Dynamic spectrum frequency hopping communication method and device between target drone and ground station

Abstract

This invention relates to the field of unmanned aerial vehicle (UAV) communication technology, and discloses a dynamic spectrum frequency hopping communication method and apparatus between a target drone and a ground station. The dynamic spectrum frequency hopping communication method includes: collecting status information; evaluating signal quality indicators based on power data; dividing the total spectrum range into N sub-bands based on the total spectrum range and signal quality indicators to obtain spectrum allocation results; electing a master target drone based on task priority and communication quality indicators; each target drone sending an orthogonal probe sequence in its corresponding time slot to obtain channel status information; converting the channel status information into a frequency hopping key; and generating a frequency hopping sequence based on the frequency hopping key. This invention employs adaptive frequency hopping technology based on channel status, effectively avoiding interference bands by real-time monitoring of channel quality and generating dynamic frequency hopping sequences; and fundamentally avoiding signal overlap through a time-division multiplexing channel probe mechanism, ensuring continuous and stable operation in complex dynamic environments.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) communication technology, and more specifically, to a method and apparatus for dynamic spectrum frequency hopping communication between a target drone and a ground station. Background Technology

[0002] With the increasing demands of modern training activities and equipment testing, unmanned target drone systems have become important training equipment. In typical target drone training scenarios, multiple target drones are usually required to perform different flight missions simultaneously, including complex tasks such as single-target simulation, formation flight, and coordinated maneuvers. This requires the establishment of a stable and secure communication link between the target drones and the ground station.

[0003] Currently, target drone communication systems mainly employ fixed-frequency or simple frequency-hopping communication methods. Fixed-frequency communication is susceptible to interception and interference, while traditional frequency-hopping technology requires pre-setting frequency-hopping patterns, posing numerous technical challenges when multiple target drones are operating collaboratively.

[0004] A Chinese patent with authorization announcement number CN113765541B discloses a frequency hopping communication method, device, computer equipment, and storage medium for unmanned aerial vehicles (UAVs). It provides a frequency hopping communication scheme based on channel estimation technology to dynamically generate consensus frequency hopping sequences. That is, without relying on a dedicated frequency hopping system, frequency hopping sequences that do not need to be agreed upon in advance can be dynamically generated between ordinary ground stations (or UAVs). This can meet the dynamic networking requirements between different UAVs and ground stations. It enables the UAVs to communicate by continuously switching frequencies according to time slices based on the generated frequency hopping sequences. It also greatly increases the confidentiality of UAVs using frequency hopping communication based on this dynamic generation method of frequency hopping sequences, preventing eavesdroppers from stealing information.

[0005] However, existing technologies are only applicable to point-to-point communication scenarios. When faced with one-to-many communication between multiple target drones and ground stations, and multiple target drones simultaneously sending channel probe signals to the ground station, signal overlap and mutual interference will occur, affecting the accurate calculation of channel state information. Summary of the Invention

[0006] The purpose of this invention is to provide a dynamic spectrum frequency hopping communication method and apparatus between a target drone and a ground station in order to solve the above-mentioned problems.

[0007] This invention provides a dynamic spectrum frequency hopping communication method between a target drone and a ground station, comprising the following steps:

[0008] Collect status information, including task priority and total spectrum range;

[0009] Real-time monitoring of power data; evaluation of signal quality indicators based on power data; division of the total spectrum range into N sub-bands based on the total spectrum range and signal quality indicators to obtain spectrum allocation results;

[0010] Calculate communication quality indicators based on power data, elect the master target machine based on task priority and communication quality indicators, and establish a timing benchmark.

[0011] Based on the timing reference, the time axis is divided into N+1 non-overlapping time slots. Each target device sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information.

[0012] Channel state information is converted into a frequency hopping key; a frequency hopping sequence is generated based on the frequency hopping key.

[0013] Furthermore, the frequency hopping sequence includes a master frequency hopping sequence and a slave frequency hopping sequence from the target device. The steps for generating the frequency hopping sequence include:

[0014] The main target device uses the generated main frequency hopping key to perform a modulo operation on the size of the first frequency point set, and adds one to the remainder as the index value. It selects the frequency point at the corresponding position from the first frequency point set as the frequency hopping frequency at the current moment. The frequency hopping frequencies at multiple moments form the main frequency hopping sequence. The first frequency point set is the set of frequency points that satisfy the minimum interval constraint and the signal quality index is greater than the preset red threshold.

[0015] The target device generates a slave frequency hopping key based on the received channel state information; processes the slave frequency hopping key using a SHA256 hash function; receives the master frequency hopping sequence broadcast by the master target device; and applies an orthogonal mapping function to generate a slave frequency hopping sequence by combining the hash-processed slave frequency hopping key and the master frequency hopping sequence.

[0016] Furthermore, the frequency hopping key includes a primary frequency hopping key and a secondary frequency hopping key. The frequency hopping key is generated using a multi-channel feature fusion method to extract the key information of the path with the largest complex gain coefficient magnitude in the channel state information. The key information includes the magnitude of the complex gain coefficient, the argument of the complex gain coefficient, and the time rate of change of the channel state. The key information is quantized, and the quantization results are concatenated to obtain the basic key. The basic key is then shifted and XORed to obtain the hybrid key. The hybrid key is then subjected to multiple shift and concatenation operations to obtain the frequency hopping key.

[0017] The magnitude of the complex gain coefficient is the sum of the squares of the real and imaginary parts of the complex gain coefficient, and then the square root of the sum is taken. The argument of the complex gain coefficient is the arctangent function of the imaginary part divided by the real part of the complex gain coefficient. The time change rate of the channel state is obtained by the ratio of the magnitude of the change in the complex gain coefficient at adjacent time points to the time interval. The change in the complex gain coefficient at adjacent time points is obtained by subtracting the complex gain coefficient at the previous time point from the complex gain coefficient at the current time point.

[0018] Furthermore, the orthogonal mapping function converts the hashed frequency hopping key, the current frequency hopping frequency in the master frequency modulation sequence, the slave target number, and the current timestamp into 32 bits and then performs an XOR operation to obtain the composite key value. The composite key value is then moduloed with the number of first frequency points in the slave target, and the remainder is incremented by one to serve as the base frequency point index. The frequency point at the corresponding position is selected from the set of first frequency points of the corresponding slave target using the base frequency point index as the frequency hopping frequency at the current moment.

[0019] Furthermore, the specific steps for dividing the time into N+1 non-overlapping time slots include:

[0020] All target drones and ground stations are synchronized based on a timing reference.

[0021] Divide the continuous time axis into periodic time frames;

[0022] Each time frame is divided into N+1 time slots of equal length. The first N time slots are allocated to N target drones, and the N+1 time slot is allocated to the ground station.

[0023] Furthermore, the detection signal is the signal after the orthogonal detection sequence has been propagated through the wireless channel. The orthogonal detection sequence is a signal sequence with mutually orthogonal characteristics. The channel state information includes the channel impulse response, the number of multipath taps, and the multipath propagation parameters. The channel impulse response consists of the complex gain coefficients of multiple multipath components. The multipath propagation parameters include the time delay, magnitude attenuation, and phase offset of each path.

[0024] The detection signal is correlated with a known orthogonal detection sequence. The correlation operation is performed by integrating the detection signal and the conjugate of the orthogonal detection sequence at different time delay positions. The integration result is the channel impulse response corresponding to the time delay position.

[0025] The number of multipath taps is determined by a power threshold detection method. Multipath components that exceed a preset threshold are judged as valid taps, and the number of valid taps is the number of multipath taps.

[0026] The path delay is determined by the position of the channel impulse response in the time domain. The delay is equal to the time position of the peak occurrence minus the arrival time of the direct path. The magnitude attenuation is calculated from the magnitude of the peak. The phase offset is calculated from the phase of the peak.

[0027] Furthermore, a weighted summation is used to calculate the comprehensive score of each target drone, and the target drone with the highest comprehensive score is selected as the main target drone. The comprehensive score is the sum of the task priority item and the communication quality index item, where the task priority item is the normalized task priority multiplied by the preset task priority weight, and the communication quality index item is the normalized communication quality index multiplied by the preset communication quality weight. The communication quality is the ratio of signal-to-noise ratio to bit error rate, where the bit error rate is the distance factor multiplied by the bit error rate, the signal-to-noise ratio is the ratio of useful signal power to noise power, the distance factor is the propagation loss caused by the distance from the target drone to the ground station, and the bit error rate is the ratio of the number of received erroneous bits to the total number of transmitted bits.

[0028] Furthermore, the steps for establishing a timing reference include:

[0029] Obtain GPS or BeiDou time from ground stations;

[0030] Multiple standard frequency components of different frequencies are generated based on GPS or BeiDou time, including 10MHz, 1MHz, and 100kHz.

[0031] The standard frequency components are superimposed and synthesized in the frequency domain to generate a timing reference signal;

[0032] The timing reference signal is broadcast to all target devices.

[0033] Furthermore, the power data includes signal power, noise power, and interference power. The signal quality index is the ratio of signal power to the total noise. The total noise is the sum of noise power and interference power. The spectrum allocation results include the sub-band allocated to each target drone and the set of first frequency points within the corresponding sub-band.

[0034] This invention provides a dynamic spectrum frequency hopping communication device for a target drone and a ground station, which stores computer instructions and executes the aforementioned dynamic spectrum frequency hopping communication method for the target drone and the ground station when the computer instructions are read. The device includes:

[0035] The collection module collects status information, including task priority and total spectrum range.

[0036] The allocation module monitors power data in real time and evaluates signal quality indicators based on the power data; based on the total spectrum range and signal quality indicators, it divides the total spectrum range into N sub-bands to obtain the spectrum allocation result.

[0037] The timing module calculates communication quality indicators based on power data, elects the master target machine based on task priority and communication quality indicators, and establishes a timing benchmark.

[0038] The channel module divides the time axis into N+1 non-overlapping time slots based on the timing reference. Each target device sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information.

[0039] The frequency hopping module converts channel state information into a frequency hopping key; and generates a frequency hopping sequence based on the frequency hopping key.

