Satellite communication adaptive frequency hopping method based on time slot constraint and spectrum sensing coupling

Abstract

The application discloses a satellite communication adaptive frequency hopping method based on time slot constraint and spectrum sensing coupling, which comprises the following steps: acquiring time slot resource elements generated by network timing scheduling at the master station side; collecting wideband spectrum covered by a reverse link and performing state sensing and interference determination to obtain availability states of candidate frequency hopping frequency points; performing adaptive closed loop updating based on the availability states to obtain a currently effective available frequency point set; performing position mapping in the currently effective available frequency point set based on the time slot resource elements to determine a target frequency hopping frequency point corresponding to a current time slot; and performing burst frequency hopping communication based on the target frequency hopping frequency point and the time slot resource elements. The application can improve the closed loop synchronization problem under the long space-ground delay, realizes the natural cooperation of frequency point sensing and time slot sequence, and improves the communication anti-interference ability under the complex electromagnetic environment.

Description

Technical Field

[0001] This invention relates to the field of satellite communication technology, and in particular to an adaptive frequency hopping method and system for satellite communication based on time slot constraints and spectrum sensing coupling. Background Technology

[0002] With the evolution of high-throughput satellite communication systems, communication links face increasingly severe anti-interference challenges in complex electromagnetic environments. As the backbone channel for space-to-ground information exchange, the stability of the satellite backhaul link directly determines the availability of the entire communication network. In broadband coverage scenarios, effectively avoiding co-channel interference, adjacent-channel leakage, and sudden external electromagnetic spectrum intrusions to ensure high reliability and low bit error rate transmission of the satellite-to-ground link is a core technical foundation in the design of the underlying physical layer and media access control layer of current broadband satellite communication systems.

[0003] Current conventional solutions for satellite link interference mitigation mostly employ traditional frequency hopping scheduling. In traditional frequency hopping systems, end stations typically rely on local independent clocks or semi-statically configured frequency transfer tables to generate frequency hopping sequences, or use conventional spectrum awareness mechanisms based on single-state detection. In practice, the evolution of traditional frequency hopping sequences is often independent of the on-board network's time slot resource allocation plan. This not only incurs additional signaling overhead to maintain system-level synchronization, but also, when facing geostationary orbit communication channels with propagation delays of hundreds of milliseconds, can easily lead to severe time misalignment between end-to-end interference state feedback and terminal local frequency point execution. Furthermore, when facing fluctuating non-stationary background noise and transient pulses across a wide frequency band, existing judgment schemes based on uniform static thresholds across the entire frequency band are prone to missed detections and frequent system oscillations under dynamic conditions.

[0004] In summary, existing frequency-hopping communication methods struggle to balance timing consistency with dynamic interference resistance when handling high-bandwidth, long-latency satellite-ground collaborative networks. Furthermore, existing methods still lack robustness and environmental adaptability under complex, time-varying electromagnetic conditions. Therefore, it is necessary to investigate a communication method that can improve resource scheduling accuracy and link transmission robustness in long-latency, complex electromagnetic scenarios. Summary of the Invention

[0005] Purpose of the invention: To provide an adaptive frequency hopping method for satellite communication based on time slot constraints and spectrum sensing coupling, in order to solve the above-mentioned problems in the prior art.

[0006] Technical solution: An adaptive frequency hopping method for satellite communication based on time slot constraints and spectrum sensing coupling, comprising: Obtain time slot resource elements generated by network timing scheduling on the main station side; The broadband spectrum covered by the reverse link is collected, and the broadband spectrum is subjected to state perception and interference judgment to obtain the availability status of each candidate frequency hopping point. Adaptive closed-loop update is performed based on the availability status of each candidate frequency hopping point to obtain the currently effective set of available frequency points; Based on time slot resource elements, perform location mapping in the currently effective set of available frequency points to determine the target frequency hopping point corresponding to the current time slot; The target frequency hopping point and corresponding time slot resource elements are sent to the terminal station through the forward link, and the terminal station equipment is controlled to perform burst frequency hopping communication.

[0007] Beneficial effects: This invention directly couples the generation of frequency hopping sequences to existing time slot resources in the forward link, eliminating the need for additional dedicated synchronization signaling. By utilizing the network's own timing structure to drive frequency hopping switching, it can improve the closed-loop synchronization problem of frequency hopping sequences between the master station and the terminal station under long satellite-to-ground delays. Through adaptive threshold construction based on local neighborhood statistics and a dual-threshold hysteresis filtering mechanism, it can reduce false detections and frequent oscillations of the full-band static threshold in non-stationary noise environments. The constructed dual-buffered superframe boundary alignment mechanism locks the effective time of sensing updates at the whole superframe boundary, completely absorbing the uncertain satellite-to-ground propagation delay within the reserved time period, which is beneficial for the deterministic timing coordination of spectrum sensing results and frequency hopping sequence execution. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of the adaptive frequency hopping process in satellite communication coupled with time slot constraints and spectrum sensing.

[0009] Figure 2 This is a schematic diagram of the process for obtaining the target frequency hopping point.

[0010] Figure 3 This is a schematic diagram of the process for determining the state of the broadband spectrum and handling interference.

[0011] Figure 4 This is a schematic diagram of the frequency adaptive update and double-buffered timing alignment process.

[0012] Figure 5 This is a schematic diagram of the overall frequency hopping scheme.

[0013] Figure 6 A schematic diagram of an interference detection module according to an embodiment of the present invention.

[0014] Figure 7 A platform diagram of one embodiment of the present invention.

[0015] Figure 8 A schematic diagram of subband segmentation according to an embodiment of the present invention.

[0016] Figure 9A schematic diagram of frequency domain analysis according to an embodiment of the present invention.

[0017] Figure 10 A schematic diagram illustrating the specific operation steps of interference detection according to an embodiment of the present invention.

[0018] Figure 11 A schematic diagram outlining the time consumption for interference detection according to an embodiment of the present invention. Detailed Implementation

[0019] To more accurately describe the solution in this application, the relevant concepts are defined as follows: Superframe: A superframe is defined as a set of reverse link resources used by a specified remote station. Each superframe has an identifier ID, and multiple temporally consecutive superframes can share a single superframe ID. In this way, the frequency and time of the reverse link can be divided into several superframes and managed centrally by the central station. The time slot allocation table for the superframes is broadcast through the forward link. Each remote station receives the time slot allocation table and can only send bursts in the time slots allocated to it or in randomly accessed time slots.

