A method for unmanned aerial vehicle signal spectrum sensing access

Abstract

The application discloses a kind of unmanned aerial vehicle signal spectrum sensing access methods, belong to spectrum sensing technical field, specifically include: acquisition unmanned aerial vehicle three-dimensional velocity and acceleration, synchronous reception baseband signal and construct multidimensional parameter set;According to three-dimensional velocity and acceleration, calculate frequency shift compensation factor, extract channel fading sequence by envelope detection;According to the fluctuation frequency of channel fading sequence and frequency shift compensation factor, determine dynamic sensing period, and combine fading depth with current ambient noise level to generate dynamic sensing threshold;In dynamic sensing period, frequency offset correction is carried out to baseband signal, the energy characteristics of corrected signal are calculated, and the spectrum occupation state is determined according to the comparison result of energy characteristics and dynamic sensing threshold;When spectrum is idle, tune the radio frequency unit of unmanned aerial vehicle to nominal center frequency to establish link;The application can reduce the lag and misjudgment of spectrum sensing result under high dynamic channel, improve spectrum sensing accuracy and spectrum access success rate.

Description

Technical Field

[0001] This invention relates to the field of spectrum sensing technology, and specifically to a method for accessing the spectrum sensing of unmanned aerial vehicle (UAV) signals. Background Technology

[0002] Unmanned aerial vehicle (UAV) wireless communication typically involves receiving and processing spatial electromagnetic signals via an onboard radio frequency (RF) unit, and selecting an appropriate frequency band to establish a communication link based on the channel environment. Spectrum sensing is a crucial foundation for UAVs to implement dynamic spectrum access. It involves collecting baseband signals within candidate frequency bands, extracting signal energy, power spectrum, or cyclic characteristics, and comparing these extracted features with pre-set sensing thresholds to determine whether the candidate frequency band is occupied or idle. In existing technologies, UAVs can repeatedly perform spectrum sensing according to a set sensing cycle. When a candidate frequency band is determined to be idle, the onboard RF unit is controlled to tune to the corresponding frequency, and spectrum access is completed by sending link handshake data.

[0003] However, in complex three-dimensional spatial scenarios where UAVs cruise at high speeds or maneuver agilely, the high dynamic movement of the UAV causes significant Doppler shifts and rapid fading of received signals, resulting in rapid changes in the wireless channel state. Existing spectrum sensing technologies are typically designed for static or low-speed moving environments, often employing fixed sensing thresholds and long sensing periods. These technologies struggle to adapt to the rapidly changing time-varying characteristics of high-dynamic channels, easily causing spectrum sensing results to lag behind the actual spectrum state. When a UAV performs channel access based on lagging spectrum sensing results, it may misjudge the frequency band of an already active licensed user as an idle band, leading to signal collisions with licensed users. It may also misjudge short-lived idle spectrum as occupied, missing available idle spectrum resources, thus reducing spectrum sensing accuracy and spectrum access success rate. Summary of the Invention

[0004] The purpose of this invention is to provide a method for unmanned aerial vehicle (UAV) signal spectrum sensing and access, thereby solving the problems in the background art.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A method for unmanned aerial vehicle (UAV) signal spectrum sensing and access includes the following steps:

[0007] S1: Collect the three-dimensional velocity and acceleration of the UAV, synchronously receive the baseband signal, and construct a multi-dimensional parameter set to characterize the high dynamic channel state;

[0008] S2: Based on a multi-dimensional parameter set, the frequency shift compensation factor is calculated using three-dimensional velocity and acceleration, and envelope detection is performed on the baseband signal to extract the channel fading sequence;

[0009] S3: Use the fluctuation frequency of the channel fading sequence to set the basic sensing window, use the frequency shift compensation factor as a weight to weight the basic sensing window, and output the dynamic sensing period.

[0010] S4: Calculate the dynamic sensing threshold for adapting to high dynamic channels by combining the fading depth of the channel fading sequence and the current environmental noise level.

[0011] S5: During the dynamic sensing period, the baseband signal is frequency offset corrected using the frequency shift compensation factor, and the energy characteristics of the corrected signal are calculated. The corrected signal is then compared with the dynamic sensing threshold to determine the current spectrum occupancy status.

[0012] S6: When the spectrum occupancy status is determined to be idle, obtain the pre-configured nominal center frequency, tune the UAV radio frequency unit to the nominal center frequency to establish a link, and complete the spectrum access.

[0013] As a further aspect of the present invention: in step S1, the process of constructing a multidimensional parameter set for characterizing the high-dynamic channel state is as follows:

[0014] The airborne inertial measurement unit and satellite positioning equipment are activated to obtain three-dimensional velocity and acceleration, and the radio frequency receiving front-end is scheduled in parallel to perform down-conversion and discrete sampling of space electromagnetic waves to generate baseband signals;

[0015] The timing pulses output by satellite positioning equipment are extracted to construct a time reference, and a unified timestamp is assigned to the three-dimensional velocity, acceleration and baseband signals. Based on the time series alignment and fusion, a multi-dimensional parameter set is constructed to characterize the high dynamic channel state.

[0016] As a further aspect of the present invention: in step S2, the process of calculating the frequency shift compensation factor based on the multidimensional parameter set using three-dimensional velocity and acceleration, and performing envelope detection operation on the baseband signal to extract the channel fading sequence is as follows:

[0017] The three-dimensional velocity and acceleration of the multi-dimensional parameter set are analyzed, and the three-dimensional velocity and acceleration are superimposed by directional projection along the signal transmission path to calculate the instantaneous radial velocity of the UAV.

[0018] Divide the instantaneous radial velocity by the speed of light to obtain the change ratio, multiply the change ratio by the center carrier frequency to calculate the Doppler frequency shift, and invert the Doppler frequency shift to construct the frequency shift compensation factor.

[0019] The baseband signal within the multidimensional parameter set is extracted, and the square root of the sum of the squares of the in-phase and quadrature components of the baseband signal is calculated to obtain the instantaneous amplitude waveform. The instantaneous amplitude waveform is then subjected to low-pass filtering and smoothing to extract the channel fading sequence.

[0020] As a further aspect of the present invention: In step S3, the process of setting a basic sensing window using the fluctuation frequency of the channel fading sequence, weighting the basic sensing window with the frequency shift compensation factor as a weight, and outputting the dynamic sensing period is as follows:

[0021] Calculate the mean amplitude of the channel fading sequence, count the number of intersections of the channel fading sequence across the mean amplitude, and divide the number of intersections by the total duration of the sequence to obtain the fluctuation frequency.

[0022] The reference duration is obtained by performing a reciprocal operation on the fluctuation frequency. The reference duration is set as the perception cutoff boundary. The basic perception window is set by dividing the time axis interval according to the perception cutoff boundary.

[0023] The absolute value of the frequency shift compensation factor is extracted and normalized to generate a weighting coefficient. The weighting coefficient is summed with the set period offset constant to generate an adjustment parameter. The adjustment parameter is multiplied by the reference duration of the basic sensing window and weighted to output the dynamic sensing period.