[0040] The beneficial effects of this invention are as follows: This invention achieves flexible sharing and optimized configuration of spectrum resources through dynamic spectrum allocation and intelligent frequency hopping technology. It employs spectrum allocation optimization based on a genetic algorithm, dynamically adjusting frequency band allocation according to real-time spectrum occupancy and task priorities, thereby improving spectrum utilization. It utilizes adaptive frequency hopping technology based on channel state, effectively avoiding interfering frequency bands by monitoring channel quality in real time and generating dynamic frequency hopping sequences. The multi-frequency modulated composite timing reference signal has strong anti-interference characteristics; even if some frequency components are interfered with, other components can still maintain time synchronization.

[0041] By using a time-division multiplexing channel detection mechanism, the continuous time axis is divided into N+1 non-overlapping time slots, ensuring that each target device sends detection signals within a dedicated time slot. This fundamentally avoids signal overlap and guarantees the accuracy of channel state information calculation.

[0042] By using orthogonal frequency hopping sequence generation technology, the frequency points used by different target drones at any given time are ensured to be completely orthogonal. By combining the SHA256 hash function and the orthogonal mapping function, frequency conflicts between multiple target drones are fundamentally eliminated. This invention establishes a timing reference based on GPS or BeiDou time and uses an adaptive broadband phase-locked loop to achieve accurate synchronization under high dynamic conditions, thereby improving synchronization accuracy and ensuring coordinated operation of all parts of the system. Attached Figure Description

[0043] Figure 1 This is a flowchart illustrating the dynamic spectrum frequency hopping communication method between the target drone and the ground station according to the present invention.

[0044] Figure 2 This is an example diagram illustrating the acquisition of channel state information using the dynamic spectrum frequency hopping communication method between the target drone and the ground station according to the present invention.

[0045] Figure 3 This is an example of the frequency hopping sequence generation process of the dynamic spectrum frequency hopping communication method between the target drone and the ground station according to the present invention. Figure 1 ;

[0046] Figure 4 This is an example of the frequency hopping sequence generation process of the dynamic spectrum frequency hopping communication method between the target drone and the ground station according to the present invention. Figure 2 ;

[0047] Figure 5This is a module example diagram of the dynamic spectrum frequency hopping communication device between the target drone and the ground station of the present invention. Detailed Implementation

[0048] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0049] A dynamic spectrum frequency hopping communication method and apparatus between a target drone and a ground station, including the following embodiments:

[0050] Example 1:

[0051] Dynamic spectrum frequency hopping communication methods between target drones and ground stations, such as Figure 1 As shown, it includes the following steps:

[0052] Step 101: Establish basic data link connections between N target drones and the ground station in the training airspace using the standard data link protocol; collect status information, including target drone status information and spectrum environment information, to provide basic data for subsequent master-slave target drone election and spectrum allocation.

[0053] When the communication method is initiated, a basic communication connection needs to be established and the working environment fully perceived. First, within the training airspace, N target drones (N≥2) establish basic data link connections with the ground station. The number of target drones N is determined based on actual training needs, typically ranging from 2 to 20, to balance training effectiveness and implementation complexity. The training airspace refers to a pre-defined aerial training area, usually a rectangular or circular area with sides of 50-200 kilometers, ranging in altitude from ground level to 12,000 meters, designated by the training management department according to the type of training mission, safety requirements, and airspace control regulations.

[0054] The communication method employs a standard data link protocol to establish an initial connection, comprising three phases: device discovery, authentication, and connection verification. The advantage of using a standard data link protocol lies in its mature security mechanisms and high reliability, ensuring connection stability and security even in complex electromagnetic environments. The standard data link protocol includes a complete protocol stack encompassing the physical layer, data link layer, network layer, and application layer, supporting functions such as encryption authentication, error control, and traffic management. The ground station discovers available target devices via broadcast, equipping each target device with a unique identifier for authentication. After connection establishment, communication quality testing is performed. The ground station maintains a real-time updated target device registry, monitoring the connection status and communication quality of each target device.

[0055] After successfully establishing a basic communication connection, it is necessary to collect status information to lay the foundation for subsequent coordinated communication. Status information includes target device status information and spectrum environment information; the specific collection methods are as follows:

[0056] Methods for collecting target drone status information include:

[0057] Position coordinate information: The precise position coordinates of each target drone in three-dimensional space, including longitude, latitude, and altitude, are obtained in real time via GPS or BeiDou navigation systems. The necessity of choosing GPS or BeiDou navigation systems lies in their global coverage and high-precision positioning characteristics, which can provide an accurate position reference for subsequent Doppler effect calculations and channel predictions. Coordinate accuracy must reach the meter level, with an update frequency of 10-100 times per second. The airborne navigation equipment is equipped with differential GPS or BeiDou receivers, and positioning accuracy is improved through calibration using ground reference stations.

[0058] Motion state information: The real-time motion state of the target drone is calculated using an inertial navigation system, GPS, or BeiDou differential positioning, including velocity components and acceleration information in three directions. Acquiring motion state data is crucial for predicting future position and Doppler shift, as these parameters directly affect the changing trends of channel characteristics. The inertial navigation system, based on gyroscopes and accelerometers, is an autonomous navigation system that calculates position and attitude by measuring the angular velocity and acceleration of the vehicle. It has advantages such as independence from external signals and strong anti-interference capabilities, but suffers from error accumulation and requires periodic calibration using GPS or BeiDou signals. Motion state data is processed through multi-sensor fusion, combining gyroscope, accelerometer, and GPS or BeiDou data to provide accurate dynamic information.

[0059] Task Priority Information: Task priority values ​​are assigned based on the importance and urgency of the tasks performed by the target drones. The value range is typically 1-10, with higher values ​​indicating higher priority. Setting task priorities is fundamental to achieving differentiated quality of service, ensuring that important tasks receive better communication resources and higher service quality. Task importance refers to the degree of impact the task has on achieving the overall training objectives; for example, a simulation task targeting the primary target has a priority of 9-10, an auxiliary jamming task has a priority of 5-7, and a reconnaissance / surveillance task has a priority of 3-5. Urgency refers to the time sensitivity of the task execution; for example, a threat avoidance task requiring immediate response has a priority of 8-10, and a scheduled patrol task has a priority of 2-4. Task priorities are dynamically adjusted by the task control center according to the training plan, taking into account factors such as task type, execution difficulty, and time requirements.

[0060] Methods for collecting spectrum environment information include:

[0061] Total spectrum range information: The frequency range in which communication methods operate is determined through spectrum monitoring equipment, typically covering the L-band to Ku-band (1GHz-18GHz), and dynamically adjusted according to usage licenses and environmental conditions. Accurate determination of the spectrum range is a prerequisite for avoiding unauthorized occupation and interference, and also provides available resource boundaries for subsequent intelligent spectrum allocation. Specifically, the L-band refers to the 1GHz-2GHz frequency range, mainly used for long-distance communication and navigation; the S-band refers to the 2GHz-4GHz frequency range, mainly used for radar and satellite communication; the C-band refers to the 4GHz-8GHz frequency range, mainly used for satellite communication and microwave relay; the X-band refers to the 8GHz-12GHz frequency range, mainly used for radar and high-speed data transmission; and the Ku-band refers to the 12GHz-18GHz frequency range, mainly used for satellite communication and high-resolution radar. Usage licenses clearly specify the frequency range, transmit power limits, usage time, and geographical scope. Spectrum management devices monitor the spectrum usage license status in real time to ensure compliant use.

[0062] Frequency spacing information: The minimum spacing between adjacent first frequency points is set based on signal bandwidth and interference tolerance, typically 1.2-2 times the signal bandwidth, to ensure that adjacent signals do not interfere with each other. A reasonable frequency spacing setting can effectively prevent adjacent channel interference and improve spectrum utilization efficiency. Signal bandwidth refers to the frequency range occupied by the communication signal; interference tolerance is the ratio of adjacent channel interference power that the receiver can tolerate to the useful signal power, typically requiring adjacent channel interference to be 20-40 dB lower than the useful signal. Spacing settings consider factors such as transmitter frequency stability, receiver selectivity, and propagation environment.

[0063] Real-time spectrum occupancy information: Continuously monitor the occupancy status of each frequency point at the current moment, identify external interference sources and internal signal conflicts, and establish a dynamic spectrum occupancy database. The necessity of real-time monitoring lies in the dynamic changing characteristics of the spectrum environment; only by understanding the real-time status can correct spectrum allocation decisions be made. External interference sources include radiation from civilian communication equipment, industrial equipment, radar systems, and other communication systems, which are classified using signal feature identification algorithms. Internal signal conflicts refer to the overlapping use of frequencies by different users within the same communication system, which are identified through power detection and time-domain analysis. The monitoring device uses a broadband receiver and a spectrum analyzer, updating the spectrum status every 100 milliseconds.

[0064] After completing the collection of status information, the communication method obtains the basic data for intelligent spectrum management, providing a basis for decision-making in the next step of dynamic spectrum allocation.

[0065] Step 102: Monitor power data in real time and evaluate signal quality indicators based on power data; based on the total spectrum range and signal quality indicators, divide the total spectrum range into N sub-bands to obtain the spectrum allocation result.

[0066] Specific implementation methods include:

[0067] The ground station's broadband spectrum analysis device simultaneously monitors the spectrum status across three key dimensions to obtain power data. The first dimension is signal power monitoring, which uses a square-law detector to measure the useful signal power intensity at each frequency point in real time. The second dimension is noise power assessment, obtaining background noise levels through power statistics during idle periods. The third dimension is interference power identification, using signal feature analysis algorithms to distinguish and quantify the power of various interference signals. The square-law detector is a nonlinear device whose output voltage is proportional to the input signal power, commonly implemented using diodes or field-effect transistors. Idle periods refer to times when no useful signal is transmitted; pure noise power measurements are obtained through timed sampling with a sampling interval of 10 milliseconds and a sampling duration of 1 millisecond. The signal feature analysis algorithms include power spectral density analysis, modulation identification, and time-frequency analysis, capable of distinguishing different types of interference, such as continuous wave interference, impulse interference, and modulated signal interference. The spectrum status monitoring method represents the state of each frequency point at a specific moment as a three-dimensional data combination containing signal power, noise power, and interference power. This description comprehensively reflects the spectrum's usage status and communication quality potential, providing an accurate decision-making basis for subsequent intelligent allocation.