[0020] Frames: A superframe consists of multiple frames, each with its own frame number. Frame length is not fixed; a single frame can fill an entire superframe, or it can contain only one frequency point and one time slot. Frame usage is flexible, and the system can modify it based on the remote station's hardware configuration, link conditions, and bursty service types.

[0021] Time slot: A time slot is the smallest unit of time allocation and the basic unit that makes up a frame. A frame can contain a maximum of 2048 time slots. Each time slot has its own unique identifier, which consists of the superframe ID number, superframe count, frame number, and time slot number. A frame ID points to a group of managed time slots.

[0022] like Figure 1 , Figure 6 and Figure 7 As shown in Example 1: This example provides an adaptive frequency hopping method for satellite communication based on the coupling of time slot constraints and spectrum sensing. The method includes the following steps: Step 101: Obtain the time slot resource elements generated by the network timing schedule on the master station side; in other words, receive the communication continuous stream sent by the master station through the forward link, and parse and obtain the time slot resource elements from the communication continuous stream. Specifically, the master station maintains continuous forward beam radiation, transmitting a continuous communication stream. End station equipment receives this continuous communication stream through its configured antennas and LNBs. Physically, the continuous communication stream is a continuous time-division multiplexed large carrier signal, carrying system control signaling, a network clock reference field, and the frequency and time resource allocation scheme for the reverse link at the network control protocol layer. Time slot resource elements are structured network resource data used for timing orchestration of reverse link communication, specifically including superframe identifiers, frame identifiers, and time slot identifiers. The master station's NCR generation module is responsible for generating and maintaining the network clock reference. The resource scheduling module directly generates time slot resource elements based on this clock reference for subsequent frequency hopping sequence generation and position mapping.

[0023] Step 102: Collect the broadband spectrum covered by the reverse link, and perform state awareness and interference determination on the broadband spectrum to obtain the availability status of each pre-configured candidate frequency hopping point.

[0024] Specifically, a spectrum sensing module is deployed on the main station side to collect radio frequency signals from the broadband spectrum covered by the reverse link in real time. To balance the coverage of the broadband band with the frequency resolution of digital signal processing, the target coverage range of the broadband spectrum is qualitatively constrained to the overall physical frequency band between 950MHz and 2450MHz. A multi-channel parallel orthogonal sampling architecture is used for signal acquisition. The system deploys five radio frequency receiving channels, each covering five sub-bands that partially overlap in frequency range within the broadband spectrum. The orthogonal sampling method is used, with the system's complex sampling rate set to 491.52Msps, thus obtaining a maximum analyzable bandwidth of 491.52MHz. The acquired RF digital data stream is divided into five sub-bands that partially overlap in frequency range for spectrum estimation processing. The analysis bandwidth of each sub-band is set to 400MHz, and adjacent sub-bands maintain a 100MHz frequency overlap. Candidate frequency hopping points refer to multiple discrete central transmission frequency points pre-configured by the system and distributed within the reverse link communication band. State awareness and interference determination refers to identifying whether the frequency is currently experiencing malicious external interference or adjacent channel leakage by extracting the energy characteristics of each frequency point and comparing them with a decision threshold. A one-dimensional state identifier array is used, where a value of 1 indicates that the corresponding frequency point is not interfered with, and a value of 0 indicates that the corresponding frequency point is interfered with.

[0025] Optionally, the availability status of each candidate frequency hopping point can also be represented as a one-dimensional state flag array of the same length as the total number of candidate frequency hopping points. A value of 0 indicates no interference at the corresponding frequency point, a value of 1 indicates interference at the corresponding frequency point, and a value of 2 indicates that the corresponding frequency point is not detected. The "not detected" state is used to identify frequency points that the system knows are currently occupied by other systems. These frequency points do not participate in interference determination during the spectrum sensing stage and are not included in the available frequency point set during the frequency hopping sequence mapping stage.

[0026] Step 103: Perform adaptive closed-loop update based on the availability status of each candidate frequency hopping point to obtain the set of currently effective available frequency points.

[0027] Specifically, the system performs time-series statistical analysis and aggregation on the availability status obtained from single or multiple spectrum sensing operations to generate an updated set of currently active available frequencies. This set of active available frequencies is dynamically maintained in a buffer on the master station side. By constructing a closed-loop adaptive mechanism, when the external electromagnetic environment dynamically evolves, the system can broadcast the latest generated available frequency information to all end-station devices across the network in real time via the forward channel. This allows the master station and end-stations to seamlessly synchronize the frequency hopping sequence calculation reference pool under the premise of sharing the same external constraints, thereby achieving a dynamic feedback adjustment process of detection, updating, and synchronization.

[0028] Step 104: Based on the time slot resource elements, perform location mapping in the currently effective set of available frequency points to determine the target frequency hopping point corresponding to the current time slot.

[0029] The generation of frequency hopping sequences is deeply coupled with the network's own temporal characteristics. Specifically, the system does not rely on any additional dedicated synchronization security signaling. Instead, it directly inputs the superframe identifier, frame identifier, and time slot identifier from the time slot resource elements into a pre-set composite temporal feature conversion operator to directly calculate a time index that corresponds one-to-one with the current transmission time and evolves monotonically. Subsequently, the time index is input into a pseudo-random sequence generator for fast addressing and permutation to calculate the pseudo-random index value corresponding to the current time slot. Finally, the pseudo-random index value is used as an addressing pointer to perform mapping calculations within the currently effective set of available frequency points, thereby uniquely locking the single transmission frequency used by the current transmission time slot, i.e., obtaining the target frequency hopping frequency point. Since the mapping calculation is only performed within the currently effective set of available frequency points, the generated target frequency hopping frequency point naturally avoids all frequencies marked as being in a state of interference, achieving an intrinsic coupling between the sensing results and sequence generation.

[0030] Step 105: Send the target frequency hopping point and corresponding time slot resource elements to the terminal station through the forward link, and control the terminal station equipment to perform burst frequency hopping communication.