[0024] As a further aspect of the present invention: in step S4, the process of calculating the dynamic sensing threshold adapted to the high dynamic channel by combining the fading depth of the channel fading sequence and the current environmental noise level is as follows:

[0025] The local envelope peak and local envelope valley values ​​are extracted by traversing the channel fading sequence. The local envelope peak values ​​are subtracted from the local envelope valley values ​​to calculate each amplitude difference component. The mean of all amplitude difference components is calculated as the fading depth of the channel fading sequence.

[0026] Extract spatial noise samples under passive radiation conditions to construct a noise sequence, perform variance calculation on the noise sequence to obtain the noise floor fluctuation amplitude, and set the noise floor fluctuation amplitude as the current environmental noise level;

[0027] The fading depth is multiplied by a preset margin adjustment coefficient to generate a dynamic bias term. The current ambient noise level is summed with the dynamic bias term in the linear domain to obtain the basic decision threshold. The basic decision threshold is output as a dynamic sensing threshold adapted to the high dynamic channel.

[0028] As a further aspect of the present invention: in step S5, the process of correcting the frequency offset of the baseband signal using a frequency shift compensation factor within the dynamic sensing period and calculating the energy characteristics of the corrected signal is as follows:

[0029] A time truncation window is set with the dynamic sensing period, and a time domain truncation operation is performed on the baseband signal to extract the current sensing segment. The time series of the current sensing segment is extracted, and a complex exponential compensation sequence is constructed by combining the time series with the frequency shift compensation factor.

[0030] The complex exponential compensation sequence is multiplied element-wise with the current sensing segment to complete the frequency offset correction of the current sensing segment and output the corrected signal.

[0031] The discrete amplitude sequence of the corrected signal is analyzed, and the discrete amplitude sequence is squared to generate instantaneous power data. The instantaneous power data is accumulated and summed to output the energy characteristics.

[0032] As a further aspect of the present invention: the specific content of comparing it with the dynamic sensing threshold to determine the current spectrum occupancy status is as follows:

[0033] Extract the energy characteristics and dynamic sensing threshold of the corrected signal, subtract the value of the dynamic sensing threshold from the value of the energy characteristics, and perform a difference operation to generate the state deviation.

[0034] The sign characteristics of the analytical state deviation are analyzed. If the sign characteristic is positive, the current spectrum occupancy status is determined to be occupied; if the sign characteristic is negative or zero, the current spectrum occupancy status is determined to be idle.

[0035] As a further aspect of the present invention: In step S6, when the spectrum occupancy status is determined to be idle, the process of obtaining the pre-configured nominal center frequency, tuning the UAV radio frequency unit to the nominal center frequency to establish a link, and completing the spectrum access is as follows:

[0036] Obtain the determination result that the spectrum occupancy status is idle, and extract the nominal center frequency pre-configured for the current sensing frequency band;

[0037] The nominal center frequency is input into the frequency synthesizer of the UAV's radio frequency unit, which drives the voltage-controlled oscillator to generate a local oscillator carrier with the same nominal center frequency to complete the tuning. The radio frequency front end is then controlled to transmit handshake data packets based on the local oscillator carrier to establish a link and complete spectrum access.

[0038] The beneficial effects of this invention are:

[0039] This invention synchronously acquires the three-dimensional velocity, acceleration, and baseband signals of a UAV and uses a unified timestamp for alignment and fusion, enabling accurate correlation between motion state data and corresponding wireless channel data. Based on this, instantaneous radial velocity and frequency shift compensation factors are determined according to the three-dimensional velocity and acceleration, while channel fading sequences are extracted from the baseband signals. The fluctuation frequency of the channel fading sequence is then used to set a basic sensing window, and the dynamic sensing period is adjusted in conjunction with the frequency shift compensation factor. This allows for timely shortening of the sensing interval when the UAV's speed increases or channel fluctuations accelerate, reducing the lag in spectrum state updates caused by a fixed sensing period and improving the temporal consistency between the sensing results and the current spectrum state.

[0040] This invention generates a dynamic sensing threshold by combining the channel fading depth and the current ambient noise level, enabling the decision threshold to adapt to different fading levels and noise floor fluctuations, reducing misjudgments of occupancy status caused by fixed thresholds in deep fading environments. During the dynamic sensing period, a frequency shift compensation factor is used to correct the frequency offset of the baseband signal, and energy features are extracted under a unified statistical scale for spectrum occupancy status determination, reducing signal energy deviation from the sensing frequency band caused by Doppler shift. When the spectrum is determined to be idle, the UAV's radio frequency unit tunes to the pre-configured nominal center frequency and establishes a link through a handshake confirmation, thereby reducing the possibility of signal collisions with authorized users who have resumed activity and improving the utilization capability of short-term idle spectrum and the spectrum access success rate. Attached Figure Description

[0041] The invention will now be further described with reference to the accompanying drawings.

[0042] Figure 1 This is a flowchart illustrating a method for unmanned aerial vehicle (UAV) signal spectrum sensing and access according to the present invention.

[0043] Figure 2 This is a schematic diagram illustrating the construction process of a multi-dimensional parameter set for a UAV signal spectrum sensing access method according to the present invention.

[0044] Figure 3 This is a schematic diagram of the frequency shift compensation factor and channel fading sequence extraction process in a UAV signal spectrum sensing access method according to the present invention.

[0045] Figure 4 This is a schematic diagram of the dynamic sensing cycle generation process of a UAV signal spectrum sensing access method according to the present invention.

[0046] Figure 5 This is a schematic diagram of the dynamic sensing threshold generation process of a UAV signal spectrum sensing access method according to the present invention.

[0047] Figure 6 This is a schematic diagram of the frequency offset correction and spectrum occupancy status determination process of a UAV signal spectrum sensing access method according to the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Please see Figure 1 As shown, the present invention is a method for unmanned aerial vehicle (UAV) signal spectrum sensing and access, comprising the following steps:

[0050] S1: Collect the three-dimensional velocity and acceleration of the UAV, synchronously receive the baseband signal, and construct a multi-dimensional parameter set to characterize the high dynamic channel state;

[0051] S2: Based on a multi-dimensional parameter set, the frequency shift compensation factor is calculated using three-dimensional velocity and acceleration, and envelope detection is performed on the baseband signal to extract the channel fading sequence;

[0052] S3: Use the fluctuation frequency of the channel fading sequence to set the basic sensing window, use the frequency shift compensation factor as a weight to weight the basic sensing window, and output the dynamic sensing period.

[0053] S4: Calculate the dynamic sensing threshold for adapting to high dynamic channels by combining the fading depth of the channel fading sequence and the current environmental noise level.

[0054] S5: During the dynamic sensing period, the baseband signal is frequency offset corrected using the frequency shift compensation factor, and the energy characteristics of the corrected signal are calculated. The corrected signal is then compared with the dynamic sensing threshold to determine the current spectrum occupancy status.

[0055] S6: When the spectrum occupancy status is determined to be idle, obtain the pre-configured nominal center frequency, tune the UAV radio frequency unit to the nominal center frequency to establish a link, and complete the spectrum access.