[0068] Signal quality indices are used to quantify the availability of each frequency point. The signal quality index is calculated by using signal power as the numerator and the sum of noise power and interference power as the denominator. The ratio reflects the signal quality of that frequency point. A higher signal quality index indicates better signal quality, suitable for important communications; a lower index indicates severe interference, and usage should be avoided or protective measures should be taken. Spectrum utilization is defined as the ratio of the actual used spectrum bandwidth to the total available spectrum bandwidth, used to measure the efficiency of spectrum resource utilization. Useful signal power refers to the expected received communication signal power, measured in decibels per milliwatt (dBm), with a typical range of -100 to -30 dBm. Noise power refers to the sum of thermal noise and equipment noise, calculated as ten times the common logarithm of the product of Boltzmann's constant, absolute temperature, and signal bandwidth. Absolute temperature is typically taken as 290 Kelvin, and signal bandwidth is measured in Hertz. Interference power refers to the unwanted signal power generated by external interference sources and internal signal conflicts, separated from the total received power using a signal identification algorithm.

[0069] To improve evaluation accuracy, the processing device employs a digital signal processor for real-time calculations, combining moving average and exponentially weighted average methods to smooth the measurement results. Frequency points are classified into three levels based on the signal quality index calculation results to obtain the frequency point quality assessment results. High-quality frequencies are marked in green, indicating excellent communication quality; average frequencies are marked in yellow, indicating usability but requiring attention; and low-quality frequencies are marked in red, indicating unsuitability for use. Moving average refers to the arithmetic mean of the most recent M sample values, where M is typically 10-50, effectively smoothing random fluctuations. Exponentially weighted average assigns decreasing weights to historical data; the calculation method is that the current average value equals the smoothing factor multiplied by the current measurement value plus or minus the smoothing factor multiplied by the previous average value, where the smoothing factor is typically 0.1-0.3. The thresholds for the three-level classification are set as follows: the signal quality index of green frequencies is greater than the green threshold (default value of 10 dB); the signal quality index of yellow frequencies is between the green and red thresholds; and the signal quality index of red frequencies is less than the red threshold (default value of 0 dB).

[0070] Based on signal quality metrics and real-time spectrum occupancy information, a genetic algorithm is employed to optimize frequency allocation, achieving intelligent allocation of spectrum resources. The genetic algorithm is a global optimization algorithm based on the principles of biological evolution. By simulating natural selection, crossover, and mutation processes, it searches for the optimal solution in a complex solution space. The advantage of using a genetic algorithm lies in its strong global search capability, its ability to handle multi-objective optimization problems, and its suitability for solving complex combinatorial optimization problems such as spectrum allocation.

[0071] The optimization objective function of the genetic algorithm comprehensively considers three key indicators: maximizing overall efficiency, optimizing load balancing, and minimizing interference level. Maximizing overall efficiency means maximizing the overall communication capacity and service quality of the system under given spectrum resource constraints; optimizing load balancing means ensuring the load distribution across frequency bands is as even as possible, avoiding overload in some bands while others are idle; minimizing interference level means reducing mutual interference and improving communication quality through reasonable frequency allocation. These three objectives are weighted and summed to form a comprehensive objective function, with the weighting coefficients adjusted according to actual application requirements.

[0072] Based on the optimization objective, the genetic algorithm divides the total spectrum range into N sub-bands, each of which is allocated to a corresponding target machine. The band segmentation employs an adaptive strategy, grouping consecutive frequencies with similar quality into the same band based on the quality assessment results and task priority information, ensuring relatively uniform quality within each band. Target machines for high-priority tasks are preferentially allocated to sub-bands with better quality, while target machines for low-priority tasks are allocated to the remaining available sub-bands.

[0073] Sub-band allocation is achieved through chromosome encoding using a genetic algorithm. Chromosome encoding employs a one-dimensional array structure, where each element represents the sub-band index assigned to the corresponding target machine. The fitness function comprehensively considers three factors: frequency quality, task priority matching degree, and interference level, calculating the fitness value of each chromosome through weighted summation. Genetic operations include three steps: selection, crossover, and mutation. The selection operation uses a roulette wheel selection method, choosing superior individuals based on their fitness values. The crossover operation uses single-point crossover to generate new sub-band allocation schemes. The mutation operation increases population diversity by randomly altering the sub-band allocations of some target machines.

[0074] The parameters of the genetic algorithm are set as follows: population size of 50-100 individuals, number of generations of evolution of 100-200, crossover probability of 0.7-0.9, and mutation probability of 0.01-0.1. The convergence criterion of the algorithm is that the optimal fitness value changes by less than 0.1% for 10 consecutive generations, or the maximum number of generations of evolution is reached.

[0075] Based on the sub-band allocation, a first set of frequency points within a dedicated sub-band is established for each target drone. The process of establishing the first set of frequency points includes: firstly, extracting all frequency points within the sub-band range allocated to the target drone; then, based on the frequency point quality classification results, removing inferior frequency points marked in red; and finally, ensuring that adjacent selected frequency points meet the minimum spacing constraint according to the frequency point spacing requirements, thus forming the final first set of frequency points for the target drone.

[0076] The minimum spacing constraint is determined based on the frequency spacing information collected in step 101. Specifically, the minimum frequency spacing between adjacent first frequency points is equal to 1.2-2 times the signal bandwidth. The signal bandwidth is determined according to the type of communication service; digital data transmission bandwidth is typically 10-100kHz, and high-definition video transmission bandwidth is 1-10MHz. For example, when the signal bandwidth is 50kHz, the minimum spacing constraint is 60-100kHz; when the signal bandwidth is 2MHz, the minimum spacing constraint is 2.4-4MHz. The purpose of this constraint is to prevent adjacent channel interference, ensure that signals on adjacent frequency points do not affect each other, and improve spectrum utilization efficiency and communication quality.

[0077] The specific implementation method of the minimum interval constraint is as follows: In the candidate frequency point set, all first frequency points are arranged in ascending order of frequency. Then, a greedy algorithm is used to select frequency points sequentially, ensuring that the frequency difference between each newly selected frequency point and all previously selected frequency points is greater than or equal to the minimum interval constraint value. The algorithm starts from the lowest frequency point. If the frequency difference between the current frequency point and the previous selected frequency point satisfies the minimum interval constraint, it is added to the final first frequency point set; otherwise, the frequency point is skipped and the next frequency point is checked. This method ensures that the finally selected frequency points are uniformly distributed in the frequency domain and do not interfere with each other.

[0078] The genetic algorithm outputs spectrum allocation results, including the dedicated sub-band range allocated to each target drone and the first set of frequencies within that sub-band. These spectrum allocation results provide a dedicated spectrum resource foundation for the subsequent establishment of the master-slave architecture and the generation of frequency hopping sequences.

[0079] After optimization, the ground station broadcasts the spectrum allocation results to all target drones via the control channel, including information such as the dedicated frequency band range, specific frequency point set, and usage time for each target drone. Upon receiving the allocation information, each target drone configures its local frequency synthesizer and filters, preparing to communicate on the designated spectrum resources. The spectrum allocation information is protected using digital signature technology to prevent malicious tampering and forgery.

[0080] To ensure the dynamic adaptability of spectrum allocation, a spectrum allocation update mechanism is established. The spectrum reallocation process is automatically triggered when there are significant changes in the spectrum environment, adjustments to target device task priorities, deterioration in communication quality, or the addition of a new target device. Significant changes in the spectrum environment are defined as changes in the total spectrum range exceeding 20%, changes in signal quality indicators exceeding 3 dB, or the addition of a new interference source leading to an increase in interference power exceeding 10 dB. Adjustments to target device task priorities are defined as changes in task priority values ​​exceeding two levels or changes in the number of high-priority tasks exceeding 30%. Deterioration in communication quality is defined as a bit error rate exceeding 10^-4, a decrease in signal-to-noise ratio exceeding 5 dB, or more than 3 communication interruptions per hour. The addition of a new target device refers to the integration of a new target device into the system, requiring the allocation of dedicated spectrum resources and a re-optimization of the overall spectrum allocation scheme. The update cycle is set to 1-5 minutes, balancing adaptability and stability requirements. During reallocation, current communication is maintained without interruption, and a gradual switching method is used to progressively migrate communication to the new spectrum resources.

[0081] Through the aforementioned intelligent spectrum management process, optimal allocation of spectrum resources was achieved, providing a high-quality dedicated spectrum resource foundation for subsequent master-slave architecture establishment and coordinated communication. The allocation of sub-bands and the determination of the first set of frequency points ensured the rationality and effectiveness of resource allocation, meeting the complex requirements of dynamic communication among multiple target devices.

[0082] Step 103: Calculate communication quality indicators based on power data, and elect the master target machine based on task priority and communication quality indicators; establish a timing benchmark.

[0083] Specific methods include:

[0084] A comprehensive scoring method was used to select the primary target machine, with the score based on two core elements: task priority and communication quality indicators. The calculation of the communication quality indicators comprehensively considered three key parameters: signal-to-noise ratio (SNR) reflecting the basic quality of the communication link, distance factor considering the impact of propagation loss, and bit error rate (BER) reflecting the actual communication performance. Specifically, the SNR was used as a positive factor in the quality indicator, the distance factor multiplied by the BER was used as a negative factor in the numerator, and the overall communication quality evaluation was obtained through ratio calculation in the denominator. The signal-to-noise ratio (SNR) is the ratio of useful signal power to noise power, measured in decibels (dB). A typical range is 10-40 dB. It is calculated as SNR equal to 10 times the signal power multiplied by the logarithm of the noise power. The distance factor is the free-space propagation loss caused by the distance between the target drone and the ground station, also measured in decibels. The loss is calculated as 32.45 plus the logarithm of 20 times the communication frequency plus the logarithm of 20 times the distance. Frequency is measured in MHz, distance in km, and the logarithm is base 10. The bit error rate (BER) is the ratio of the number of received erroneous bits to the total number of transmitted bits. A typical range is 10⁻⁶ to 10⁻³. It is detected through cyclic redundancy check (CRC) or forward error correction coding and obtained through observation and statistics during real-time communication.