[0031] Specifically, the master station encapsulates the determined target frequency hopping point and corresponding time slot resource elements into a forward link communication continuous stream and sends it down. The end station equipment receives and parses the communication continuous stream to obtain the target frequency hopping point and legal transmission time window corresponding to the current time slot. According to the specified transmission time slot start time, the end station equipment calls its internal RF frequency synthesizer and RF local oscillator module to switch its transmission local oscillator frequency to the physical frequency corresponding to the target frequency hopping point. After completing the local oscillator switching and waiting for the phase-locked loop to lock, the end station equipment up-converts the frequency and modulates and transmits service burst data packets within the data segment of the current time slot. In order to offset the physical time required for hardware local oscillator switching, phase-locked loop frequency stabilization, and power ramp-up, a specific length of protection period is reserved in the time slot structure of burst frequency hopping communication. Each burst time slot strictly corresponds to one hop frequency. By continuously and incrementally sending data packets in each time slot in time, the reverse link exhibits an adaptive frequency hopping communication mode that is strictly constrained by time slot resources and dynamically avoids interference.

[0032] Example 2: Due to the fluctuations in broadband channel noise floor and burst pulse interference faced by the reverse link of satellite communication, this example provides an adaptive spectrum sensing mechanism based on local noise floor modeling and neighborhood linkage determination. In one possible implementation, the following steps are included: Step 201: Divide the broadband spectrum into multiple sub-bands that partially overlap in frequency range.

[0033] refer to Figure 8 Since the frequency band to be detected is relatively wide, in order to take into account the resolution of the spectrum, the wide-band signal cannot be directly analyzed. Therefore, it is necessary to divide it into multiple narrow sub-bands and then perform interference recognition on each sub-band.

[0034] Specifically, the system acquires broadband baseband data through five parallel receiving channels, each with a resampling rate of 491.52 Msps, corresponding to a maximum analyzable bandwidth of 491.52 MHz per channel. The broadband spectrum is divided into five sub-bands in the frequency domain by the five channels, each with an analysis bandwidth of 400 MHz. To eliminate the spectral energy cutoff effect at sub-band boundaries, a 100 MHz frequency overlap region is established between adjacent sub-bands. Data redundancy in the overlap region controls the omission of interference signal characteristics crossing sub-band boundaries.

[0035] Step 202: Perform segmented spectrum estimation processing on the complex sampled data of each sub-band to calculate the power spectral density at each frequency point.

[0036] Specifically, for the complex-sampled data streams of each sub-band, windowed segmented Fast Fourier Transform (FFT) or periodogram method is used for spectral estimation. During the calculation, the acquired data is divided into multiple data segments, and a discrete-time Fourier Transform is performed independently on each segment, with the amplitude squared. The calculation results from multiple data segments are then averaged. The segmented processing clock is 122.88 MHz, and each segment undergoes a 1024-point FFT operation, corresponding to a frequency resolution of 12 kHz.

[0037] Step 203: Construct a corresponding neighborhood window for each frequency point based on the frequency hopping step interval; In this embodiment, the frequency hopping step interval refers to the physical frequency distance between adjacent candidate frequency hopping points configured by the system. For the frequency point currently undergoing interference determination, a frequency interval is constructed by extending half the number of frequency points on both sides of the frequency point as the center frequency. This frequency interval is the neighborhood window of the corresponding frequency point.

[0038] Step 204: Sort the power spectral density within the neighborhood window in ascending order of numerical value, and extract the power spectral density corresponding to the specified percentile as the local noise floor estimate for the current frequency point.

[0039] Because the background noise level of broadband channels exhibits baseline drift across different frequency bands, the power spectral density values ​​of all discrete frequency points within a neighborhood window are extracted and compiled into a one-dimensional dataset. This one-dimensional dataset is then sorted in ascending order using a sorting algorithm. Within the sorted sequence, corresponding values ​​are extracted based on a pre-configured percentile threshold. This pre-configured percentile threshold is set between 70% and 80%, for example, using the 75th percentile. Data values ​​ranking at the 75th percentile are extracted and assigned to the local noise floor estimate for the current frequency point. Through sorting statistics extraction, suspected narrowband interference signals with high energy intensity are stripped away, retaining statistical values ​​that reflect the true noise floor of the current local frequency band.

[0040] like Figure 3 As shown, in step 205, the detection threshold margin is calculated based on the preset target false alarm probability and the chi-square distribution characteristics of the spectrum estimation.

[0041] In segmented spectrum estimation, the power spectrum of noise follows a chi-square distribution, and its degrees of freedom are determined by the number of segments in the segmented averaging. Based on the system's required false alarm tolerance, a preset target false alarm probability is configured. Then, the required threshold bias is calculated using the inverse function of the chi-square distribution. The calculation formula is as follows: δ margin =10×log10(F {chi2_inv} (1-P fa ,2×K) / (2×K)); Where, δmargin For the calculated detection threshold margin, log10 is the logarithm base 10, F {chi2_inv} Let P be the inverse cumulative distribution function of the chi-square distribution. fa denoted as the preset false alarm probability, K is the number of non-overlapping data segments used in the segmented spectrum estimation process, and 2×K is the degree of freedom characterizing the chi-square distribution.

[0042] In this embodiment, the number of segments K is set to 16, and the preset false alarm probability P is... fa Set to 10 - ³. Those skilled in the art can determine the specific values ​​of the above parameters using conventional detection theory design methods, based on the actual system's detection sensitivity requirements and acceptable false alarm costs.

[0043] Step 206: Add the local noise floor estimate to the detection threshold margin to construct the adaptive decision threshold for the current frequency point.

[0044] Obtain the extracted local noise floor estimate and the calculated detection threshold margin. For each frequency point, perform the following calculations: T th(i) =N floor(i) +δ margin ; Among them, T th(i) N is the adaptive decision threshold for the i-th frequency point. floor(i) Let δ be the local noise floor estimate for the i-th frequency point. margin To standardize the calculation of the detection threshold margin, the above operations are represented as linear addition in the logarithmic field.

[0045] Step 207: When the power spectral density of the current frequency point is greater than the adaptive decision threshold, mark the current frequency point as being in an interference state.

[0046] The calculated power spectral density value at the corresponding frequency point is compared with the adaptive decision threshold. If the power spectral density value is greater than the adaptive decision threshold, it is determined that there is radio frequency energy injection exceeding the background noise in the current frequency band, and the status register is updated to mark the current frequency point as interfered. If the power spectral density value is less than or equal to the adaptive decision threshold, it is marked as interference-free.