[0056] Please see Figure 2 As shown, in one embodiment of the present invention, the process of acquiring the three-dimensional velocity and acceleration of the UAV, synchronously receiving baseband signals, and constructing a multi-dimensional parameter set to characterize the high-dynamic channel state in step S1 is as follows:

[0057] In this embodiment, the airborne inertial measurement unit adopts a six-axis microelectromechanical inertial measurement unit, which includes a three-axis accelerometer and a three-axis gyroscope. The satellite positioning equipment adopts a dual-mode positioning receiver that supports BeiDou satellite navigation signals and Global Positioning Satellite Navigation signals. The radio frequency receiving front end adopts a software radio receiver that can digitally receive preset sensing frequency bands.

[0058] The airborne inertial measurement unit and satellite positioning equipment are activated. The airborne inertial measurement unit acquires the acceleration data of the UAV in the three coordinate axes at a sampling frequency of 200 Hz, and uses the internal attitude calculation function to perform coordinate transformation and gravity component subtraction on the acceleration data to generate three-dimensional acceleration. The satellite positioning equipment acquires the velocity data of the UAV in the three coordinate axes at an output frequency of 20 Hz to generate three-dimensional velocity.

[0059] While acquiring three-dimensional velocity and acceleration, a field-programmable gate array (FPGA) is used to schedule the radio frequency (RF) receiving front-end in parallel. The RF receiving front-end includes a receiving antenna, a bandpass filter, a low-noise amplifier, a downconverter, and an analog-to-digital converter (ADC). The receiving antenna receives spatial electromagnetic waves, the bandpass filter selects spatial electromagnetic waves within a preset sensing frequency band, the low-noise amplifier amplifies the selected spatial electromagnetic waves with low noise, the downconverter converts the amplified spatial electromagnetic waves to baseband, and the ADC performs discrete sampling on the converted signal to generate a baseband signal containing in-phase and quadrature components.

[0060] The satellite positioning equipment outputs a timing pulse once per second, inputs the timing pulse into the timing interface of the field-programmable gate array (FPGA), and uses the hardware counter inside the FPGA to construct a time reference. The hardware counter is calibrated each time a timing pulse arrives, and a stable reference clock is used to keep the hardware counter continuously counting between two adjacent timing pulses to obtain continuous time information.

[0061] When the field-programmable gate array receives a three-dimensional velocity, it reads the current count value of the hardware counter and converts the current count value into a corresponding unified timestamp. When it receives a three-dimensional acceleration, it uses the same method to calibrate the unified timestamp. When the RF receiving front end starts to collect baseband signals, it records the sampling start time and calibrates the unified timestamp for each sampling point in the baseband signal according to the sampling start time and the discrete sampling frequency.

[0062] Before data acquisition, the data interface of the airborne inertial measurement unit and the sampling interface of the radio frequency receiving front-end are triggered separately using the same trigger signal. The data output delay of the airborne inertial measurement unit, the satellite positioning equipment and the radio frequency receiving front-end are measured. The measured data output delay is stored in the field programmable gate array. When calibrating a unified timestamp, the corresponding data output delay is deducted to reduce the impact of data transmission time of different devices on time alignment.

[0063] It should be noted that the three-dimensional velocity, acceleration, and baseband signal have different data update frequencies and cannot be directly matched according to the data sequence number. Therefore, the baseband signal is divided into continuous baseband signal segments according to the sampling time of the baseband signal, and the center time of each baseband signal segment is used as the alignment time. The three-dimensional velocities located on both sides of the alignment time are selected in the three-dimensional velocity time series and time interpolation is performed to obtain the three-dimensional velocity corresponding to the alignment time. The acceleration falling within the duration of the corresponding baseband signal segment is extracted in the acceleration time series, and the average result of the extracted acceleration is calculated to obtain the acceleration corresponding to the alignment time.

[0064] Following the chronological order of unified timestamps, the three-dimensional velocity, acceleration, and baseband signal segments corresponding to each alignment moment are combined into a data record. The continuously generated data records are sorted and stored to form a multi-dimensional parameter set for characterizing the high-dynamic channel state. When the timestamp of any data record is missing, the data output delay exceeds the allowable range, or the corresponding three-dimensional velocity and acceleration cannot be found, the data record is discarded to ensure the time correspondence between the three-dimensional velocity, acceleration, and baseband signal within the multi-dimensional parameter set.

[0065] Please see Figure 3 As shown, in one embodiment of the present invention, step S2, the process of calculating the frequency shift compensation factor based on a multi-dimensional parameter set using three-dimensional velocity and acceleration, and performing envelope detection operation on the baseband signal to extract the channel fading sequence, is as follows:

[0066] In this embodiment, three-dimensional velocity, acceleration, and baseband signals with the same unified timestamp are sequentially read from the multi-dimensional parameter set. The airborne digital signal processor is used to perform directional projection, Doppler frequency shift calculation, envelope detection operation, and low-pass filtering smoothing. The airborne digital signal processor is a digital signal processing chip with floating-point operation function. The baseband signal is output by the radio frequency receiving front end, and the center carrier frequency is read from the frequency configuration data of the radio frequency receiving front end.

[0067] The three-dimensional velocity in the multidimensional parameter set is analyzed into velocity components along the three coordinate axes, and the acceleration is analyzed into acceleration components along the three coordinate axes. The three-dimensional velocity and acceleration are transformed to the same coordinate system where the signal transmission path is located. The unit direction vector is determined according to the signal transmission path, and the three direction components of the unit direction vector satisfy the condition that the sum of the squares is one.

[0068] Taking the direction from the UAV pointing to the signal source as the positive direction of the directional projection, within a baseband signal segment, the projection result of the three-dimensional velocity along the signal transmission path is taken as the radial velocity at the beginning of the baseband signal segment. The projection result of the acceleration along the signal transmission path is accumulated according to the corresponding time interval to obtain the radial velocity increment generated by the acceleration within the time interval. The radial velocity is superimposed with the radial velocity increment to obtain the instantaneous radial velocity corresponding to each sampling moment within the baseband signal segment. The instantaneous radial velocity is calculated according to the following formula:

[0069] ;

[0070] in, Indicates the first in the baseband signal The instantaneous radial velocity corresponding to each sampling moment. Indicates the sampling sequence number of the baseband signal. This indicates the sequence number of the current data record in the multidimensional parameter set. and These represent the three velocity components in the current data record. and These represent the three acceleration components in the current data record. and These represent the three unit direction components of the signal transmission path. Indicates the first The time interval between each sampling time and the corresponding time of the current data record.

[0071] It should be noted that the directional projection superposition does not directly add the three-dimensional velocity and acceleration. Instead, the acceleration is first directionally projected along the signal transmission path, and then the projected acceleration is converted into radial velocity increments according to the corresponding time interval. Finally, the radial velocity increments are superimposed with the directional projection results of the three-dimensional velocity. This ensures that all data involved in the superposition are velocity data. During the duration of the current baseband signal segment, the acceleration is considered to remain constant. When the change in acceleration exceeds a preset range, the duration of the baseband signal segment is shortened and the time-varying radial velocity is recalculated.