[0085] The weighting of the election algorithm reflects the task-oriented principle. Task priority is typically weighted at 0.6, reflecting the importance of task execution; communication quality is weighted at 0.4, ensuring the selected primary target machine possesses good communication capabilities. A weighted summation is used to calculate the comprehensive score of each target machine, and the target machine with the highest comprehensive score is selected as the primary target machine. The advantage of using a comprehensive scoring method is that it balances task importance and communication capability, avoiding selection bias caused by a single indicator. The comprehensive score is calculated as follows: the comprehensive score equals 0.6 multiplied by the normalized task priority plus 0.4 multiplied by the normalized communication quality index. Normalization maps the original value to the range of 0 to 1, calculated by subtracting the minimum value from the original value and then dividing by the maximum value minus the minimum value.

[0086] The election employs a distributed voting mechanism, with each target machine voting independently based on its own observations, avoiding the impact of single points of failure. The ground station, acting as the central node for vote aggregation and result announcement, collects all voting information and calculates the final result. To prevent malicious interference, digital signature technology is used to verify the authenticity and integrity of the votes. The distributed voting mechanism means that each node independently votes, avoiding the single point of failure risk of centralized decision-making; the digital signature technology uses RSA or Elliptic Curve Digital Signature Algorithm (ECDSA) with a key length of 2048-4096 bits, ensuring the unforgeability and non-repudiation of voting information. The term of the master target machine is set at 5-10 minutes; if performance significantly degrades during its term, a re-election can be triggered early.

[0087] Once the primary target drone is selected, a timing reference is immediately established in collaboration with the ground station. The timing reference is a standardized time reference signal used to ensure strict time consistency among all target drones; specifically, it is a multi-frequency modulated composite signal. This timing reference signal has the following specific technical characteristics: signal frequency accuracy on the order of 10 to the power of -12; time synchronization accuracy controlled within ±100 microseconds; signal coverage of at least 100 kilometers; and interference immunity margin greater than 20 dB. Specifically, signal frequency accuracy refers to frequency stability, using a cesium or rubidium atomic clock as a reference, with short-term stability ranging from 10 to the power of -11 to 10 to the power of -12; time synchronization accuracy refers to the deviation of each node's clock from the standard time, achieved through bidirectional time transfer and propagation delay compensation; signal coverage refers to the effective propagation distance of the timing reference signal, affected by transmit power, antenna gain, and propagation environment; and interference immunity margin refers to the ability to operate normally under interference conditions, improved through spread spectrum technology and error correction coding.

[0088] The steps for obtaining the timing reference include:

[0089] The ground station obtains high-precision GPS or BeiDou time through a time-frequency receiver, which serves as the fundamental source of the entire timeline reference. It's important to note that GPS time and BeiDou time (BDT) have a fixed deviation; therefore, the system design should consistently use the same time standard. The corresponding time-frequency receiver is a specialized GPS or BeiDou receiving device.

[0090] The digital signal generator generates multiple standard frequency components of different frequencies based on the GPS or BeiDou time obtained in the first step, including international standard frequencies such as 10MHz, 1MHz, and 100kHz. The corresponding digital signal generator employs direct digital frequency synthesis technology.

[0091] The signal processing unit performs frequency domain superposition and synthesis of various standard frequency components using a digital signal processor. This involves digitally superimposing multiple standard frequency components, such as 10MHz, 1MHz, and 100kHz, to generate the final timing reference signal. The corresponding signal processing unit employs a high-performance DSP chip or FPGA, achieving a processing speed of hundreds of billions of floating-point operations per second.

[0092] The timing reference signal is broadcast to all target drones via a high-power transmitter, ensuring that every target drone within the coverage area can receive it.

[0093] The timing reference employs multi-frequency modulation composite signal technology, which simultaneously modulates multiple sinusoidal signals of different frequencies onto a single carrier to form a reference signal with high-precision time characteristics. The advantage of multi-frequency modulation technology lies in its strong anti-interference capability; even if some frequency components are interfered with, the accuracy of the time reference can still be guaranteed for other components. Multi-frequency modulation refers to simultaneously modulating multiple baseband signals of different frequencies onto a single carrier signal. Mathematically, the modulation signal equals the carrier magnitude multiplied by a cosine function. The parameters of the cosine function are twice pi multiplied by the carrier frequency multiplied by time, plus the sum of the modulation exponents of each modulation component multiplied by a second cosine function. The parameters of the second cosine function are twice pi multiplied by the frequency of each modulation component multiplied by time.

[0094] The timing reference signal is based on a unified time system (GPS time or BeiDou time), upon which multiple modulation frequency components are superimposed. Each component has a specific time offset, modulation frequency, and initial amplitude parameter. The time offset is set to microsecond precision, different international standard frequencies are selected for the modulation frequencies, and the initial amplitude is generated through a pseudo-random sequence to enhance the signal's anti-interference capability. Specifically, the timing reference signal can be expressed as: the reference time equals the unified time plus a weighted combination of each modulation component, where the weighting coefficients are determined through an optimization algorithm to ensure signal stability and anti-interference capabilities. The time offset, with a precision of 1 microsecond, is the time difference set to distinguish different modulation components; international standard frequencies include 1MHz, 5MHz, and 10MHz, as specified by the International Telecommunication Union; the pseudo-random sequence uses an m-sequence or a Gold sequence with a period length of [missing information]. , where n is the sequence order.

[0095] The digital signal generator employs a high-stability crystal oscillator and GPS or BeiDou discipline technology to ensure required frequency accuracy. Digital signal generation utilizes direct digital frequency synthesis technology, supporting arbitrary waveform generation and precise frequency control. Transmit power is adaptively adjusted based on propagation distance and environmental conditions to ensure all slave targets receive a signal strength sufficient for synchronization. Slave targets are those other than the master target. The high-stability crystal oscillator refers to a temperature-compensated crystal oscillator or a cryogenic crystal oscillator; GPS or BeiDou discipline technology utilizes satellite signals to correct the local oscillator frequency; and direct digital frequency synthesis technology directly generates the desired frequency signal digitally, achieving a frequency resolution at the microhertz level.

[0096] The digital signal generation device comprises five main components: a GPS or BeiDou receiver module, a high-stability reference oscillator, a digital signal processor, a power amplifier, and an antenna system. The GPS or BeiDou receiver module receives satellite signals and provides a standard time reference; the high-stability reference oscillator provides a local clock maintenance function when GPS or BeiDou signals are lost; the digital signal processor performs multi-frequency modulation signal synthesis operations; the power amplifier ensures sufficient transmission power; and the antenna system is responsible for signal transmission and coverage. Specifically, the GPS or BeiDou receiver module includes an RF front-end, a baseband processor, and a clock output interface, capable of simultaneously tracking 8-12 satellites; the high-stability reference oscillator employs dual-furnace temperature control; the digital signal processor uses a multi-core ARM or DSP architecture, supporting parallel processing; the power amplifier is a linear amplifier; and the antenna system uses an omnidirectional or directional antenna array.

[0097] The transmitted timing reference signal employs spread spectrum technology to enhance anti-interference capabilities and includes a periodic synchronization identifier code to facilitate rapid synchronization and amplitude locking at the receiver. The signal transmission power is adaptively adjusted based on propagation distance and environmental conditions to ensure that all target devices receive a signal strength sufficient for synchronization. Signal propagation utilizes line-of-sight propagation, taking into account the effects of Earth's curvature and atmospheric refraction on the propagation path. Spread spectrum techniques include direct sequence spread spectrum and frequency hopping spread spectrum, with a processing gain of 10-30 dB. The synchronization identifier code is a specific digital sequence used for frame synchronization and clock recovery, typically employing Barker codes or m-sequences. Adaptive power adjustment is based on signal strength indications from the receiver, with an adjustment step of 1 dB. Line-of-sight propagation refers to the straight-line propagation of electromagnetic waves, with propagation distance limited by antenna height and Earth's curvature.

[0098] Each device receives timing reference signals from the target unit and achieves precise synchronization with the time reference through an adaptive wideband phase-locked loop (PLL) structure. The PLL device comprises three core components: a digital phase detector, an adaptive loop filter, and a high-performance voltage-controlled oscillator (VCO). The digital phase detector detects the amplitude difference and frequency offset between the local clock and the reference signal; the adaptive loop filter adjusts its filtering parameters in real time according to the channel's dynamic characteristics; and the high-performance VCO provides a wide tuning range and fast response capability. The adaptive wideband PLL is a tracking system specifically designed for highly dynamic environments, capable of rapidly tracking rapid changes in frequency and amplitude. The digital phase detector is implemented using a complex multiplier and the CORDIC algorithm. The adaptive loop filter employs a variable parameter design, automatically adjusting its bandwidth based on the Doppler frequency shift rate. The high-performance VCO utilizes direct digital frequency synthesis technology.

[0099] After receiving the timing reference signal from the target drone, the signal is first preprocessed through a bandpass filter and automatic gain control circuit, and then converted into a digital baseband signal by a digital downconverter. The adaptive phase-locked loop system includes a Doppler frequency shift estimation module, which calculates and compensates for frequency shifts caused by high-speed motion in real time. A digital phase detector compares the compensated reference signal with the local clock signal in terms of amplitude and frequency, generating an error control signal. The adaptive loop filter dynamically adjusts the loop bandwidth according to the current motion state and channel variation characteristics, achieving an optimal balance between tracking speed and noise suppression. The high-performance voltage-controlled oscillator quickly adjusts the frequency and amplitude of the local clock according to the filtered control signal, ultimately achieving high-precision synchronization with the reference signal. The Doppler frequency shift estimation uses a maximum likelihood estimation algorithm; the adaptive bandwidth adjustment algorithm determines the loop bandwidth based on the rate of change of speed. The calculation formula is: adaptive bandwidth equals base bandwidth plus the rate of change of speed multiplied by an adjustment coefficient, where the base bandwidth is 50Hz, and the adjustment coefficient is determined through Kalman filtering optimization.