[0047] Step 208: For any candidate frequency hopping point, detect whether there are any frequency points marked as being interfered with within the frequency hopping step interval centered on it.

[0048] Actual interference signals exhibit energy sidelobe leakage characteristics in the frequency domain. For each member in the pre-configured candidate frequency hopping pool, a continuous frequency review window is formed by extending the frequency hopping step interval by half to both the high and low ends, centered on that candidate frequency hopping point. All physically discrete frequency points contained within this frequency review window are scanned, and their status flags are read one by one.

[0049] Step 209: If it exists, determine that the availability status of the candidate frequency hopping point is unavailable; if it does not exist, determine that the availability status of the candidate frequency hopping point is available.

[0050] During the frequency review process, whenever any frequency point is detected as having an interference-affected status, a logical OR operation is performed to change the availability status of that central candidate frequency hopping point to unavailable. Only when all frequency points within the review window are marked as interference-free will the availability status of the candidate frequency hopping point remain available.

[0051] In some alternative implementations, the frequency hopping step interval can be dynamically configured based on the main lobe bandwidth of the modulated signal for service links employing different modulation and coding schemes. For higher-order modulation signals that occupy a wider bandwidth, the frequency hopping step interval can be increased, thereby expanding the span of the frequency review window and reducing the impact of adjacent channel interference on service demodulation performance.

[0052] In some embodiments, the detection process is triggered by frequency hopping related parameters and the detection range is limited, so that the output result is no longer simply power spectrum information, but a set of frequency availability constraints directly mapped to the frequency hopping frequency space.

[0053] like Figure 10 and 11 As shown, in a given scenario, the steps and time consumption for interference detection are roughly as follows: The data acquisition rate is set at 491.52MHz, divided into 5 sub-bands, each with a bandwidth of 400M.

[0054] The data processing clock is 122.88MHz. Each sub-band processes 16384 data points for a duration of 33μs. Welch spectrum estimation is performed in four segments, with 4096 FFT points and a frequency resolution of 120kHz. The processing time for a single sub-band is approximately 200μs, and the total detection time for a single test is approximately 1ms. The interference result is 4096*5 in length, taking approximately 170μs to upload via the DMA channel.

[0055] Example 3: Due to the presence of impulse noise and intermittent interference in satellite communication channels, the single determination result can easily jump frequently between usable and unusable. This example provides a dual-threshold hysteresis filtering control method to combat intermittent interference. In one possible implementation, it includes the following steps: Step 301: Establish a historical decision window of a specified length for each candidate frequency hopping point.

[0056] Specifically, in the storage medium on the master station side, a fixed-length memory space is allocated to construct a circular buffer for each candidate frequency hopping point in the pre-configured candidate frequency hopping point pool. This circular buffer serves as the historical decision window for the corresponding frequency point. The specified length of this historical decision window corresponds to the system's set superframe period and the period of a single interference detection. Typically, the time span of the historical decision window is constrained to be less than or equal to one complete superframe period.

[0057] In one optional implementation, the system performs 32 spectrum sensing detections within one superframe period, and correspondingly sets the specified length of the historical decision window to 32 detection samples to ensure the temporal correlation of the state data. This historical decision window is used to store the identifiers of multiple consecutive single state decisions in the historical time series.

[0058] Step 302: In each detection period, perform an interference determination on the broadband spectrum once to obtain a single interference determination result for the candidate frequency hopping point, and update the single interference determination result for the candidate frequency hopping point to the corresponding historical decision window in sequence.

[0059] At the end of each detection cycle of the system, the single interference determination result of each candidate frequency hopping point is acquired. Using a sliding window update mechanism, the latest acquired single interference determination result is written into the circular buffer of the corresponding frequency point.

[0060] The write operation adopts a first-in, first-out (FIFO) strategy, that is, the latest judgment data is used to overwrite the oldest judgment data in the buffer, ensuring that a fixed number of the latest sample data are always maintained in the historical judgment window.

[0061] Step 303: Calculate the probability of occurrence of the state being judged as being disturbed within the historical judgment window.

[0062] For each updated historical decision window, read all decision results stored in the buffer. Iterate through the data elements within the window, accumulating the number of samples marked as being in an interfered state. Calculate the proportion based on the statistically obtained count; specifically, the probability P of the current frequency point being in an interfered state. occ =C int / L win ; Among them, C int L represents the total number of samples recorded as being in a disturbed state within the historical decision window. win Specifies the length of the historical judgment window.

[0063] Step 304: Perform a double-threshold hysteresis decision based on the occurrence probability.

[0064] Specifically, the probability of occurrence calculated in the above steps is obtained and compared with the pre-configured dual threshold parameters in the system. The dual threshold parameters include a higher upper limit decision boundary and a lower lower limit decision boundary, which form a no-operation hysteresis interval. Depending on the interval in which the calculated value falls, the system enters different conditional branches to execute subsequent actions.

[0065] Under the first judgment condition, when the probability of occurrence is greater than or equal to the preset high threshold, the availability status of the candidate frequency hopping point is marked as unavailable.

[0066] Specifically, when the duty cycle of intermittent interference increases, causing a denser appearance of interfering samples within the window, the calculated probability of occurrence will rise. The system then compares this probability with a preset high threshold.

[0067] In one scenario, the historical decision window length is normalized to 100 detection samples, and the preset high threshold normalization is 0.7. When the cumulative number of interference samples within the window reaches 70 or more, the probability of occurrence is greater than or equal to 0.7. At this time, the system triggers a judgment action, determining that the frequency band is continuously occupied or that there is high-frequency pulse interference, thereby modifying the status register and flipping the availability status of the candidate frequency hopping point from available to unavailable.

[0068] Under the second determination condition, when the probability of occurrence is less than or equal to a preset low threshold, the availability status of the candidate frequency hopping point is restored to available.

[0069] Specifically, when the interference source stops transmitting or leaves the coverage area, the number of continuously detected interference-free samples increases, and the probability of occurrence decreases accordingly. A preset low threshold normalization characterization is set to 0.2. Only when the cumulative number of interference samples within the window drops to 20 or less, i.e., the probability of occurrence is less than or equal to 0.2, does the system determine that the channel environment has returned to stability, and then performs a recovery action, re-marking the availability status of the candidate frequency hopping point as available, allowing it to re-participate in the mapping of frequency hopping sequences.