[0072] After obtaining the instantaneous radial velocity, dividing the instantaneous radial velocity by the speed of light yields a ratio characterizing the relative change in carrier frequency caused by the UAV's motion. This ratio is calculated using the following formula:

[0073] ;

[0074] in, Indicates the first The ratio of changes corresponding to each sampling time point Indicates the first The instantaneous radial velocity corresponding to each sampling moment. In this embodiment, the speed of light is represented. The value is 299,792,458 meters per second, and the change ratio is a dimensionless value.

[0075] The center carrier frequency is extracted from the frequency configuration data of the radio frequency receiving front end. The change ratio is multiplied by the center carrier frequency to obtain the Doppler frequency shift at the corresponding sampling time. The Doppler frequency shift is calculated according to the following formula:

[0076] ;

[0077] in, Indicates the first Doppler frequency shift corresponding to each sampling time Indicates the first The ratio of changes corresponding to each sampling time point Indicates the center carrier frequency. Indicates the first The instantaneous radial velocity corresponding to each sampling moment. The speed of light is represented by the Doppler frequency shift, and the Doppler frequency shift uses the same frequency unit as the center carrier frequency.

[0078] In this embodiment, the instantaneous radial velocity is 60 meters per second, and the center carrier frequency is 2.4. At Hertz, the change ratio is approximately 2.001 × 10⁻⁶. The calculated Doppler frequency shift is approximately 480.3 Hz, indicating that the baseband signal has a frequency shift of approximately 480.3 Hz relative to the center carrier frequency.

[0079] Inverting the Doppler frequency shift yields the frequency shift compensation factor, which is determined according to the following formula: ;in, Indicates the first Frequency shift compensation factor corresponding to each sampling time. Indicates the first The Doppler frequency shift corresponding to each sampling time is negative. The negative sign indicates that the compensation direction of the frequency shift compensation factor is opposite to the Doppler frequency shift direction. When the Doppler frequency shift is positive, the frequency shift compensation factor is negative, and when the Doppler frequency shift is negative, the frequency shift compensation factor is positive.

[0080] The baseband signal corresponding to the current time within the multidimensional parameter set is extracted. The in-phase and quadrature components of the baseband signal are input into the airborne digital signal processor (DSP). The DSP performs squaring operations on the in-phase and quadrature components respectively. The two squaring results are added together, and the square root operation is performed on the sum to obtain the instantaneous amplitude waveform. The instantaneous amplitude waveform is calculated according to the following formula:

[0081] ;

[0082] in, Indicates the first in the baseband signal The instantaneous amplitude of each sampling point Indicates the first The in-phase components at each sampling point Indicates the first Orthogonal components of each sampling point Indicates the sampling sequence number of the baseband signal. Indicates the first The square of the complex amplitude at each sampling point.

[0083] In this embodiment, when the in-phase component of a baseband signal sampling point is 0.6 and the quadrature component is 0.8, the square of the in-phase component is 0.36 and the square of the quadrature component is 0.64. The sum of the two square operation results is 1. The instantaneous amplitude obtained by performing a square root operation on the sum is 1. All sampling points in the baseband signal are processed in the same way to form an instantaneous amplitude waveform arranged according to the sampling time.

[0084] The instantaneous amplitude waveform is input to a finite impulse response (FIR) low-pass filter implemented by an airborne digital signal processor. This FIR low-pass filter employs sixty-four filter coefficients, with the sum of all coefficients set to one. This suppresses rapid noise fluctuations in the instantaneous amplitude waveform and rapid envelope changes in the baseband signal. The low-pass filtering smoothing process is performed according to the following formula:

[0085] ;

[0086] in, This indicates the number of lines after low-pass filtering and smoothing. The channel fading sequence values ​​corresponding to each sampling point Indicates that it is located at the th The instantaneous amplitude before each sampling point, The first finite impulse response low-pass filter represents the... Each filter coefficient This represents the number of filter coefficients in a finite impulse response (FIR) low-pass filter. This indicates the index of the filter coefficients participating in the current low-pass filter operation.

[0087] It should be noted that the stop frequency of the finite impulse response (FIR) low-pass filter is set higher than the expected channel fading frequency but lower than the rapid envelope change frequency of the baseband signal. This is to preserve the slow amplitude changes caused by channel fading and suppress the effects of noise and rapid envelope changes on the channel fading sequence, thus continuously obtaining... Arranged according to a unified timestamp, a channel fading sequence is formed that corresponds to the three-dimensional velocity, acceleration, and baseband signal hold time.

[0088] Please see Figure 4 As shown, in one embodiment of the present invention, in step S3, the process of setting a basic sensing window using the fluctuation frequency of the channel fading sequence, weighting the basic sensing window with the frequency shift compensation factor as a weight, and outputting the dynamic sensing period is as follows:

[0089] In this embodiment, an airborne digital signal processor performs calculations on the channel fading sequence and frequency shift compensation factor. The airborne digital signal processor uses a digital signal processing chip with a floating-point arithmetic unit and a data buffer unit, and is configured with a hardware timer. The data buffer unit is used to store the channel fading sequence and frequency shift compensation factor according to a unified timestamp, and the hardware timer is used to generate a spectrum sensing start signal according to the dynamic sensing period.

[0090] Extract a channel fading sequence within a continuous time period from the data buffer unit, sort the amplitude values ​​in the channel fading sequence according to a unified timestamp, read the first and last unified timestamps of the channel fading sequence, and determine the time difference between the two unified timestamps as the total duration of the sequence.

[0091] The airborne digital signal processor accumulates all amplitude values ​​in the channel fading sequence and divides the accumulated result by the number of amplitude values ​​contained in the channel fading sequence to obtain the mean amplitude of the channel fading sequence. The mean amplitude is calculated according to the following formula:

[0092] ;

[0093] in, This represents the mean amplitude of the channel fading sequence. Represents the first in the channel fading sequence Amplitude value, This indicates the number of amplitude values ​​contained in the channel fading sequence. This indicates the index of the amplitude value in the channel fading sequence.

[0094] After obtaining the mean amplitude, the airborne digital signal processor compares the magnitude of two adjacent amplitude values ​​with the mean amplitude in chronological order. When an amplitude value is less than the mean amplitude and its next adjacent amplitude value is greater than the mean amplitude, an intersection point is recorded. When an amplitude value is greater than the mean amplitude and its next adjacent amplitude value is less than the mean amplitude, an intersection point is also recorded. After completing the comparison of all adjacent amplitude values, the number of intersection points where the channel fading sequence crosses the mean amplitude is obtained.

[0095] When multiple consecutive amplitude values ​​in a channel fading sequence are equal to the mean amplitude, these multiple consecutive amplitude values ​​are treated as an equal value interval. The amplitude values ​​before and after the equal value interval are compared. When the two amplitude values ​​are on different sides of the mean amplitude, the equal value interval is recorded as an intersection point. When the two amplitude values ​​are on the same side of the mean amplitude, the equal value interval is not recorded as an intersection point, thereby avoiding repeated statistics for the same traversal.