[0100] An adaptive broadband phase-locked loop (PLL) design is employed, with the loop bandwidth dynamically adjusted according to the target drone's motion state to ensure rapid tracking capability under high-speed maneuvering conditions. For high-speed target drone scenarios, the system integrates a Doppler pre-compensation algorithm, using motion state information to predict Doppler frequency shift and actively perform frequency pre-correction. Through multi-stage closed-loop feedback control, it can automatically compensate for clock offsets caused by various factors such as propagation delay variations, Doppler effects, temperature drift, and oscillator aging. Achieving synchronization accuracy relies on intelligent optimization of adaptive control parameters, including real-time adjustment of dynamic loop bandwidth, variable time constant, and adaptive damping coefficient. After multiple compensations and adaptive control, even under high-speed maneuvering conditions, the deviation between the target drone's local clock and the master timing reference can still be controlled within a preset deviation threshold, meeting the stringent timing requirements of coordinated communication. The dynamic loop bandwidth refers to the tracking bandwidth that is adjusted in real time according to the dynamic characteristics of the input signal. Narrow bandwidth is used in low dynamic environments to reduce noise, while wide bandwidth is used in high dynamic environments to improve tracking speed. Doppler pre-compensation predicts frequency offset through carrier motion parameters. Multi-level closed-loop control includes a coarse adjustment loop and a fine adjustment loop. The coarse adjustment loop is responsible for large-range frequency acquisition, while the fine adjustment loop is responsible for high-precision amplitude tracking. The intelligent optimization of adaptive control parameters adopts a recursive least squares algorithm to adjust control parameters in real time based on historical performance data.

[0101] Accurate calculation of clock offset requires consideration of the combined effects of various dynamic factors. The calculation methods include: base offset calculation is the target's local clock time minus the received synchronization time reference; propagation delay compensation is calculated based on real-time distance and radio wave propagation speed, with the compensation value equal to the distance divided by the speed of light; Doppler effect compensation is calculated based on relative motion speed to account for the time deviation caused by frequency offset; atmospheric propagation correction considers the influence of the ionosphere and troposphere on signal propagation speed. The final clock offset equals base offset minus propagation delay compensation minus Doppler effect compensation minus atmospheric propagation correction. Real-time distance is obtained through two-way ranging technology; Doppler effect compensation is calculated as frequency offset divided by carrier frequency multiplied by the time interval; atmospheric propagation correction uses a standard atmospheric model with a correction accuracy of 0.1% of the propagation delay.

[0102] A multi-mode adaptive controller is employed for precise clock offset correction. The controller includes three operating modes: fast acquisition mode, precise tracking mode, and high dynamic tracking mode. Fast acquisition mode is used for initial synchronization establishment, employing a large bandwidth for rapid convergence; precise tracking mode is used for maintaining high-precision synchronization in stable environments; and high dynamic tracking mode is specifically designed for high-speed maneuvering conditions. The controller automatically selects the optimal operating mode based on the current error magnitude, rate of change, and predicted trend. This multi-mode control strategy significantly improves the system's adaptability to different motion states, ensuring a synchronization accuracy of ±1 millisecond under various conditions. The mode switching criteria are as follows: fast acquisition mode is entered when the synchronization error is greater than 10 milliseconds; precise tracking mode is entered when the synchronization error is less than 1 millisecond and the rate of change is less than 0.1 milliseconds / second; and high dynamic tracking mode is entered when the velocity change rate is greater than 10 meters per square second. Control parameters are dynamically adjusted in different modes: the loop bandwidth is 1 kHz in fast acquisition mode, 50 Hz in precise tracking mode, and the loop bandwidth in high dynamic tracking mode is adaptively adjusted based on dynamic parameters.

[0103] The system provides real-time evaluation of the overall performance of the master target device, automatically triggering a master-slave switchover when performance drops below a preset overall threshold. Performance evaluation employs multi-dimensional metrics, including signal-to-noise ratio (SNR) stability, communication reliability, and coverage area. Communication reliability is calculated through statistical analysis of bit error rate data; coverage is calculated by comparing the already covered area with the total area requiring coverage. The three metrics are weighted and summed to calculate the overall performance evaluation, with SNR stability having a weight of 0.4, communication reliability a weight of 0.4, and coverage area a weight of 0.2. The overall performance evaluation value ranges from 0 to 1. Signal-to-noise ratio (SNR) stability refers to the magnitude of SNR variation, measured by calculating the standard deviation of SNR within a certain time window. A standard deviation less than 2 dB is rated as 1, and greater than 10 dB is rated as 0. Communication reliability refers to the stability of the communication link, assessed by statistically analyzing the number and duration of communication interruptions. A bit error rate less than 10^-5 with no interruptions is rated as 1, and a bit error rate greater than 10^-3 or an interruption lasting more than 5 seconds is rated as 0. Coverage area refers to the geographical area where the main target drone can effectively communicate. Affected by transmission power, antenna pattern, and terrain, a coverage rate of 95% or higher is rated as 1, and less than 60% is rated as 0.

[0104] Performance monitoring performs a comprehensive evaluation every 30 seconds and establishes a multi-level threshold alarm mechanism. When the overall performance evaluation falls below the warning threshold of 0.7, an early warning signal is issued; when it falls below the danger threshold of 0.5, a master-slave switchover process is automatically triggered. The default value for the preset overall threshold is 0.5. The switchover process employs seamless switchover technology, with the pre-synchronized backup master target machine immediately taking over to ensure uninterrupted communication services.

[0105] After establishing stable timing synchronization, in order to obtain accurate channel state information for subsequent frequency hopping sequence generation and optimization, the communication method needs to perform high-precision time-division multiplexing channel detection.

[0106] Step 104: Based on the timing reference, the time axis is divided into N+1 non-overlapping time slots. Each target device sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information.

[0107] Specifically, such as Figure 2 As shown, it includes:

[0108] Timeline unification and synchronization: All target drones and ground stations are strictly synchronized based on the timing reference established in step 103 to ensure that all nodes in the system have a completely consistent understanding of the timeline. The timing reference provides a unified time reference and achieves microsecond-level time synchronization.

[0109] The frame structure design, based on a unified timing reference, divides the continuous time axis into periodic time frames. The length of each time frame is determined according to the number of target devices and the required detection accuracy, typically ranging from 10 to 100 milliseconds. The start time of each time frame is strictly determined according to the timing reference to ensure that the frame boundaries of all participating nodes are perfectly aligned.

[0110] Time slots are divided and allocated as follows: each time frame is divided into N+1 time slots of equal length. The length of each time slot is determined by dividing the time frame length by the total number of time slots. The first N time slots are allocated to N target drones for dedicated channel probing, and the (N+1)th time slot is allocated to the ground station for data aggregation and processing. Time slot allocation strictly follows a timing reference. The start time of the time slot for the i-th target drone is equal to the drone's sequence number minus one, multiplied by the length of a single time slot, ensuring that the time slots are strictly arranged in timing order and do not overlap.

[0111] A time-slot protection mechanism is established based on timing references to monitor in real time whether each target device strictly transmits probe signals according to the allocated time slots. Any unauthorized use of time slots will be detected and corrected immediately to ensure the strict implementation of time-division multiplexing.

[0112] A time slot allocation table is established and broadcast to all participating nodes via dedicated control signaling. The table contains crucial information such as the precise start time, duration, and assigned target device number for each time slot. A strict time slot protection mechanism is implemented to monitor in real time whether each target device accurately transmits probe signals according to its allocated time slot, preventing time slot conflicts and unauthorized occupation.

[0113] Target devices that misuse time slots will be temporarily disabled by the communication method until the problem is resolved and re-verified. Time slot usage statistics are recorded in real time, providing a basis for communication method optimization and fault diagnosis.

[0114] Each target drone transmits an orthogonal probe sequence within its allocated dedicated time slot. The ground station calculates the detailed channel state of each link based on the received probe signals, obtaining channel state information. The target drones transmit known orthogonal probe sequences, and the probe signals received by the ground station are the signals after the orthogonal probe sequences have propagated through the wireless channel—that is, the result of convolving the original probe sequence with the channel impulse response. The ground station extracts channel characteristic information by comparing the known orthogonal probe sequences with the actual received probe signals. Channel state information includes the channel impulse response, the number of multipath taps, and multipath propagation parameters. The channel impulse response consists of the complex gain coefficients of multiple multipath components, and the multipath propagation parameters include the delay, magnitude attenuation, and phase offset of each path.

[0115] The propagation and reception process of the probe signal specifically includes: the orthogonal probe sequence sent by the target device has known time and frequency domain characteristics; when the probe sequence propagates in the wireless channel, it is affected by factors such as multipath propagation, fading, noise, and interference; the probe signal received by the ground station is the result of convolving the original probe sequence with the channel impulse response and then adding noise, that is, the received signal is equal to the response of the orthogonal probe sequence after propagation through the channel plus environmental noise; the received signal contains channel characteristic information, but it is no longer the original orthogonal probe sequence, but a signal after channel distortion; the signal processor of the ground station extracts the channel impulse response using reverse engineering methods based on the known relationship between the orthogonal probe sequence and the received signal.

[0116] The specific methods for obtaining the channel impulse response include: channel estimation is achieved by performing matched filtering and correlation processing on the received signal. First, the received probe signal is correlated with a known orthogonal probe sequence. The correlation operation is performed by integrating the conjugate of the probe signal and the orthogonal probe sequence at different time delay positions. The integration result is the estimated value of the channel impulse response corresponding to that time delay position. Due to the multipath effect in the wireless propagation environment, the transmitted signal will arrive at the receiver through different paths with different time delays. Therefore, the result of the correlation operation is represented in the time domain as peaks at multiple time delay positions. Each multipath component corresponds to a peak output by the correlation function. The time position of the peak reflects the propagation delay of the path. The complex value of the peak is the complex gain coefficient of the path, which includes magnitude and argument information. In this context, the peak value refers to the local maximum point of the correlation function output in the time domain. Specifically, it refers to the peak point reached by the correlation function at a specific time delay. The amplitude and location of this peak point contain key information about multipath propagation: the peak position corresponds to the propagation delay of the corresponding multipath component, and the complex value of the peak contains the complex gain coefficient of that multipath component. The peak formation mechanism is that when the received signal and the orthogonal detection sequence are aligned in time delay, the correlation function output reaches its maximum value, forming a peak. When they are not aligned, the correlation function output is smaller, thus exhibiting a clear peak characteristic on the time-domain correlation function curve. The number of peaks is equal to the number of effective multipath components. Each peak corresponds to a propagation path, the height of the peak reflects the signal strength of that path, and the width of the peak is affected by the autocorrelation characteristics of the detection sequence. The number of multipath taps is determined using a power threshold detection method. Multipath components exceeding a preset threshold are considered effective taps. The preset threshold is set to -20 dB relative to the power of the strongest path. The number of taps is usually controlled between eight and sixteen to balance estimation accuracy and computational complexity, and the specific number is dynamically adjusted according to the complexity of the propagation environment. The strongest path refers to the path with the highest power among all propagation paths in a multipath propagation environment.