[0070] Under the third determination condition, when the probability of occurrence is between the preset high threshold and the preset low threshold, the availability status of the candidate frequency hopping point remains unchanged.

[0071] Specifically, when the probability of occurrence falls within the hysteresis range between 0.2 and 0.7, the current interference is determined to be transient or in a transitional period of interference decay. At this time, the system employs state maintenance logic, ignoring the flip-flop disturbance of a single detection result and maintaining the availability status data of that frequency point from the previous processing cycle. By setting a high-low threshold difference, the system isolates the frequency point availability status oscillations caused by a single or a few false detections, reducing the signaling overhead during the closed-loop update process.

[0072] In some alternative implementations, multiple sets of different high and low threshold pairs can be configured for end-station devices with different service priorities.

[0073] For high-reliability control signaling links, the preset high threshold and the preset low threshold are lowered to improve sensitivity to interference and extend the observation period for frequency recovery. For ordinary broadband data links, the preset high threshold should be appropriately increased to improve the capacity utilization of the candidate frequency hopping pool.

[0074] In some embodiments, the neighborhood determination method based on frequency hopping step interval means that for any candidate frequency hopping point, if there is any interfering frequency point in its corresponding frequency neighborhood, the candidate frequency point is determined to be unusable. This achieves an effective mapping of the continuous spectrum to the discrete frequency hopping point space, avoids the problem of missed interference detection at the frequency point boundary, and enables the interference detection results to directly participate in the frequency hopping sequence generation and screening process.

[0075] Example 4: Due to the long propagation delay of hundreds of milliseconds in geosynchronous orbit satellite communication, the conventional instantaneous activation mechanism will cause the frequency hopping sequence at both ends of the communication to be misaligned. Therefore, a double-buffered superframe alignment and seamless switching strategy that is compatible with the long propagation delay between satellite and ground is provided.

[0076] In one possible implementation, the following steps are included: like Figure 4 As shown, in step 401, the currently active buffer and the preset buffer are maintained synchronously.

[0077] Specifically, two identical memory spaces are allocated in the control and processing unit of the main station equipment as a double-buffered storage entity.

[0078] One of the memory spaces is set as the currently active buffer, used to latch the bitmap of the available frequency points currently in use; Another memory space is set as a preset buffer to receive and store the status bitmap of available frequency points to be used in the next scheduling cycle.

[0079] By using two physically or logically isolated storage structures, the data read operation being performed can be decoupled from the data update operation in the background.

[0080] Step 402: Using the set of currently active available frequency points stored in the currently active buffer, perform frequency hopping sequence generation for the current superframe period to obtain the target frequency hopping frequency points corresponding to each time slot.

[0081] During any network scheduling cycle, the frequency hopping sequence generator uses the currently active buffer as the only valid addressing data source.

[0082] After reading the time slot resource elements that it is authorized to send, the terminal equipment directly maps them to the currently effective buffer to extract the legal physical frequency parameters, and then controls the radio frequency local oscillator module to execute the corresponding frequency point burst frequency hopping communication.

[0083] Before the end of the current superframe cycle, the data in this buffer is set to read-only by the system, thereby ensuring the static stability of the frequency hopping pool within the current cycle.

[0084] Step 403: Within the current superframe period, the master station writes the availability status of each candidate frequency hopping point after comprehensive judgment into the preset buffer.

[0085] While the system maintains the communication link using the currently active buffer, the spectrum sensing module and filtering module on the master station side continue to run in the background.

[0086] The latest availability status data streams of each candidate frequency hopping point, after undergoing dual-threshold hysteresis decision stabilization processing, are directed to be written into a preset buffer on the main station side. This writing action updates the status bitmap in the preset buffer in real time without causing any data overwriting interference to the currently active buffer that is undertaking the foreground communication guidance task.

[0087] Step 404: During the reserved time period before the end of the current superframe cycle, the master station determines the effective superframe number corresponding to the next update and broadcasts the contents of the preset buffer and the effective superframe number to each end station through the forward link.

[0088] In this step, the master station extracts the set of available frequency points that have been updated in the preset buffer, packages them together with the target effective superframe number into a signaling control frame, and sends it down through the forward broadcast carrier.

[0089] To ensure data delivery under long propagation delays, the system has established constraint rules regarding reserved time periods. Specifically: T pre >=T prop +T proc ; Among them, T pre To set a reserved time period before the end of the current superframe cycle, T prop For the one-way propagation delay of the satellite link, T procThis is to reduce the delay in signaling parsing and baseband processing for terminal station equipment.

[0090] Due to the physical limitations of geostationary orbit satellites, the one-way propagation delay can be set to 270 milliseconds, and the terminal station processing delay can be set to 50 milliseconds. The total time required for system loop closure is calculated to be 320 milliseconds. The system sets the superframe period to 600 milliseconds and configures the reserved time period to 400 milliseconds.

[0091] Specifically, 400 milliseconds before the end of the current superframe period, the master station triggers a broadcast action, ensuring that all end stations have sufficient physical time to receive and parse data packets containing the frequency hopping parameters for the next superframe before the current superframe ends. After receiving the data, the end station loads it into a local preset buffer and puts it in a standby state.

[0092] Step 405: When the system time reaches the superframe start boundary corresponding to the effective superframe number, the master station and each end station synchronously trigger buffer switching, and copy the contents of the preset buffer to the currently effective buffer to absorb the satellite-to-ground propagation delay.

[0093] Specifically, the master station and each end station device maintain a high-precision time synchronizer using the network clock reference field. The control processing unit of each node continuously monitors the timestamp. When the parsed network system time is strictly aligned with the physical time boundary corresponding to the received effective superframe number, a hardware trigger pulse is generated.

[0094] The trigger pulse drives the storage controllers inside the master station and each end station to copy the status bitmap data in the preset buffer to the currently active buffer, or directly swap the read and write pointer identifiers of the dual buffers.

[0095] Through this synchronous triggering mechanism, the effective delay of the sensing results is constrained to a constant one whole superframe period, and the uncertain propagation delay between the satellite and the ground is completely absorbed by the superframe boundary, realizing seamless switching of global closed-loop parameters.

