown in the table below:

parameter value parameter value parameter value R/m 340 T/s 0.1 π ₁ 0.3 H/m 60 N 8 _L 3 f _s /kHz 300 d/m 5 t _r /ms 0.1 P _r /mW 10 P _U /W 6 _f 0.01 P _S /mW 40 P _P /W 10 v/m·s ^-1 10 P _fly /W 5 P _zero /W 0 f _c /GHz 2.4 m _tot /kg 0.6 r _p /m 0.2 alpha ₀ π/3

(2) Perform simulation according to the parameters given in the above table, and obtain Figure 2 and Figure 3. From the figure, it can be seen that the normalized perception method adopts the "2-out-of-N" method The judgment method is better than the "1-out-of-N" judgment method adopted by the existing energy detection algorithm, so the subsequent fusion threshold is set to 2 for simulation.

(3) In order to further observe whether the normalized perception method can be applied to the energy efficiency optimization problem in the noisy high dynamic scene, the signal-to-noise ratio of the signal received by the UAV is randomly set from 5dB to 10dB, and the obtained results are shown in the figure 4.

(4) According to the previous simulation results, it can be concluded that both the perception time τ and the segment number U have an impact on the energy efficiency of the normalization algorithm. The best perception time and energy efficiency under different number of segments obtained by simulation are shown in Figure 5 and Figure 6, respectively.

Example 2

Based on the same inventive concept as Embodiment 1, this embodiment provides a normalized UAV spectrum sensing system based on energy efficiency, including:

Cognitive UAV network. The cognitive UAV network includes N cognitive UAVs and a central UAV. The cognitive UAV senses the frequency band and reports the sensing data to the central UAV one by one. Human-machine, and the central drone averages and normalizes the sensing data and judges. If the judgment result is idle, then access the perceived frequency band for communication, otherwise wait until the next frame to start sensing the frequency band again;

The perception result classification module is used to classify the spectrum sensing results of the UAV, and calculate the average throughput and average energy consumption of each frame of the cognitive UAV network according to the classification results of the spectrum sensing;

The energy efficiency feedback module is used to define the energy efficiency as the reciprocal of the energy consumed per bit under the unit bandwidth, calculate the energy efficiency, and adjust the UAV perception time according to the energy efficiency.

Among them, the cognitive drone network filters the received signals at the sensing nodes of each cognitive drone to obtain signals in the sensing frequency range; samples the signals in the sensing frequency range to obtain continuous sensing frequency ranges discrete samples of the inner signal;

Performing periodogram power spectrum estimation on the discrete samples collected within the range of the sensing frequency band to obtain the spectral line intensity of the discrete samples;

Take half of the spectral line intensities for summing and segmentation to represent the power spectrum; upload the power spectrum sampled by N cognitive drones to the central drone to obtain the estimated periodic spectrum of all drones average value;

Divide all UAV periodic atlases evenly into several groups to obtain the spectral line intensity of each group; perform normalization according to the spectral line intensity of each group and the sum of estimated average values of all UAV periodic atlases to obtain normalized chemical spectrum;

Preset the false alarm probability of a single-segment normalized signal, obtain the judgment threshold under the preview setting, and obtain the detection probability of a single-segment normalized signal according to the judgment threshold;

The central UAV adopts a fusion strategy to make a final judgment, and obtains false alarm probability and detection probability.

For a more detailed implementation process of implementing spectrum sensing through various modules in this embodiment, refer to Embodiment 1, and details are not repeated here.

Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be construed as limitations to the present invention. Variations, modifications, substitutions, and modifications to the above-described embodiments are possible within the scope of the present invention.

Claims (9)

1. The energy efficiency-based normalized unmanned aerial vehicle frequency spectrum sensing method is characterized by comprising the following steps of:

the cognitive unmanned aerial vehicle perceives the frequency bands, the perceiving data are reported to the central unmanned aerial vehicle one by one, the central unmanned aerial vehicle carries out average and normalization processing judgment on the perceiving data, if the judgment result is idle, the perceived frequency bands are accessed for communication, otherwise, the next frame starts to perceive the frequency bands again;

classifying the spectrum sensing result of the unmanned aerial vehicle, and calculating the average throughput sum of each frame of the cognitive unmanned aerial vehicle network according to the classification result of the spectrum sensing;

and defining the energy efficiency as the reciprocal of energy consumed per bit under the unit bandwidth, calculating the energy efficiency, and adjusting the perception time of the unmanned aerial vehicle according to the energy efficiency.