[0117] The specific methods for obtaining multipath propagation parameters include: the time delay of each path is determined by the position of the peak, and the time delay is equal to the time position of the peak minus the arrival time of the direct path, with measurement accuracy reaching the nanosecond level. Interpolation algorithms are used to improve the resolution of time delay estimation; the magnitude attenuation is calculated from the magnitude of the peak, and the attenuation value is equal to 20 times multiplied by the received signal magnitude divided by the common logarithm of the transmitted signal magnitude, with the unit being decibels, and a typical value range of 0-40 dB; the phase offset is calculated from the phase angle of the peak, and the phase offset is equal to the received signal phase angle minus the transmitted signal phase angle, with a range of -π to π radians, and the phase angle measurement accuracy is 0.01 radians. The location of the peak refers to its coordinates on the time delay axis, determined by a peak detection algorithm. This algorithm first finds the zero point by differentiating the relevant function, then uses the second derivative to determine if it is a peak. The magnitude of the peak refers to the amplitude of its complex value, calculated by the square root of the sum of the squares of the real and imaginary parts of the complex number. The argument of the peak refers to its phase angle, calculated by the ratio of the imaginary to the real part of the complex number using the arctangent function, with a range limited to -π to π. The accuracy of peak detection directly affects the accuracy of multipath parameter estimation. Multiple threshold detection and noise suppression algorithms are used to ensure the reliability of peak detection, while a super-resolution algorithm is used to separate similar peaks. The measurement of all the above parameters considers the influence of the Doppler effect, which is compensated and corrected using carrier motion state information.

[0118] Orthogonal probe sequences (OPS) are a set of signal sequences that are mutually orthogonal. Mathematically, the cross-correlation function of any two different sequences is zero at zero time delay, while the autocorrelation function has a significant peak at zero time delay. This orthogonality allows the receiver to effectively separate and identify signals from different sources, even if these signals arrive at the receiver simultaneously. OPS are primarily obtained through pseudo-random code generation. Commonly used sequences include Gold sequences, Kasami sequences, and Walsh codes. Gold sequences are generated by adding two m-sequences modulo-2, resulting in a sequence length of 2 to the power of n, suitable for medium-sized systems. Kasami sequences are generated by decimating and combining m-sequences, exhibiting a smaller maximum cross-correlation value, suitable for systems requiring strict orthogonality. Walsh codes are generated based on the Hadamard matrix, with a sequence length of 2 to the power of n, suitable for synchronization systems. In practical applications, the selection of sequences must consider factors such as system size, anti-interference requirements, and computational complexity.

[0119] Channel states are mathematically described using the Finite Impulse Response (FIR) model, which treats a complex wireless channel as a linear time-invariant system consisting of multiple propagation paths with different time delays. Channel impulse response is a time-domain representation of channel characteristics, while the FIR model is a specific implementation that limits the system's memory length, ensuring that the system responds only to inputs within a finite time interval.

[0120] The communication method employs a correlation receiver architecture for accurate channel impulse response estimation. The process first transmits a known probe sequence, then calculates the correlation between this probe sequence and the actual received signal, extracting the channel impulse response characteristics from the correlation results. In practical implementation, the correlator uses Fast Fourier Transform (FFT) technology to convert time-domain convolution into frequency-domain multiplication, significantly improving computational efficiency and real-time processing capabilities.

[0121] For complex multipath environments, traditional correlation methods have limitations in terms of time delay resolution. Therefore, super-resolution algorithms (such as MUSIC or ESPRIT) are further employed for accurate multipath component estimation. These algorithms can identify and separate multipath components with very similar time delays, overcoming the resolution limitations of conventional Fourier analysis and providing higher estimation accuracy than traditional methods. The complete multipath parameter estimation results include information such as the time delay, magnitude attenuation, and phase offset of each path. This information constitutes accurate channel impulse response parameters, providing a reliable foundation for subsequent frequency hopping sequence generation and optimization.

[0122] The complete channel estimation process includes: each target drone strictly transmits known orthogonal probe sequences according to the allocated time slots to ensure the timing accuracy and sequence orthogonality of signal transmission; the ground station receives all probe signals and performs precise time slot separation processing to extract signals belonging to different target drones; the signals within each time slot undergo specialized correlation processing, using the known probe sequences as matched filters to extract the channel impulse response of that link from the received signals; the channel impulse response is optimized using the minimum mean square error estimation method to reduce the impact of noise and interference on estimation accuracy; all channel estimation results are statistically analyzed and quality assessed, then stored in the channel state database to obtain channel state information, which is then broadcast to the relevant target drone nodes in a timely manner.

[0123] The channel estimation processor is implemented using a dedicated digital signal processing chip or a field-programmable gate array (FPGA) to ensure it meets the stringent requirements of real-time processing. The communication method requires the normalized mean square error (MSE) of the channel estimation to be less than 0.01. This threshold was determined through extensive simulation experiments, ensuring the accuracy of subsequent frequency hopping sequence generation. The estimation accuracy is continuously verified using dedicated pilot signals. These pilot signals employ QPSK modulation with a power of -10 dBm, transmitted every 100 milliseconds for a duration of 1 millisecond. A complete historical channel state database is established, using a time-series storage structure to store channel state data for the most recent 24 hours, with a storage capacity of 10 GB. This provides reliable data support for subsequent frequency hopping sequence generation, collision prediction, and communication method optimization.

[0124] After obtaining high-precision channel state information, and combining the spectrum allocation results and timing references, the communication method has all the necessary conditions to generate adaptive frequency hopping sequences.

[0125] Step 105: Convert the channel state information into a frequency hopping key; generate a frequency hopping sequence based on the frequency hopping key and the spectrum allocation result.

[0126] Specifically, such as Figure 3 and Figure 4 As shown, it includes:

[0127] Frequency hopping key generation employs a multi-channel feature fusion method to improve the randomness and unpredictability of the key. Key information is extracted from the channel state information obtained in step 104 for key generation. The key information includes the magnitude of the complex gain coefficient, the argument of the complex gain coefficient, and the time rate of change of the channel state. The magnitude of the complex gain coefficient reflects the amplitude fading characteristics of the channel; the argument of the complex gain coefficient reflects the phase change characteristics of the channel; and the time rate of change of the channel state reflects the dynamic change characteristics of the channel.

[0128] The complex gain coefficient mentioned here specifically refers to the complex gain coefficient of the strongest multipath component in the channel impulse response. The method for selecting the strongest multipath component is as follows: sort all multipath components in the channel impulse response obtained in step 104 by power, and select the component with the largest complex gain coefficient magnitude as the strongest multipath component. This component usually corresponds to the direct path or the main reflection path, and has the highest signal-to-noise ratio and the most stable estimation accuracy.

[0129] The time-domain channel impulse response is obtained through correlation receiver processing in step 104. The channel impulse response consists of multiple complex gain coefficients, each corresponding to a multipath component. The complex gain coefficients are in complex form, consisting of a real part and an imaginary part. The real part reflects the in-phase component, and the imaginary part reflects the quadrature component. The complex gain coefficients are calculated using a digital signal processor. The squares of the real and imaginary parts of the complex gain coefficients are added together, and the square root of the result is taken to obtain the absolute value of the complex number, which is the magnitude of the complex gain coefficient. The imaginary part of the complex gain coefficient is divided by the real part, and the arctangent function of the result is calculated to obtain the phase angle of the complex number, which is the argument of the complex gain coefficient.

[0130] The time rate of change of the channel state is obtained by the ratio of the magnitude of the change in the complex gain coefficient at adjacent time points to the time interval.

[0131] The specific methods for obtaining the change in the complex gain coefficient between adjacent time steps include:

[0132] The cache stores the complex gain coefficients of the strongest multipath components in the last 10 sampling times. The cache adopts a first-in-first-out (FIFO) queue structure with a queue length of 10 complex elements, and each element occupies 8 bytes of storage space.

[0133] To reduce the impact of noise, a three-point moving average filter is applied to the original complex gain coefficient data. The filtering method is to multiply the complex gain coefficient at the previous time step by 0.25, the complex gain coefficient at the current time step by 0.5, and the complex gain coefficient at the next time step by 0.25, and then add the three together to obtain the filtered complex gain coefficient.

[0134] The change in the complex gain coefficient between adjacent time steps is obtained by subtracting the filtered complex gain coefficient from the previous time step from the current time step, where the time interval is the sampling interval. The change is calculated in the complex domain, including both real and imaginary changes, and the result is also in complex form.

[0135] The key information is quantized independently. The magnitude of the complex gain coefficient is converted into an integer value between 0 and 255 using an 8-bit quantizer. The quantization method is to multiply the magnitude by 255 and then round down to obtain the digital representation of the magnitude. The argument of the complex gain coefficient is mapped from negative π to positive π radians to an integer value between 0 and 255. The quantization method is to first add π to the argument, then multiply by 255, then divide by 2π, and finally round down to obtain the digital representation of the argument. The rate of change of time is logarithmically transformed and normalized before being quantized by 8 bits. The quantization method is to first add one to the rate of change of time, then take the logarithm to base 10, then multiply by 255, then divide by the result of adding one to the historical maximum rate of change of time, then take the logarithm, and finally round down to obtain the digital representation of the rate of change of time.

[0136] The three 8-bit quantization results are combined through a bit concatenation operation to generate a 24-bit base key. The modulus quantization result is used as the high 8 bits, the argument quantization result as the middle 8 bits, and the time rate of change quantization result as the low 8 bits. These are concatenated in sequence to form the 24-bit base key. The modulus quantization result occupies bits 23 to 16, the argument quantization result occupies bits 15 to 8, and the time rate of change quantization result occupies bits 7 to 0.

[0137] To enhance the randomness and diffusion of the key, the 24-bit base key is shifted and XORed. The 24-bit base key is XORed with the result of shifting it 8 bits to the right, and then XORed with the result of shifting it 16 bits to the right, resulting in the mixed key. This process ensures that every 8 bits of the original 24-bit key influence each other, thus enhancing the random distribution characteristics of the key.