[0096] like Figure 2 As shown, in step 406, the total number of available frequency points in the updated set of currently active available frequency points is counted; After completing the buffer switching or background update operation, the system performs numerical statistics on the one-dimensional status identifier array corresponding to the currently effective set of available frequency points, accumulates the number of data bits with values ​​representing available status, and calculates the total number of physical frequency points that can actually be used for frequency hopping mapping in the current network.

[0097] Step 407: When the total number of available frequency points is less than the preset minimum available frequency point threshold, based on the power spectral density of each frequency point obtained from state awareness and interference determination, the control system exits the frequency hopping communication mode and switches to fixed-frequency time-division multiple access communication by selecting a single frequency point with the lowest power spectral density. When the total number of available frequency points is greater than or equal to the preset minimum available frequency point threshold, the adaptive frequency hopping communication mode continues to operate.

[0098] After obtaining the total number of available frequency points, a unified conditional judgment node is introduced to compare this total number with a preset minimum available frequency point threshold. The preset minimum available frequency point threshold is set based on ensuring that frequency hopping points do not experience duplicate reuse conflicts within a superframe period. The system executes corresponding logical path routing based on the comparison result.

[0099] In the first execution path, when the total number of available frequency points is less than the preset minimum available frequency point threshold, the system determines that the current electromagnetic environment has deteriorated and the available spectrum resources are insufficient to support the anti-collision and anti-interception requirements of random frequency hopping.

[0100] Specifically, the master station issues control signaling, requiring the entire network to back off and exit frequency-hopping communication mode. The control system searches the historical decision window or the current power spectral density matrix to extract the single physical frequency point with the lowest relative background noise or that is not affected by high-power interference. The system instructs all end stations to lock their RF local oscillators to this single physical frequency point and switch to performing conventional fixed-frequency time-division multiple access communication to maintain the connectivity of the basic link.

[0101] In the second execution path, when the total number of available frequency points is greater than or equal to the preset minimum available frequency point threshold, the system determines that the currently available resources are within the safety index range.

[0102] Specifically, the system continues to use the double-buffered superframe alignment and seamless switching strategy to keep the adaptive frequency hopping communication mode running continuously according to the predetermined timing.

[0103] In some scenarios, the closed-loop adaptive update mechanism of the frequency hopping sequence includes: generating an initial frequency hopping sequence based on the superframe structure and time slot parameters; obtaining a set of frequency point availability constraints through the spectrum sensing module and modifying the initial frequency hopping sequence by removing or replacing interfered frequency points; forming updated frequency hopping parameters and broadcasting or synchronizing them, so that the master station and the terminal station can complete the reconstruction and consistent update of the frequency hopping sequence under the condition of shared parameters; and realizing the adaptive evolution process of the frequency hopping sequence as the spectrum environment changes dynamically.

[0104] Example 5: In one possible implementation, based on time slot resource elements, location mapping is performed in the currently effective set of available frequency points to determine the target frequency hopping point corresponding to the current time slot, including the following steps: Step 501: Merge the superframe identifier, frame identifier and time slot identifier in the time slot resource elements into a unified time index.

[0105] Specifically, the system receives the time slot resource elements from the communication protocol stack parsing output and extracts the time-level parameters to establish mapping constraints between time series and frequency sequences.

[0106] The current superframe identifier, frame identifier, and time slot identifier are concatenated and used as input to the hash module. The three-dimensional network time coordinates are uniformly used as parameters, so that each communication time slot is assigned a unique numerical index.

[0107] Step 502: Obtain the pre-configured frequency hopping seed and construct a composite key based on the pre-configured frequency hopping seed and time slot resource elements.

[0108] Specifically, the pre-configured frequency hopping seed and time slot resource elements are read from non-volatile memory.

[0109] Step 503: Based on the composite key and time index, calculate the pseudo-random index value corresponding to the current time slot using the sequence direct addressing algorithm.

[0110] Specifically, a pseudo-random sequence generator is constructed using a linear feedback shift register. The calculated composite key is loaded into the linear feedback shift register as the initial state. The bit width and feedback polynomial of the linear feedback shift register can be selected by those skilled in the art based on the period length of the desired pseudo-random sequence.

[0111] The LFSR bit width should satisfy the requirement that its maximum period length is not less than the total number of time slots within the validity period of a superframe packet number in the system. The feedback polynomial is selected as the primitive polynomial corresponding to the bit width to ensure the longest period. The lookup table for the primitive polynomial is a well-known technical document in this field.

[0112] To avoid the computational delay caused by the conventional stepwise recursive algorithm when crossing long time slot intervals, the system calls the sequence direct addressing algorithm based on matrix fast exponentiation.

[0113] This algorithm transforms the state transition process of the shift register into a power operation of the state transition matrix, determines the power based on the time index, and then directly calculates the register output value at the corresponding time node.

[0114] The output value is extracted as the pseudo-random index value corresponding to the current time slot. This algorithm decouples the timing dependencies between time slots, allowing the system to independently reconstruct frequency hopping parameters in any specified time slot.

[0115] Step 504: Based on the pseudo-random index value, perform position mapping in the currently effective set of available frequency points to obtain the target frequency hopping point.

[0116] The pseudo-random index values ​​calculated in the above steps are distributed within a discrete space with fixed numerical boundaries. Simultaneously, the currently active set of available frequencies constitutes another dynamically changing physical frequency pool. By establishing a mapping channel between the two, the system transforms dimensionless random values ​​into specific radio frequency transmission frequencies.

[0117] Step 505: Determine the total number of available frequency points in the currently active set of available frequency points.

[0118] The system reads the available frequency point status bitmap latched in the currently active buffer of the double-buffered control entity. It iterates through this one-dimensional status identifier array, performs an accumulation and statistical operation, calculates the number of available feature bits for each status value, and outputs this statistical count as the total number of available frequency points. This total number reflects the scale of available physical channels that the network can use to carry services in a specific time slot.

[0119] Step 506: Use the pseudo-random index value to perform a modulo operation on the total number of available frequency points to obtain the mapping index.

[0120] Obtain the pseudo-random index value calculated in the above steps and the total number of available frequency points, and perform a division and remainder operation to limit the addressing range.