[0096] Dividing the number of intersections by the total duration of the sequence yields the fluctuation frequency of the channel fading sequence, which is calculated using the following formula: ;in, This indicates the frequency of fluctuations in the channel fading sequence. This represents the number of intersections of the channel fading sequence across the mean amplitude. The total duration of the sequence is represented by the fluctuation frequency, which represents the number of times the channel fading sequence crosses the average amplitude per unit time.

[0097] It should be noted that the fluctuation frequency is used to characterize the amplitude change rate of the channel fading sequence. The larger the fluctuation frequency, the more frequently the amplitude changes of the channel fading sequence occur per unit time. The smaller the fluctuation frequency, the slower the amplitude changes of the channel fading sequence. When the number of intersections is zero, the fluctuation frequency is set to a predetermined minimum non-zero fluctuation frequency to avoid performing the reciprocal operation on zero.

[0098] The reference duration is obtained by taking the reciprocal of the fluctuation frequency, and the reference duration is calculated according to the following formula: ;in, Indicates the reference duration. The frequency of fluctuation is indicated by the reference duration, which represents the average time scale between the events in which the average amplitude of two adjacent events crosses. The higher the frequency of fluctuation, the shorter the reference duration, and vice versa.

[0099] The reference duration is set as the sensing cutoff boundary. Starting from the first unified timestamp of the channel fading sequence, multiple sensing cutoff boundaries are determined sequentially according to the reference duration. The time axis interval between two adjacent sensing cutoff boundaries is set as a basic sensing window. When the duration of the last time axis interval is less than the reference duration, the last time axis interval is not set as a basic sensing window separately, so as to ensure that each basic sensing window has a consistent reference duration.

[0100] After setting the basic sensing window, frequency shift compensation factors with the same time range as the basic sensing window are extracted from the data cache unit. The absolute values ​​of each frequency shift compensation factor are taken, and the average of the absolute values ​​of all frequency shift compensation factors within the basic sensing window is calculated. The average result is calculated according to the following formula:

[0101] ;

[0102] in, This represents the average of the absolute values ​​of the frequency shift compensation factor within the basic sensing window. Indicates the first in the basic perception window One frequency shift compensation factor, This indicates the number of frequency shift compensation factors within the basic sensing window. This indicates the sequence number of the frequency shift compensation factor.

[0103] Based on the maximum frequency offset range that the RF receiving front end can compensate, a normalization upper limit for the frequency shift compensation factor is preset. The average result of the absolute value of the frequency shift compensation factor is divided by the normalization upper limit, and the result is reverse normalized to generate a weighting coefficient. The weighting coefficient is calculated according to the following formula.

[0104] ;

[0105] in, Indicates the weighting coefficient. This represents the average result of the absolute value of the frequency shift compensation factor. This represents the normalized upper limit of the frequency shift compensation factor. This indicates that the smaller value is selected from the values ​​being compared, and the weighting coefficient is neither less than zero nor greater than one.

[0106] It should be noted that the reverse normalization process is used so that the weight coefficient is smaller when the absolute value of the frequency shift compensation factor is larger, and the weight coefficient is larger when the absolute value of the frequency shift compensation factor is smaller. This results in a shorter dynamic sensing period when the frequency offset caused by the UAV's movement increases, thereby reducing the time deviation between the spectrum sensing result and the current spectrum state.

[0107] A pre-set period bias constant is read, and the period bias constant is added to a weighting coefficient to generate an adjustment parameter; in this embodiment, the period bias constant is set to 0.2. The adjustment parameter is multiplied by the reference duration of the basic sensing window to obtain the dynamic sensing period.

[0108] In this embodiment, when the total sequence duration is 1 second, the number of intersections is 8, the fluctuation frequency is 8 times per second, the reference duration is 0.125 seconds, the average result of the absolute value of the frequency shift compensation factor is 600 Hz, the normalization upper limit is 1000 Hz, the weighting coefficient is 0.4, the weighting coefficient is added to the period offset constant set to 0.2 to obtain an adjustment multiplier of 0.6, the adjustment multiplier is multiplied by the reference duration to obtain a dynamic sensing period of 0.075 seconds.

[0109] When the calculated dynamic sensing period is less than the shortest timing duration that the hardware timer can stably execute, the dynamic sensing period is set to the shortest timing duration. When the calculated dynamic sensing period is greater than the preset longest sensing duration, the dynamic sensing period is set to the longest sensing duration. This avoids the problem of insufficient baseband signal sampling due to an excessively short dynamic sensing period, or the problem of the spectrum sensing result not reflecting the current spectrum status in a timely manner due to an excessively long dynamic sensing period.

[0110] Please see Figure 5As shown, in one embodiment of the present invention, in step S4, the process of calculating the dynamic sensing threshold for adapting to a high-dynamic channel by combining the fading depth of the channel fading sequence and the current ambient noise level is as follows:

[0111] In this embodiment, an airborne digital signal processor performs calculations on the channel fading sequence and spatial noise floor samples. The airborne digital signal processor uses a data processing chip with a floating-point arithmetic unit and is configured with a high-speed data memory and a threshold register. The high-speed data memory is used to store the channel fading sequence and spatial noise floor samples according to a uniform timestamp, and the threshold register is used to store the calculated dynamic sensing threshold.

[0112] The airborne digital signal processor extracts the channel fading sequence within a dynamic sensing period from the high-speed data memory, sorts the amplitude values ​​in the channel fading sequence according to a unified timestamp, and compares the magnitude relationship between the current amplitude value and its previous and next amplitude values, starting from the first amplitude value.

[0113] When the current amplitude value is greater than the previous amplitude value and greater than the next amplitude value, the current amplitude value is determined as a local envelope peak. When the current amplitude value is less than the previous amplitude value and less than the next amplitude value, the current amplitude value is determined as a local envelope trough. When the current amplitude value is equal to the adjacent amplitude value, the amplitude values ​​on both sides of the equal amplitude value are compared, and the presence of a local envelope peak or a local envelope trough is determined based on the direction of change of the amplitude values ​​on both sides.

[0114] Local envelope peaks and valleys are paired according to the order of unified timestamps. A local envelope peak is paired with the first local envelope valley that is after the local envelope peak and before the next local envelope peak. When there are multiple local envelope valleys between two local envelope peaks, the local envelope valley with the smallest amplitude value is selected for pairing. When there are unpaired local envelope peaks at the end of the channel fading sequence, the unpaired local envelope peaks are discarded.

[0115] For each peak-valley pair, the local envelope peak value is subtracted from the local envelope valley value to obtain the corresponding amplitude difference component. The amplitude difference component is a non-negative value. After calculating the amplitude difference components of all peak-valley pairs, all amplitude difference components are accumulated, and the accumulated result is divided by the number of amplitude difference components to obtain the fading depth of the channel fading sequence.