[0138] The mixed key is expanded to 128 bits through multiple shift and concatenation operations. Specifically, bits 127 to 104 are the original 24-bit value of the mixed key; bits 103 to 80 are the lower 24 bits of the mixed key after shifting left by 8 bits; bits 79 to 56 are the lower 24 bits of the mixed key after shifting left by 16 bits; bits 55 to 32 are the lower 24 bits of the mixed key after shifting left by 24 bits; bits 31 to 8 are the lower 24 bits of the mixed key after bitwise inversion; and bits 7 to 0 are filled with the lower 8 bits of the mixed key.

[0139] The final 128-bit frequency hopping key is used in the subsequent frequency hopping sequence generation process to ensure that the key has sufficient length and good randomness.

[0140] The master frequency hopping sequence is a time-varying frequency sequence, represented as: ,in Indicates the first Frequency hopping frequency of each time slot, sequence length The number of frequency points is typically determined by the communication period and frequency hopping rate, ranging from 1000 to 10000. The main frequency hopping sequence is generated using a direct key-to-frequency-point mapping method. The core algorithm of the mapping function is to perform a modulo operation on the size of the first frequency point set using the generated frequency hopping key, add one to the remainder as the index value, and select the corresponding frequency point from the pre-established first frequency point set as the frequency hopping frequency for the current moment. The complete main frequency hopping sequence is generated by repeatedly applying this mapping method over consecutive time slots. The frequency hopping key for each time slot is dynamically updated based on real-time channel state information, ensuring the sequence's adaptability and anti-predictability.

[0141] This mapping method ensures the uniform distribution and good randomness of frequency selection. The mapping function is implemented using an efficient lookup table method, pre-establishing and periodically updating the key-to-frequency mapping table, which is 65536×4 bytes in size, and adjusting the frequency allocation strategy in a timely manner according to changes in the spectrum environment.

[0142] The method for generating the frequency hopping sequence from the target device includes: the target device generating a frequency hopping key based on the channel state information it receives, with a key length of 128 bits; processing the frequency hopping key using a SHA256 hash function to enhance randomness and anti-prediction capability; receiving the master frequency hopping sequence information broadcast by the master target device, which includes parameters such as the current frequency hopping point, sequence index, and timestamp; applying a specially designed orthogonal mapping function, combining the hashed frequency hopping key and the master frequency hopping sequence information to generate the frequency hopping sequence from the target device; and performing the final frequency point selection and confirmation within the dedicated sub-frequency band allocated in step 102.

[0143] The hash function uses the SHA256 algorithm to process the key from the target machine. The SHA256 algorithm uses a series of complex bitwise operations, including circular left shift, logical XOR, and modulo operations, to convert an input key of arbitrary length into a hash value of fixed 256 bits. The SHA256 algorithm was chosen based on its cryptographic security and uniform output distribution, and is currently a widely accepted secure hash algorithm.

[0144] Specific implementation methods of orthogonal mapping functions include:

[0145] Based on the spectrum allocation results of step 102, each drone obtains a corresponding dedicated sub-band frequency range from the target drone, and the set of first frequency points within the sub-band. Represented as ,in For the first The first target drone The first frequency point For the first The number of target drones at the first frequency point is set, and the frequency points are arranged in ascending order of frequency to facilitate subsequent indexing operations.

[0146] The generation of frequency hopping points from the target machine employs a four-element XOR hybrid mapping method. Specifically, this involves converting the target machine's hashed frequency hopping key, the current frequency hopping frequency in the master frequency hopping sequence, the eight-bit left shift result of the target machine's number, and the current timestamp into a 32-bit data format, then performing an XOR operation to obtain a composite key value. Next, a modulo operation is performed between the composite key value and the number of first frequency points of the target machine to obtain the base frequency point index. Since the array index starts from zero while the frequency point number starts from one, the base index needs to be incremented by one to obtain the final frequency point selection index. Finally, this index value is used to select the corresponding frequency point from the set of first frequency points of the target machine as the current frequency hopping frequency.

[0147] The slave target number is a unique numerical designation assigned to each slave target device. It is used to distinguish different slave targets in the orthogonal mapping function and generate differentiated frequency hopping sequences. The slave target number uses an eight-bit unsigned integer format, with a value ranging from 1 to 255. Identifiers are assigned sequentially: the first slave target device to join the system is assigned identifier 1, the second is assigned identifier 2, and so on, ensuring that each slave target device has a unique digital identity. Identifiers are uniformly assigned by the ground station and issued via security signaling when the slave target device initializes its access to the system. After receiving the identifier, the slave target device confirms its receipt, establishing a binding relationship between the identifier and the actual device. Once assigned, the identifier remains unchanged throughout the communication session. When a slave target device leaves the system, its identifier is reclaimed and can be assigned to a newly joined slave target device.

[0148] To ensure the correct execution of the XOR operation, the four participating values ​​need to be uniformly converted to a 32-bit unsigned integer format. The specific conversion methods include: taking the lower 32 bits of the SHA256 algorithm output from the hashed frequency hopping key as the first operand; truncating or padding the integer part of the frequency value in the main frequency hopping sequence to 32 bits as the second operand; performing an eight-bit left shift operation on the target machine number to obtain a 16-bit value, then padding the high bits with zeros to 32 bits as the third operand; and taking the lower 32 bits of the system time from the current timestamp or truncating the high bits to 32 bits as the fourth operand. After the four 32-bit values ​​are uniformly converted, a bitwise XOR operation is performed, with each bit being XORed separately, ultimately resulting in a 32-bit composite key value used for subsequent modulo operations and frequency index calculations.

[0149] The lower 32 bits of the hashed slave frequency hopping key provide the basic source of randomness; the current frequency value of the master frequency hopping sequence ensures its correlation and synchronization with the master sequence; the result of shifting the target machine number eight bits to the left guarantees the difference and independence of the mapping results for different slave targets; the current timestamp adds variability and dynamic characteristics in the time dimension. The four parameters are fully mixed through bitwise XOR operations to ensure a statistically uniform distribution and high degree of randomness in the mapping results.

[0150] Since each slave target drone has its own independent sub-frequency band and a non-overlapping set of first frequency points, meaning that there are no common frequency points between the frequency point sets of any two different slave target drones, the frequency hopping points generated by different slave target drones through the orthogonal mapping function will not be the same at any time, thus ensuring complete orthogonality. The master target drone uses a dedicated sub-frequency band different from that of the slave target drones, further ensuring the orthogonality between the master and slave sequences.

[0151] After each mapping calculation, the system automatically verifies the validity of the selected frequency point, including whether the frequency point is within the allocated sub-band range, whether it meets signal quality requirements, and whether it conflicts with the frequency points of other target devices at the current time. Verification is achieved through real-time table lookup and comparison operations. If an invalid frequency point is found, the system automatically selects the next candidate frequency point from the first frequency point set, attempting a maximum of 3 times to ensure the continuity and reliability of the frequency hopping sequence.

[0152] The frequency hopping period length is dynamically adjusted within a preset time range based on system load and real-time requirements, matching the synchronization accuracy of the timing reference to ensure timing consistency. When a sequence violating orthogonality constraints is detected, the communication method immediately marks it as invalid and triggers a regeneration process. A dedicated orthogonality statistics database is established, recording the results of each orthogonality check, including check time, number of sequences involved in the check, and number of collisions found. This database continuously analyzes sequence quality and potential collision probabilities, providing data support for optimizing the communication method.

[0153] A specially designed handshake protocol ensures precise synchronization of the frequency hopping sequences across all nodes. Sequence verification employs a cyclic redundancy check (CRC) method, performing polynomial operations on the generated frequency hopping sequences to produce a 16-bit checksum. The receiving end uses the same calculation method to generate a checksum and compares it to the received sequence to verify its integrity and correctness.

[0154] The Cyclic Redundancy Check (CRC) polynomial adopts the CRC-16-CCITT standard, with a generator polynomial of power 16 + power 12 + power 5 + 1. This standard is widely used in communication systems and has good error detection capabilities. The handshake protocol uses a three-way handshake mechanism to ensure the reliability of sequence synchronization. The handshake process includes three stages: sequence request, sequence acknowledgment, and synchronization completion. When an error is detected during verification, the communication method automatically requests retransmission, retrying a maximum of three times with a retry interval of 100 milliseconds. If the process still fails, a sequence regeneration process is triggered.

[0155] After generating high-quality frequency hopping sequences, in order to cope with dynamic environmental changes and prevent potential conflicts, the communication method needs to establish a prediction-based dynamic sequence update and conflict avoidance mechanism.

[0156] Step 106: Based on channel state information and frequency hopping sequences, including master frequency hopping sequences and slave target frequency hopping sequences, the frequency collision probability at future time is calculated using statistical methods, and a collision probability threshold is set. When the frequency collision probability exceeds the set collision probability threshold, an adaptive adjustment mechanism is used to update the sequence; proactively preventing the occurrence of frequency hopping sequence collisions and ensuring the continuous and stable operation of the communication system in a dynamic environment.

[0157] The intelligent frequency collision probability calculation, based on channel state information and the generated frequency hopping sequence, calculates the probability of frequency collision occurring at future times. The collision probability calculation employs statistical analysis methods, calculating the probability of frequency overlap at future times for each possible target pair combination, and then summing the results to obtain the total collision probability of the system.

[0158] The conflict probability calculation employs the Monte Carlo statistical method, estimating probability values ​​through extensive random sampling to improve the accuracy and reliability of probability estimation. The number of Monte Carlo samples is set to 10,000, determined based on statistical convergence analysis, with a sampling interval of 1 millisecond, covering a 100-millisecond time window. The calculation frequency is set to 10 times per second to achieve real-time monitoring of conflict risk. The communication method sets a conflict probability threshold based on communication service quality requirements, typically between 0.05 and 0.1, with this threshold range determined based on system reliability requirements.

[0159] When the predicted frequency collision probability exceeds a set collision probability threshold, the communication method automatically employs multiple adaptive adjustment mechanisms to adjust the sequence, reducing the collision risk to an acceptable range. The selection of adaptive adjustment mechanisms is based on the principle of minimizing the impact on system performance, and includes methods such as sequence offset adjustment, backup sequence switching, and frequency point replacement.