[0121] The calculated mapping index k index =R val MODM total ; Among them, R val The input is a pseudo-random index value to the mapping module, MOD is the integer modulo operator, and M is the input. total This represents the total number of available frequency points. A modulo operation is used to generate a non-negative integer whose value is limited to zero and the total number of available frequency points minus one.

[0122] Step 507: Select the corresponding frequency point as the target frequency hopping point from the set of currently active available frequency points arranged in ascending order of frequency according to the mapping index.

[0123] Extract all physical frequency points marked as available from the currently active set of available frequency points, sort them in ascending order from low to high according to the physical frequency value of each frequency point, and construct an available frequency point lookup table for the current scheduling period.

[0124] The calculated mapping index is used as the row pointer of the lookup table to directly address and retrieve the physical frequency value stored at the corresponding location. The retrieved physical frequency value is established as the target frequency hopping point corresponding to the current time slot, and then output to the baseband and RF control unit for subsequent local oscillator switching operations.

[0125] In some alternative implementations, for business scenarios with high requirements for frequency uniformity, a first-order nonlinear perturbation function can be added before the pseudo-random index value enters the modulo operation unit. The local clustering characteristics that may exist in the original sequence can be broken by using operations such as hash confusion, thereby controlling the dispersion variance of the frequency hopping sequence in the frequency domain after modulo mapping.

[0126] A brief overview of the frequency domain analysis process can be found in [reference needed]. Figure 9 .

[0127] Example 6 provides an adaptive frequency hopping anti-avalanche scheme based on consistent hash ring remapping. The specific difference from Example 5 is that when the available frequency point set is dynamically added to or deleted from, it avoids remapping and reorganizing the entire sequence due to conventional modulo operations. The remaining steps are the same as in Example 5, and can be referred to the foregoing description, so they will not be repeated here.

[0128] In one possible implementation, the following steps are included: Step 601: Obtain the pre-constructed hash ring space, which maps virtual nodes corresponding to all candidate frequency hopping points.

[0129] Specifically, a closed linear address space with a numerical range of 0 to 4294967295 is established to form a hash ring space. All candidate frequency hopping points preset by the system are extracted, and at this stage, regardless of their current availability status, they all participate in the infrastructure mapping.

[0130] To avoid uneven distribution of frequency points in the address space due to a single hash mapping, a virtual node expansion mechanism is introduced. Multiple virtual nodes are generated for a single candidate frequency hopping point. The specific mapping operation logic is as follows: pos jv =Hash(f j ||v); Where, pos jv Here, f represents the fixed mapping position of the virtual node on the hash ring space, where Hash is the preset one-way hash algorithm function. j is the physical frequency value identifier of the j-th candidate frequency hopping point, || is the data bit concatenation operator, and v is the virtual node index number assigned to this candidate frequency hopping point.

[0131] Furthermore, the virtual node index number is set to a range of 0 to V-1, where V is the total number of virtual nodes allocated to each candidate frequency hopping point. The standard deviation of the discrete distribution of virtual nodes in the hash ring space is constrained by statistical uniformity; increasing the number of virtual nodes can improve the uniformity of frequency distribution in the hash ring space. By pre-constructing this hash ring space offline, the physical locations of all virtual nodes are in a static and fixed state, and their locations do not migrate with the dynamic updates of the available frequency set during network scheduling.

[0132] In one optional implementation, the total number of virtual nodes V allocated to each candidate frequency hopping point is set to 150. Those skilled in the art can determine the appropriate number of virtual nodes by evaluating the statistical variance of the node spacing in the hash ring space, based on the size of the candidate frequency hopping point pool and the requirement for mapping uniformity. Generally, a larger total number of virtual nodes results in more uniform coverage of each frequency point in the hash ring space, but correspondingly increases the storage overhead for node lookup.

[0133] Step 602: Using the pseudo-random index value as the starting point for the query, perform node lookup along the selected direction of the hash ring space.

[0134] Obtain the pseudo-random index value corresponding to the current time slot output by the preceding operation module. Treat this pseudo-random index value as an address request pointer on the hash ring space.

[0135] Starting from the value of the pseudo-random index, a traversal search operation is performed along the direction of numerical increment. This direction of numerical increment is the selected direction, which manifests as a unidirectional clockwise numerical increment search within the closed circular space.

[0136] Step 603: The candidate frequency hopping point corresponding to the first virtual node found along the selected direction that belongs to the currently effective set of available frequency points is determined as the target frequency hopping point.

[0137] During the step-by-step search along the selected direction, the system reads the candidate frequency hopping point identifiers bound to the virtual nodes along the path one by one. It then calls the available frequency point status bitmap latched in the currently active buffer of the double-buffered control entity to extract the current availability status of the candidate frequency hopping point.

[0138] If the state bitmap records that the frequency point is in an interference state, the system ignores the currently read virtual node and continues to step and compare in the original direction.

[0139] If the status bitmap records the frequency point as available, the system stops the search loop, confirms the currently locked virtual node as the matching node, and extracts the physical value of the candidate frequency hopping point in its reverse mapping. This physical value is then configured as the target frequency hopping point for the current time slot and output to the next-level RF control link.

[0140] Using the above address space node lookup mechanism, when the electromagnetic environment evolves and causes a specific frequency point to change from an available state to an unavailable state, only some frequency hopping time slots that were originally hashed and mapped to the virtual node corresponding to that frequency point need to find a successor node again. The proportion of frequency hopping time slots affected by remapping converges to the reciprocal of the total number of available frequency points.

[0141] The mapping paths of the remaining frequency hopping time slots remain unchanged, blocking the avalanche effect of full sequence rearrangement caused by the dynamic update process of the available frequency point set, and maintaining the anti-interception statistical properties of frequency hopping communication on the macroscopic time axis.

[0142] In some optional implementations, considering that a large number of virtual nodes are mapped on the hash ring space, to control the baseband processing latency occupied by the node lookup operation, the position values ​​of all available virtual nodes can be extracted within the terminal station to form an index array arranged in ascending order. After receiving the pseudo-random index value, a binary search algorithm is called to determine the position of the successor node corresponding to the starting point. By adopting this lookup operation logic, the number of node comparisons can be controlled to be less than 15, compressing the time consumption of time slot parameter mapping to the microsecond level, thereby meeting the timing requirements of the system's superframe scheduling.