[0116] When there is only one effective peak-valley pair in the channel fading sequence, the amplitude difference component of the effective peak-valley pair is used as the fading depth. When there is no effective peak-valley pair in the channel fading sequence, the fading depth is set to zero, and the local envelope peak value and local envelope valley value are re-extracted in the next dynamic sensing period. This avoids the situation where the averaging operation cannot be performed due to the number of peak-valley pairs being zero.

[0117] It should be noted that both the local envelope peak and local envelope valley values ​​are extracted from the channel fading sequence after low-pass filtering and smoothing. Low-pass filtering and smoothing can suppress pseudo-local envelope peaks and pseudo-local envelope valleys generated by single abnormal amplitude values. The greater the fading depth, the more obvious the envelope fluctuations in the channel fading sequence; the smaller the fading depth, the smoother the envelope changes in the channel fading sequence.

[0118] When calculating the current ambient noise level, the radio frequency (RF) transmitter unit of the UAV is controlled to stop transmitting signals, so that the UAV enters a passive radiation state. At the same time, the RF receiver front-end, receiving antenna, low-noise amplifier, downconverter and analog-to-digital converter are kept in receiving state, and the RF receiver front-end adopts the same center frequency, receiving bandwidth, receiving gain and discrete sampling frequency as the channel fading sequence acquisition process.

[0119] It should be noted that passive radiation state means that the UAV's radio frequency transmitting unit stops transmitting signals into space, but does not mean that the radio frequency receiving front end stops working. Before extracting the space noise floor sample, the airborne digital signal processor continuously checks the data output by the radio frequency receiving front end. When a persistent effective carrier component or a sudden interference component is detected, the current data is not used as the space noise floor sample. When no effective carrier component or sudden interference component is detected, the collection of the space noise floor sample begins.

[0120] Four thousand and ninety-six spatial noise floor samples are continuously collected using the analog-to-digital converter at the radio frequency receiving front end. The spatial noise floor samples are subjected to the same amplitude extraction processing as the channel fading sequence. The processed spatial noise floor samples are arranged in the order of sampling time to construct a noise sequence. When the spatial noise floor sample contains saturated data of the analog-to-digital converter or missing data, the current spatial noise floor sample is discarded and a new sample is collected.

[0121] The airborne digital signal processor first calculates the average value of all values ​​in the noise sequence, then calculates the difference between each value in the noise sequence and the average value, squares each difference, sums all the squared results, and calculates the variance of the noise sequence based on the number of values ​​it contains. The variance is calculated according to the following formula:

[0122] ;

[0123] in, This represents the variance of the noise sequence and serves as the amplitude of the noise floor fluctuation. Represents the first in the noise sequence A number, This represents the average value of all values ​​in the noise sequence. This indicates the number of numerical values ​​contained in the noise sequence. In this embodiment, the numerical index represents the sequence number within the noise sequence. Corresponding to 4,096 spatial noise samples.

[0124] The variance of the noise sequence is determined as the noise floor fluctuation amplitude, and the noise floor fluctuation amplitude is set as the current environmental noise level. The noise floor fluctuation amplitude reflects the degree to which the spatial noise floor sample changes around its average value. The larger the noise floor fluctuation amplitude, the more obvious the change in the current spatial noise floor. The smaller the noise floor fluctuation amplitude, the more stable the current spatial noise floor.

[0125] The preset margin adjustment coefficient is read from the configuration data of the airborne digital signal processor. The fading depth is multiplied by the margin adjustment coefficient to generate a dynamic bias term. The margin adjustment coefficient is determined by pre-calibration, and the dynamic bias term adopts the same linear numerical scale as the current ambient noise level.

[0126] It should be noted that the fading depth of the channel fading sequence comes from the envelope amplitude difference, and the current ambient noise level comes from the variance of the noise sequence. The two cannot be directly added without scaling. Therefore, the margin adjustment coefficient is used not only to adjust the influence of the fading depth on the dynamic sensing threshold, but also to convert the fading depth to the same linear numerical scale as the current ambient noise level, thereby ensuring that the dynamic bias term and the current ambient noise level can be summed.

[0127] The current ambient noise level and the dynamic bias term are summed in the linear domain to obtain the basic decision threshold. The linear domain refers to the summation of values ​​directly after uniform quantization and scaling transformation, without performing logarithmic transformation on the current ambient noise level and the dynamic bias term. When the stored data is represented by logarithmic values, it is first restored to linear values ​​before the summation operation is performed.

[0128] The airborne digital signal processor uses the basic decision threshold as the dynamic sensing threshold to adapt to the high dynamic channel, and writes the dynamic sensing threshold into the threshold register for reading when determining the spectrum occupancy status. When the fading depth of the channel fading sequence increases, the dynamic bias term increases accordingly, and the dynamic sensing threshold increases accordingly. When the current ambient noise level increases, the dynamic sensing threshold increases synchronously, thereby enabling the dynamic sensing threshold to be adjusted according to the channel fading status and the spatial noise floor status.

[0129] Please see Figure 6 As shown, in one embodiment of the present invention, in step S5, the process of using a frequency shift compensation factor to correct the frequency offset of the baseband signal during the dynamic sensing period, calculating the energy characteristics of the corrected signal, comparing it with the dynamic sensing threshold, and determining the current spectrum occupancy status is as follows:

[0130] In this embodiment, the field-programmable gate array (FPGA) and the airborne digital signal processor (DSP) jointly complete the time-domain truncation of the current sensing segment, the construction of the complex exponential compensation sequence, the frequency offset correction, the energy feature calculation, and the determination of the spectrum occupancy status. The FPGA is equipped with a hardware timer, a high-speed data buffer, and a complex multiplier, while the DSP is equipped with a digital channel selection filter, an energy calculation unit, and a numerical comparator.

[0131] The dynamic sensing period is read from the storage location of the dynamic sensing period, written into the hardware timer of the field programmable gate array, and converted into the corresponding number of samples according to the discrete sampling frequency of the baseband signal. When the conversion result is not an integer, the integer closest to the conversion result is selected as the actual number of samples, so that the time error between the actual time cutoff window and the dynamic sensing period does not exceed one sampling interval.

[0132] When the hardware timer reaches the spectrum sensing start time, it records the unified timestamp of the current sampling point of the baseband signal and uses the unified timestamp as the start time of the time truncation window. It continuously reads the same number of baseband signal sampling points as the actual sampling number from the high-speed data buffer to form the current sensing segment. When the number of baseband signal sampling points in the high-speed data buffer is insufficient, it continues to receive baseband signals until the number of baseband signal sampling points meets the requirements of the time truncation window.

[0133] A time series is constructed based on the unified timestamps of each baseband signal sampling point in the current sensing segment. The first unified timestamp of the current sensing segment is set as the time starting point, and the time intervals between the remaining unified timestamps and the time starting point are calculated respectively, so as to obtain a time series that corresponds one-to-one with the baseband signal sampling points in the current sensing segment.

[0134] Extract frequency shift compensation factors with the same time range as the current sensing segment from the storage location of the frequency shift compensation factors. When the update frequency of the frequency shift compensation factors is lower than the discrete sampling frequency of the baseband signal, perform time interpolation on adjacent frequency shift compensation factors according to a unified timestamp, so that each baseband signal sampling point has a frequency shift compensation factor corresponding to its unified timestamp.