[0160] Sequence offset adjustment is achieved by adding an appropriate frequency offset to the original frequency hopping sequence, maintaining the basic structure of the sequence. The frequency offset is determined by the sub-band spacing in the spectrum allocation step 102, typically 10% of the sub-band bandwidth; for example, when the sub-band bandwidth is 10MHz, the frequency offset is 1MHz. Backup sequence switching uses pre-generated and verified alternative frequency hopping sequences to ensure rapid response; the number of backup sequences is set to twice the number of primary frequency hopping sequences. The frequency point replacement method performs local adjustments for specific conflicting frequency points, minimizing the impact on overall communication performance.

[0161] The adjustment strategy employs a tiered threshold judgment mechanism, selecting appropriate countermeasures based on different conflict probability levels. When the conflict probability is high, a complete switchover is performed using a pre-prepared backup sequence; when the conflict probability is at a medium level, local modifications and optimizations are made to the sequence; when the conflict probability is low but still exceeds the safety boundary, the original sequence is maintained while the monitoring frequency is increased.

[0162] The communication method sets two key probability thresholds: a high collision probability threshold of 0.1, which immediately triggers a switchover to the backup sequence when this value is exceeded; and a low collision probability threshold of 0.05, serving as a safety boundary indicator for normal operation. These thresholds were determined based on extensive simulation experiments and actual test data. The simulation experiments included multiple independent runs, and the test data came from actual flight tests. The adjustment strategy is implemented using an expert system, making intelligent decisions based on historical experience and operational rules.

[0163] All adjustment processes employ a smooth transition mechanism to avoid communication service interruptions. Backup sequences are pre-generated and verified, stored in fast-access memory, ensuring immediate switchover when needed. The number of backup sequences is set to twice the number of primary frequency-hopping sequences, using the same generation algorithm but with a different initial key. Adjustment commands are sent to the relevant target machines via reliable control signaling, protected by BCH encoding with 3-bit error correction capability, and containing detailed switching times and new sequence parameters.

[0164] Through the complete implementation of this embodiment, an efficient and reliable coordinated dynamic frequency hopping communication system is established between multiple target drones and the ground station. Each step is interconnected, forming a complete closed-loop control system that effectively solves the problems of multi-link conflicts and synchronization.

[0165] Example 2:

[0166] See Figure 5 As shown, a dynamic spectrum frequency hopping communication device for a target drone and a ground station is provided. This device stores computer-readable instructions, which, when read, enable the execution of the aforementioned dynamic spectrum frequency hopping communication method between the target drone and the ground station. The device includes:

[0167] The collection module 701 collects status information, including task priority and total spectrum range.

[0168] The allocation module 702 monitors power data in real time and evaluates signal quality indicators based on the power data; based on the total spectrum range and signal quality indicators, it divides the total spectrum range into N sub-bands to obtain the spectrum allocation result.

[0169] Timing module 703 calculates communication quality indicators based on power data, elects the master target machine based on task priority and communication quality indicators, and establishes a timing benchmark;

[0170] The channel module 704 divides the time axis into N+1 non-overlapping time slots based on the timing reference. Each target device sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information.

[0171] The frequency hopping module 705 converts channel state information into a frequency hopping key and generates a frequency hopping sequence based on the frequency hopping key.

[0172] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A dynamic spectrum frequency hopping communication method between a target drone and a ground station, characterized in that, Includes the following steps: Collect status information, including task priority and total spectrum range; Real-time monitoring of power data; evaluation of signal quality indicators based on power data; division of the total spectrum range into N sub-bands based on the total spectrum range and signal quality indicators to obtain spectrum allocation results; Communication quality indicators are calculated based on power data, and the main target machine is elected based on task priority and communication quality indicators; a timing reference is established based on GPS or BeiDou time. Based on the timing reference, the time axis is divided into N+1 non-overlapping time slots. Each target device sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information. Convert channel state information into frequency hopping keys; A frequency hopping sequence is generated based on the frequency hopping key and spectrum allocation results. The frequency hopping sequence includes a master frequency hopping sequence and a slave target frequency hopping sequence. The steps for generating the frequency hopping sequence include: The main target device uses the generated main frequency hopping key to perform a modulo operation on the size of the first frequency point set in the spectrum allocation result. The remainder is added to one as the index value. The frequency point at the corresponding position is selected from the first frequency point set as the frequency hopping frequency at the current moment. The frequency hopping frequencies at multiple moments form the main frequency hopping sequence. The first frequency point set is the set of frequency points that satisfy the minimum interval constraint and the signal quality index is greater than the preset red threshold. The target device generates a slave frequency hopping key based on the received channel state information; processes the slave frequency hopping key using a SHA256 hash function; receives the master frequency hopping sequence broadcast by the master target device; and applies an orthogonal mapping function to generate a slave frequency hopping sequence by combining the hash-processed slave frequency hopping key and the master frequency hopping sequence. The frequency hopping key includes a master frequency hopping key and a slave frequency hopping key. The frequency hopping key is generated using a multi-channel feature fusion method to extract the key information of the path with the largest complex gain coefficient magnitude in the channel state information. The key information includes the magnitude of the complex gain coefficient, the argument of the complex gain coefficient, and the time change rate of the channel state. The key information is quantized, and the quantization results are concatenated to obtain the base key. The base key is then shifted and XORed to obtain the hybrid key. The hybrid key is then subjected to multiple shift and concatenation operations to obtain the frequency hopping key. The magnitude of the complex gain coefficient is the sum of the squares of the real and imaginary parts of the complex gain coefficient, and then the square root of the sum is taken. The argument of the complex gain coefficient is the arctangent function of the imaginary part divided by the real part of the complex gain coefficient. The time change rate of the channel state is obtained by the ratio of the magnitude of the change in the complex gain coefficient at adjacent time points to the time interval. The change in the complex gain coefficient at adjacent time points is obtained by subtracting the complex gain coefficient at the previous time point from the complex gain coefficient at the current time point.

2. The dynamic spectrum frequency hopping communication method between the target drone and the ground station according to claim 1, characterized in that, The orthogonal mapping function converts the hashed frequency hopping key, the current frequency hopping frequency in the master frequency hopping sequence, the slave target number, and the current timestamp into 32 bits, then performs an XOR operation to obtain the composite key value. The composite key value is then moduloed with the number of first frequency points in the slave target, and the remainder is incremented by one to serve as the base frequency point index. The frequency point at the corresponding position is selected from the set of first frequency points of the corresponding slave target using the base frequency point index as the frequency hopping frequency at the current moment.

3. The dynamic spectrum frequency hopping communication method between the target drone and the ground station according to claim 1, characterized in that, The specific steps for dividing the time into N+1 non-overlapping time slots include: All target drones and ground stations are synchronized based on a timing reference. Divide the continuous time axis into periodic time frames; Each time frame is divided into N+1 time slots of equal length. The first N time slots are allocated to N target drones, and the N+1 time slot is allocated to the ground station.

4. The dynamic spectrum frequency hopping communication method between the target drone and the ground station according to claim 3, characterized in that, The detection signal is the signal after the orthogonal detection sequence is propagated through the wireless channel. The orthogonal detection sequence is a signal sequence with mutually orthogonal characteristics. The channel state information includes the channel impulse response, the number of multipath taps, and the multipath propagation parameters. The channel impulse response consists of the complex gain coefficients of multiple multipath components. The multipath propagation parameters include the time delay, magnitude attenuation, and phase offset of each path. The detection signal is correlated with a known orthogonal detection sequence. The correlation operation is performed by integrating the detection signal and the conjugate of the orthogonal detection sequence at different time delay positions. The integration result is the channel impulse response corresponding to the time delay position. The number of multipath taps is determined by a power threshold detection method. Multipath components that exceed a preset threshold are judged as valid taps, and the number of valid taps is the number of multipath taps. The path delay is determined by the position of the channel impulse response in the time domain. The delay is equal to the time position of the peak occurrence minus the arrival time of the direct path. Modulus attenuation is calculated from the modulus of the peak value; The argument offset is obtained by calculating the argument of the peak value.

5. The dynamic spectrum frequency hopping communication method between the target drone and the ground station according to claim 1, characterized in that, The comprehensive score of each target drone is calculated by weighted summation, and the target drone with the highest comprehensive score is selected as the main target drone. The comprehensive score is the sum of the task priority item and the communication quality index item. The task priority item is the normalized task priority multiplied by the preset task priority weight, and the communication quality index item is the normalized communication quality index multiplied by the preset communication quality weight. The communication quality index is the ratio of signal-to-noise ratio to bit error rate. The bit error rate is the distance factor multiplied by the bit error rate. The signal-to-noise ratio is the ratio of useful signal power to noise power. The distance factor is the propagation loss caused by the distance from the target drone to the ground station. The bit error rate is the ratio of the number of received erroneous bits to the total number of transmitted bits.

6. The dynamic spectrum frequency hopping communication method between the target drone and the ground station according to claim 1, characterized in that, The steps for establishing a timing reference include: Obtain GPS or BeiDou time from ground stations; Multiple standard frequency components of different frequencies are generated based on GPS or BeiDou time, including 10MHz, 1MHz, and 100kHz. The standard frequency components are superimposed and synthesized in the frequency domain to generate a timing reference signal; The timing reference signal is broadcast to all target devices.

7. The dynamic spectrum frequency hopping communication method between the target drone and the ground station according to claim 1, characterized in that, Power data includes signal power, noise power, and interference power. The signal quality index is the ratio of signal power to the total noise. The total noise is the sum of noise power and interference power. The spectrum allocation results include the sub-band allocated to each target drone and the set of first frequency points within the corresponding sub-band.

8. A dynamic spectrum frequency hopping communication device for target drones and ground stations, characterized in that, It is used to store computer instructions, which, when read, execute the dynamic spectrum frequency hopping communication method between the target drone and the ground station as described in any one of claims 1-7, wherein the apparatus comprises: The collection module collects status information, including task priority and total spectrum range. The allocation module monitors power data in real time and evaluates signal quality indicators based on the power data; based on the total spectrum range and signal quality indicators, it divides the total spectrum range into N sub-bands to obtain the spectrum allocation result. The timing module calculates communication quality indicators based on power data, elects the master target machine based on task priority and communication quality indicators, and establishes a timing benchmark. The channel module divides the time axis into N+1 non-overlapping time slots based on the timing reference. Each target device sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information. The frequency hopping module converts channel state information into a frequency hopping key; and generates a frequency hopping sequence based on the frequency hopping key.

Images