2. The spectrum sensing method according to claim 1, wherein the process of frequency band sensing comprises the steps of:

designing a cognitive unmanned aerial vehicle network, wherein the designed cognitive unmanned aerial vehicle network comprises N cognitive unmanned aerial vehicles and a central unmanned aerial vehicle, and the cognitive unmanned aerial vehicle carries out frequency spectrum sensing on a circle taking a main user transmitter as a center;

filtering the received signals at the sensing nodes of each cognitive unmanned aerial vehicle to obtain signals in a sensing frequency range; sampling the signals in the sensing frequency range to obtain discrete samples of the signals in the continuous sensing frequency range;

performing periodic chart power spectrum estimation on the discrete samples acquired in the sensing frequency range to obtain spectral line intensity of the discrete samples;

taking half of spectral line intensity for summation and segmentation, and representing a power spectrum; uploading the power spectrums obtained by sampling to a central unmanned aerial vehicle by N cognitive unmanned aerial vehicles to obtain an average value of periodic spectrum estimation of all unmanned aerial vehicles;

uniformly dividing all the unmanned aerial vehicle periodic patterns into a plurality of groups to obtain spectral line intensity of each group; normalizing according to the spectral line intensity of each group and the sum of the estimated average values of all the unmanned aerial vehicle periodic patterns to obtain normalized spectrums;

presetting false alarm probability of a single-segment normalized signal, solving a decision threshold under preset learning setting, and obtaining detection probability of the single-segment normalized signal according to the decision threshold;

and the central unmanned plane adopts a fusion strategy to make a final judgment so as to obtain the false alarm probability and the detection probability.

3. The spectrum sensing method according to claim 2, wherein performing periodogram power spectrum estimation on discrete samples acquired within the sensing frequency band range is:

wherein the discrete samples are x _n (m),m＝0,1,…,M-1。

4. A spectrum sensing method according to claim 3, wherein the average value of the unmanned aerial vehicle periodic map estimates is:

where n=1, 2, …, N represents the number of the drone.

5. The spectrum sensing method of claim 2, wherein the spectral line intensity of each group is:

the normalized spectrum is:

wherein P is _all The sum of the average values is estimated for all unmanned aerial vehicle periodic patterns.

6. The spectrum sensing method according to claim 2, wherein for the signal received by the unmanned aerial vehicle, the distribution condition is expressed as:

x(n):

therefore, there are:

mean value P of unmanned aerial vehicle periodic map estimation when main user does not exist _avg (k) Mean and variance of (a) are respectively

P when the primary user exists _avg (k) Mean and variance of>

Let X (u) =p be a random variable _seg (u) and y=p _all The mean and variance of the random variable X (u) are respectively

The mean and variance of the random variable Y are +.>

The false alarm probability and the detection probability of the single-section normalized signal are respectively as follows:

the false alarm probability of a given single-stage signal is

The threshold γ is calculated as:

wherein the method comprises the steps of

Phi (·) is the error function, phi ^-1 (. Cndot.) is the inverse of the error function.

7. The spectrum sensing method of claim 2, wherein the false alarm probability Q _F And detection probability Q _D The method comprises the following steps of:

where k is the fusion threshold, P _f For the false alarm probability of a single segment normalized signal,

for the detection probability of the single-segment normalized signal, u=1, 2, …, U, l=k, 1, …, U-1.

8. Energy efficiency-based normalized unmanned aerial vehicle frequency spectrum sensing system, characterized by comprising:

the cognitive unmanned aerial vehicle network comprises N cognitive unmanned aerial vehicles and a central unmanned aerial vehicle, the cognitive unmanned aerial vehicle perceives the frequency bands, perceived data are reported to the central unmanned aerial vehicle one by one, the central unmanned aerial vehicle carries out average and normalization processing judgment on the perceived data, if the judgment result is idle, the perceived frequency bands are accessed for communication, otherwise, the next frame begins to perceive the frequency bands again;

the sensing result classification module is used for classifying the spectrum sensing result of the unmanned aerial vehicle and calculating the average throughput sum average energy consumption of each frame of the cognitive unmanned aerial vehicle network according to the spectrum sensing classification result;

and the energy efficiency feedback module is used for defining the energy efficiency as the reciprocal of energy consumed by each bit under the unit bandwidth, calculating the energy efficiency and adjusting the perception time of the unmanned aerial vehicle according to the energy efficiency.

9. The spectrum sensing system of claim 8, wherein the cognitive unmanned aerial vehicle network filters the received signals at the sensing nodes of each cognitive unmanned aerial vehicle to obtain signals within a sensing frequency range; sampling the signals in the sensing frequency range to obtain discrete samples of the signals in the continuous sensing frequency range;

performing periodic chart power spectrum estimation on the discrete samples acquired in the sensing frequency range to obtain spectral line intensity of the discrete samples;

taking half of spectral line intensity for summation and segmentation, and representing a power spectrum; uploading the power spectrums obtained by sampling to a central unmanned aerial vehicle by N cognitive unmanned aerial vehicles to obtain an average value of periodic spectrum estimation of all unmanned aerial vehicles;

uniformly dividing all the unmanned aerial vehicle periodic patterns into a plurality of groups to obtain spectral line intensity of each group; normalizing according to the spectral line intensity of each group and the sum of the estimated average values of all the unmanned aerial vehicle periodic patterns to obtain normalized spectrums;

presetting false alarm probability of a single-segment normalized signal, solving a decision threshold under preset learning setting, and obtaining detection probability of the single-segment normalized signal according to the decision threshold;

and the central unmanned plane adopts a fusion strategy to make a final judgment so as to obtain the false alarm probability and the detection probability.

Images