[0135] The cumulative compensation phase corresponding to each baseband signal sampling point is calculated based on the time series and frequency shift compensation factor, and a complex exponential compensation sequence is constructed using the cumulative compensation phase. The complex exponential compensation sequence is constructed according to the following formula.

[0136] ;

[0137] in, Represents the first in the complex exponentially compensated sequence A complex value, This represents the sampling point number of the baseband signal in the current sensing segment, where i represents the imaginary unit. Represents pi (π). Indicates the first Frequency shift compensation factor corresponding to each baseband signal sampling point Indicates the first The time interval between each baseband signal sampling point and its previous baseband signal sampling point This indicates the sampling point number during the cumulative compensation phase calculation process.

[0138] It should be noted that by constructing a complex exponential compensation sequence using cumulative compensation phase, each frequency shift compensation factor can generate a phase increment only within the corresponding time interval. When the motion state of the UAV changes and the frequency shift compensation factor changes with time, the complex exponential compensation sequence can continuously follow the change of the frequency shift compensation factor, avoiding the discontinuity of compensation phase caused by directly multiplying a single frequency shift compensation factor with the entire time series.

[0139] The current sensing segment and the complex exponential compensation sequence are input into the complex multiplier of the field programmable gate array. According to the same sampling point number, each complex sample value in the current sensing segment is multiplied with the corresponding complex value in the complex exponential compensation sequence to obtain all complex sample values ​​that have completed frequency offset correction. The complex sample values ​​that have completed frequency offset correction are arranged according to a unified timestamp and the corrected signal is output.

[0140] The modulus of each complex value in the complex exponential compensation sequence is one. When the complex multiplier performs element-wise multiplication, the complex exponential compensation sequence changes the phase change rate and frequency position of the current sensing segment without changing the amplitude of a single baseband signal sampling point in the current sensing segment. Thus, frequency offset correction is completed while preserving the amplitude information of the current sensing segment.

[0141] It should be noted that the complex exponential compensation sequence itself does not change the total amplitude energy of the current sensing segment. In order for frequency offset correction to have a practical effect on the extraction of energy features, before calculating the energy features, a digital channel selection filter is used to limit the corrected signal to the sensing bandwidth corresponding to the current spectrum. Frequency offset correction allows the signal components that have deviated from the sensing bandwidth due to the movement of the UAV to re-enter the sensing bandwidth. The digital channel selection filter also suppresses noise and interference outside the sensing bandwidth.

[0142] The corrected signal, processed by the digital channel selection filter, is input into the energy calculation unit. The in-phase and quadrature components of each complex sample value in the corrected signal are extracted respectively. The in-phase and quadrature components are squared respectively. The results of the two square operations are added together to obtain the instantaneous energy of each complex sample value.

[0143] The instantaneous energy of all complex sample values ​​within the current sensing segment is accumulated, and the accumulated result is divided by the number of effective samples contained in the current sensing segment to obtain the energy characteristics of the corrected signal. The average energy corresponding to a unit number of effective samples is used as the energy characteristic, which can avoid the linear change of the energy characteristic due to the increase or decrease of the number of samples when the number of samples changes due to the change of the dynamic sensing period.

[0144] Extract the dynamic sensing threshold and confirm that the dynamic sensing threshold and energy feature use the same sensing bandwidth, receiving gain and linear numerical scale. Input the energy feature and dynamic sensing threshold into a numerical comparator. Use the numerical comparator to subtract the value of the dynamic sensing threshold from the value of the energy feature to obtain the state deviation.

[0145] The sign characteristic of the state deviation is analyzed. When the state deviation is greater than zero, the sign characteristic is determined to be positive, and the current spectrum occupancy status is determined to be occupied. When the state deviation is less than zero, the sign characteristic is determined to be negative, and the current spectrum occupancy status is determined to be idle. When the state deviation is equal to zero, the current spectrum occupancy status is determined to be idle according to the pre-set boundary determination rules.

[0146] In this embodiment, the discrete sampling frequency of the baseband signal is set to one megahertz. When the dynamic sensing period is 0.075 seconds, the current sensing segment contains 75,000 baseband signal sampling points. When the energy characteristic is 0.13 and the dynamic sensing threshold is 0.10, the state deviation is positive, and the current spectrum occupancy status is determined to be occupied. When the energy characteristic is 0.08 and the dynamic sensing threshold is 0.10, the state deviation is negative, and the current spectrum occupancy status is determined to be idle.

[0147] In one embodiment of the present invention, in step S6, the process of obtaining a pre-configured nominal center frequency and tuning the UAV radio frequency unit to the nominal center frequency to establish a link and complete spectrum access when the spectrum occupancy status is determined to be idle is as follows:

[0148] In this embodiment, the UAV radio frequency unit adopts a zero intermediate frequency direct conversion structure, including a frequency synthesizer, a voltage-controlled oscillator and a radio frequency front end. The frequency synthesizer includes a temperature-compensated crystal oscillator, a phase detector, a loop filter and a frequency divider.

[0149] The system obtains the determination result that the spectrum occupancy status is idle, checks whether the determination result is within the effective time range corresponding to the dynamic sensing cycle, and if it has exceeded the effective time range, it re-obtains the spectrum occupancy status. If it is still within the effective time range, it retrieves the nominal center frequency of the current sensing band pre-configured from the non-volatile memory.

[0150] The nominal center frequency is input to the frequency synthesizer. The frequency synthesizer configures the frequency divider according to the nominal center frequency and compares the feedback signal with the reference clock provided by the temperature-compensated crystal oscillator. Based on the comparison result, a control voltage is generated. The control voltage is used to drive the voltage-controlled oscillator to generate a local oscillator carrier with the same nominal center frequency.

[0151] It should be noted that this embodiment adopts a zero intermediate frequency direct conversion structure. Therefore, the frequency of the local oscillator carrier is the same as the nominal center frequency. When the frequency synthesizer fails to complete frequency locking within the preset locking time, spectrum access is stopped. When frequency locking is completed, the local oscillator carrier is input to the radio frequency front end to complete tuning.

[0152] The radio frequency front end uses the local oscillator carrier to convert the handshake data packet to the nominal center frequency, and transmits the handshake data packet to the communication peer through the power amplifier and receiving antenna. After the transmission is completed, it switches to the receiving state and waits for the communication peer to return the handshake response data packet within a preset response time.

[0153] When the handshake response data packet passes the verification, the link is established and spectrum access is completed. When no handshake response data packet is received or the handshake response data packet fails the verification, the transmission of signals in the current sensing frequency band is stopped and the spectrum occupancy status is reacquired to avoid mistakenly identifying the completion of handshake data packet transmission as a successful link establishment.