[0143] Based on the above implementation method, the functionality of the method of the present invention was verified on a simulation platform. The simulation conditions were set as follows: a one-way delay of 270ms for the GEO satellite link, 200 candidate frequency hopping points configured within a reverse link bandwidth of 1500MHz, an overframe period of 600ms, and the system operating in a channel environment with a signal-to-noise ratio of 3dB to 15dB, with different amounts of narrowband interference signals injected. The verification results show that, under the condition of 30% frequency point interference, the method of the present invention can complete the synchronous update of the available frequency point set between the master station and the end station within one overframe period, and the frequency hopping sequence remains consistent among all end stations in the network; the dual-threshold hysteresis filtering mechanism reduces the number of frequency point state flips compared to the single-threshold determination method, and no signaling overflow phenomenon caused by frequent state oscillations occurs during system operation.

[0144] The above verification results demonstrate that the method of the present invention can effectively improve the anti-interference capability and system operational stability of satellite communication links in complex electromagnetic environments. Those skilled in the art will understand that specific performance indicators may vary depending on satellite orbital parameters, interference signal characteristics, and system configuration.

[0145] This invention no longer uses time slots only as a transmission scheduling tool, but introduces time slot number and superframe structure as the core input constraints for frequency hopping sequence generation, so that the frequency hopping frequency point is jointly determined by parameters such as superframe number, time slot number, superframe group and frequency hopping seed, that is, frequency hopping frequency point = f(superframe number, time slot number, superframe group, frequency hopping seed).

[0146] Under this mechanism, different time slots correspond to different frequency points, and the frequency hopping sequence changes dynamically with the superframe period. The master station and the terminal station do not need to be explicitly synchronized on a frequency-by-frequency-point basis. They only need to share the above parameters to reconstruct a consistent frequency hopping sequence, so that the frequency hopping process has a structured feature that combines determinism and randomness.

[0147] The forward link maintains continuous stream transmission and carries key information such as time slot plan, superframe structure and NCR time base in the continuous stream. The end station realizes time alignment, parameter synchronization and frequency hopping sequence reconstruction by parsing the forward continuous stream, thus forming a system-level coupling relationship driven by forward broadcast and frequency hopping performed by the reverse link.

[0148] The transmission behavior of the terminal station is determined by the dual constraints of time and frequency. The time slot plan is used to determine the transmission time, and the frequency hopping sequence is used to determine the transmission frequency. Both share the same input parameter system, thus ensuring that the transmission time and transmission frequency are naturally aligned during the generation stage without relying on post-processing calibration.

Claims

1. A satellite communication adaptive frequency hopping method based on time slot constraints and spectrum sensing coupling, characterized in that, include: Obtain time slot resource elements generated by network timing scheduling on the main station side; The broadband spectrum covered by the reverse link is collected, and the broadband spectrum is subjected to state perception and interference judgment to obtain the availability status of each candidate frequency hopping point. Adaptive closed-loop update is performed based on the availability status of each candidate frequency hopping point to obtain the currently effective set of available frequency points; Based on time slot resource elements, perform location mapping in the currently effective set of available frequency points to determine the target frequency hopping point corresponding to the current time slot; The target frequency hopping point and corresponding time slot resource elements are sent to the terminal station through the forward link, and the terminal station equipment is controlled to perform burst frequency hopping communication.

2. The method according to claim 1, characterized in that, Based on time slot resource elements, location mapping is performed within the currently active set of available frequency points to determine the target frequency hopping point corresponding to the current time slot, including: The superframe identifier, frame identifier, and time slot identifier in the time slot resource elements are merged into a unified time index.

3. The method according to claim 2, characterized in that, Determining the target frequency hopping point corresponding to the current time slot also includes: Obtain the pre-configured frequency hopping seed and construct a composite key based on the pre-configured frequency hopping seed; Based on the composite key and time index, the pseudo-random index value corresponding to the current time slot is calculated using the sequence direct addressing algorithm.

4. The method according to claim 3, characterized in that, Determining the target frequency hopping point corresponding to the current time slot also includes: Based on the pseudo-random index value, position mapping is performed in the currently effective set of available frequency points to obtain the target frequency hopping point.

5. The method according to claim 4, characterized in that, Based on pseudo-random index values, location mapping is performed within the currently effective set of available frequency points to obtain the target frequency hopping point, specifically including: Determine the total number of available frequency points in the currently active set of available frequency points; The mapping index is obtained by performing a modulo operation on the total number of available frequency points using the pseudo-random index value; Based on the mapping index, select the corresponding frequency point as the target frequency hopping point from the currently active and available frequency points sorted in ascending order of frequency.

6. The method according to claim 1, characterized in that, State awareness and interference determination of the broadband spectrum, including: The broadband spectrum is divided into multiple sub-bands that locally overlap in frequency range; The power spectral density at each frequency point is calculated by performing segmented spectral estimation on the complex sampled data of each sub-band.

7. The method according to claim 6, characterized in that, State awareness and interference determination of the broadband spectrum also include: Based on the preset frequency hopping step interval, a corresponding neighborhood window is constructed for each frequency point; The power spectral density within the neighborhood window is sorted in ascending order of numerical value, and the power spectral density corresponding to the specified percentile is extracted as the local noise floor estimate for the current frequency point.

8. The method according to claim 7, characterized in that, State awareness and interference determination of the broadband spectrum also include: The detection threshold margin is calculated based on the preset target false alarm probability and the chi-square distribution characteristics of the spectrum estimation. The local noise floor estimate is added to the detection threshold margin to construct the adaptive decision threshold for the current frequency point; When the power spectral density of the current frequency point is greater than the adaptive decision threshold, the current frequency point is marked as being interfered with.

9. The method according to claim 8, characterized in that, The availability status of each candidate frequency hopping point is obtained, specifically including: For any candidate frequency hopping point, detect whether there are any frequency points marked as being interfered with within the frequency hopping step interval centered on it; If it exists, the availability status of the candidate frequency hopping point is determined to be unavailable.

10. The method according to claim 1, characterized in that, Adaptive closed-loop updates are performed based on the availability status of each candidate frequency hopping point. Specifically, a double-buffered architecture is used for timing alignment, including: Simultaneously maintain the currently active buffer and the preset buffer; Using the set of currently active available frequency points stored in the currently active buffer, perform frequency hopping sequence generation for the current superframe period.

Images