[0154] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A method for accessing the spectrum of UAV signals, characterized in that, Includes the following steps: S1: Collect the three-dimensional velocity and acceleration of the UAV, synchronously receive the baseband signal, and construct a multi-dimensional parameter set to characterize the high dynamic channel state; S2: Based on a multi-dimensional parameter set, the frequency shift compensation factor is calculated using three-dimensional velocity and acceleration, and envelope detection is performed on the baseband signal to extract the channel fading sequence; S3: Use the fluctuation frequency of the channel fading sequence to set the basic sensing window, use the frequency shift compensation factor as a weight to weight the basic sensing window, and output the dynamic sensing period. S4: Calculate the dynamic sensing threshold for adapting to high dynamic channels by combining the fading depth of the channel fading sequence and the current environmental noise level. S5: During the dynamic sensing period, the baseband signal is frequency offset corrected using the frequency shift compensation factor, and the energy characteristics of the corrected signal are calculated. The corrected signal is then compared with the dynamic sensing threshold to determine the current spectrum occupancy status. S6: When the spectrum occupancy status is determined to be idle, obtain the pre-configured nominal center frequency, tune the UAV radio frequency unit to the nominal center frequency to establish a link, and complete the spectrum access.

2. The method for unmanned aerial vehicle (UAV) signal spectrum sensing and access according to claim 1, characterized in that, In step S1, the process of constructing a multidimensional parameter set to characterize the high-dynamic channel state is as follows: The airborne inertial measurement unit and satellite positioning equipment are activated to obtain three-dimensional velocity and acceleration, and the radio frequency receiving front-end is scheduled in parallel to perform down-conversion and discrete sampling of space electromagnetic waves to generate baseband signals; The timing pulses output by satellite positioning equipment are extracted to construct a time reference, and a unified timestamp is assigned to the three-dimensional velocity, acceleration and baseband signals. Based on the time series alignment and fusion, a multi-dimensional parameter set is constructed to characterize the high dynamic channel state.

3. The method for unmanned aerial vehicle (UAV) signal spectrum sensing and access according to claim 1, characterized in that, In step S2, the process of calculating the frequency shift compensation factor based on the multidimensional parameter set using three-dimensional velocity and acceleration, and performing envelope detection on the baseband signal to extract the channel fading sequence is as follows: The three-dimensional velocity and acceleration of the multi-dimensional parameter set are analyzed, and the three-dimensional velocity and acceleration are superimposed by directional projection along the signal transmission path to calculate the instantaneous radial velocity of the UAV. Divide the instantaneous radial velocity by the speed of light to obtain the change ratio, multiply the change ratio by the center carrier frequency to calculate the Doppler frequency shift, and invert the Doppler frequency shift to construct the frequency shift compensation factor. The baseband signal within the multidimensional parameter set is extracted, and the square root of the sum of the squares of the in-phase and quadrature components of the baseband signal is calculated to obtain the instantaneous amplitude waveform. The instantaneous amplitude waveform is then subjected to low-pass filtering and smoothing to extract the channel fading sequence.

4. The UAV signal spectrum sensing access method according to claim 1, characterized in that, In step S3, the process of setting a basic sensing window using the fluctuation frequency of the channel fading sequence, weighting the basic sensing window with the frequency shift compensation factor as a weight, and outputting the dynamic sensing period is as follows: Calculate the mean amplitude of the channel fading sequence, count the number of intersections of the channel fading sequence across the mean amplitude, and divide the number of intersections by the total duration of the sequence to obtain the fluctuation frequency. The reference duration is obtained by performing a reciprocal operation on the fluctuation frequency. The reference duration is set as the perception cutoff boundary. The basic perception window is set by dividing the time axis interval according to the perception cutoff boundary. The absolute value of the frequency shift compensation factor is extracted and normalized to generate a weighting coefficient. The weighting coefficient is summed with the set period offset constant to generate an adjustment parameter. The adjustment parameter is multiplied by the reference duration of the basic sensing window and weighted to output the dynamic sensing period.

5. The method for unmanned aerial vehicle (UAV) signal spectrum sensing and access according to claim 1, characterized in that, In step S4, the process of calculating the dynamic sensing threshold for adapting to the high-dynamic channel by combining the fading depth of the channel fading sequence and the current environmental noise level is as follows: The local envelope peak and local envelope valley values ​​are extracted by traversing the channel fading sequence. The local envelope peak values ​​are subtracted from the local envelope valley values ​​to calculate each amplitude difference component. The mean of all amplitude difference components is calculated as the fading depth of the channel fading sequence. Extract spatial noise samples under passive radiation conditions to construct a noise sequence, perform variance calculation on the noise sequence to obtain the noise floor fluctuation amplitude, and set the noise floor fluctuation amplitude as the current environmental noise level; The fading depth is multiplied by a preset margin adjustment coefficient to generate a dynamic bias term. The current ambient noise level is summed with the dynamic bias term in the linear domain to obtain the basic decision threshold. The basic decision threshold is output as a dynamic sensing threshold adapted to the high dynamic channel.

6. The method for unmanned aerial vehicle (UAV) signal spectrum sensing and access according to claim 1, characterized in that, In step S5, the process of correcting the frequency offset of the baseband signal using a frequency shift compensation factor and calculating the energy characteristics of the corrected signal within the dynamic sensing period is as follows: A time truncation window is set with the dynamic sensing period, and a time domain truncation operation is performed on the baseband signal to extract the current sensing segment. The time series of the current sensing segment is extracted, and a complex exponential compensation sequence is constructed by combining the time series with the frequency shift compensation factor. The complex exponential compensation sequence is multiplied element-wise with the current sensing segment to complete the frequency offset correction of the current sensing segment and output the corrected signal. The discrete amplitude sequence of the corrected signal is analyzed, and the discrete amplitude sequence is squared to generate instantaneous power data. The instantaneous power data is accumulated and summed to output the energy characteristics.

7. The UAV signal spectrum sensing access method according to claim 6, characterized in that, The specific content of comparing it with the dynamic sensing threshold to determine the current spectrum occupancy status is as follows: Extract the energy characteristics and dynamic sensing threshold of the corrected signal, subtract the value of the dynamic sensing threshold from the value of the energy characteristics, and perform a difference operation to generate the state deviation. The sign characteristics of the analytical state deviation are analyzed. If the sign characteristic is positive, the current spectrum occupancy status is determined to be occupied; if the sign characteristic is negative or zero, the current spectrum occupancy status is determined to be idle.

8. The method for unmanned aerial vehicle (UAV) signal spectrum sensing and access according to claim 1, characterized in that, In step S6, the process of obtaining the pre-configured nominal center frequency and tuning the UAV radio frequency unit to the nominal center frequency to establish a link when the spectrum occupancy status is determined to be idle, thus completing the spectrum access, is as follows: Obtain the determination result that the spectrum occupancy status is idle, and extract the nominal center frequency pre-configured for the current sensing frequency band; The nominal center frequency is input into the frequency synthesizer of the UAV's radio frequency unit, which drives the voltage-controlled oscillator to generate a local oscillator carrier with the same nominal center frequency to complete the tuning. The radio frequency front end is then controlled to transmit handshake data packets based on the local oscillator carrier to establish a link and complete spectrum access.

Images