Missile-borne navigation receiver rotation compensation method

Abstract

The invention discloses a missile-borne navigation receiver rotation compensation method, and relates to the technical field of rotation compensation. The method comprises the following steps: acquiring a complex intermediate frequency satellite signal sequence of a missile-borne navigation receiver, extracting an instantaneous total power sequence through short-time Fourier transform, identifying a rotation period according to the instantaneous total power sequence, and dividing an availability window and a sensitive window; constructing a double-layer rotation frequency spectrum model based on the divided windows, and carrying out online calibration on the double-layer rotation frequency spectrum model; quantizing the monitored abnormal signal into a rotation consistency and interference possibility score by taking the calibrated model as a reference, and determining a discrimination result of the abnormal signal; and performing rotation compensation on the receiver according to a judgment result to generate continuous navigation measurement information. According to the invention, rotation compensation of the missile-borne navigation receiver is realized based on continuous navigation measurement information.

Description

Technical Field

[0001] This invention relates to the field of rotation compensation technology, specifically a rotation compensation method for a missile-borne navigation receiver. Background Technology

[0002] In the field of missile-borne precision guidance, navigation receivers need to operate stably in the extreme environment of high-speed missile rotation. Due to the missile's spin, the pointing of its onboard navigation antenna changes continuously, causing the received satellite navigation signal to exhibit periodic deep attenuation and rapid recovery. This signal amplitude modulation with a definite period, caused by the physical laws of rotation itself, is easily misinterpreted by the receiver as random, malicious external electromagnetic interference or channel fading. This triggers unnecessary anti-interference suppression or link reacquisition procedures, directly disrupting the continuity of navigation measurements and representing a key bottleneck restricting high-precision navigation of rotating missiles.

[0003] Currently, a typical processing method in the industry is simple gain compensation based on preset parameters. This method first estimates the approximate attenuation pattern of the signal during the rotation period through experiments or a priori knowledge, and then sets a fixed gain compensation curve in the receiver's digital signal processing link to boost the signal amplitude when the signal is weak.

[0004] However, the simple gain compensation method based on preset parameters mentioned above, with its fixed gain compensation curve, cannot accurately track and model the dynamic changes such as rotation rate and attitude coupling caused by environmental and load changes during the actual flight of the missile. As a result, when the actual rotation state deviates from the preset law, the fixed gain compensation curve will quickly become mismatched, resulting in insufficient or excessive compensation. This not only fails to stabilize the signal but also introduces additional distortion, which significantly degrades the receiver's performance in complex dynamic flight profiles and makes it difficult to guarantee navigation reliability. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a rotation compensation method for missile-borne navigation receivers, thus resolving the problems in the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for rotation compensation of an airborne navigation receiver, comprising the following steps: Step S1: Obtain the complex intermediate frequency satellite signal sequence of the missile-borne navigation receiver, analyze the complex intermediate frequency satellite signal sequence using the short time Fourier transform method, calculate the instantaneous total power sequence, identify the rotation period of the missile-borne navigation receiver based on the instantaneous total power sequence, and divide the availability window and the sensitive window. Step S2: Based on the availability window and sensitivity window, construct a two-layer rotating spectrum model; calibrate the two-layer rotating spectrum model to obtain a calibrated two-layer rotating spectrum model; Step S3: Using the calibrated rotating spectrum model as a reference, the abnormal signals detected by the missile-borne navigation receiver are converted into observation vectors, and the rotational consistency score and interference probability score are calculated; based on the rotational consistency score and interference probability score, the discrimination result of the abnormal signal is determined. Step S4: Based on the discrimination result of the abnormal signal, perform rotation compensation on the missile-borne navigation receiver to generate continuous navigation measurement information.

[0007] Preferably, acquiring the complex intermediate frequency satellite signal sequence of the missile-borne navigation receiver includes: After the missile-borne navigation receiver enters operational mode, it begins acquiring satellite signals. The receiver performs down-conversion and analog-to-digital conversion on the radio frequency satellite signals received by the antenna, outputting a complex intermediate frequency (IF) satellite signal sequence. This sequence is mathematically represented as... , where characters This represents a complex intermediate frequency signal; the character 'n' represents the discrete-time index, with a value range of... N, where N is the total number of discrete time points in the complex intermediate frequency satellite signal sequence; character It is an index used to distinguish different signal processing channels.

[0008] Preferably, the instantaneous total power sequence is calculated by analyzing the complex intermediate frequency satellite signal sequence using the short-time Fourier transform method, including: For complex intermediate frequency satellite signal sequences Time-frequency analysis was performed using the short-time Fourier transform method: the long-time sequence was divided into overlapping short time segments, and a discrete Fourier transform was performed on each time frame after windowing. For the _th_ time segment... Each time frame, and its corresponding complex spectrum It is obtained by calculation using the following formula: ; in, Indicates channel The complex spectral value at the m-th time frame and the k-th frequency unit; the character m identifies the sequence number of the time frame; k identifies the sequence number of the frequency unit, k=0,1,…,N-1; This indicates a summation operation performed on L time points within the window; It is a passage In the complex intermediate frequency signal sequence, the sample value corresponds to the r-th point of the m-th frame; the character H is the frame shift length, that is, the time interval between the start points of adjacent time frames; It is a window function of length L; character is the kernel function of the Discrete Fourier Transform, j is the imaginary unit; character N is the length of the Discrete Fourier Transform; Next, by calculating the square of the modulus of each complex spectral value, a time-frequency energy distribution matrix representing the signal power distribution in the two-dimensional time-frequency plane is obtained. Its elements are ; The time-frequency energy distribution matrix of each channel Summing along the frequency dimension k generates a one-dimensional instantaneous total power sequence. .

[0009] Preferably, based on the instantaneous total power sequence, the rotation period of the missile-borne navigation receiver is identified, and the availability window and sensitive window are divided, including: In the generated instantaneous total power sequence In this study, by performing periodogram analysis or autocorrelation function peak detection on the sequence, the basic rotation period with regular fluctuations in energy distribution is identified. Its unit is time frames; After identifying the rotation period Based on this, within a complete rotation cycle, the power sequence of the main channel is... The morphology is analyzed to divide feature windows. To ensure the robustness of the analysis, the basis for this is... Periodic samples should satisfy the following: the periodicity of power fluctuations and the identified fundamental rotation period. The data meets the criteria and shows no obvious sudden distortion, thus dividing the data into two distinct sub-intervals based on a fixed power threshold. This threshold was determined through experimental statistics; Availability Window Corresponding to Values ​​consistently above the threshold The fluctuations are relatively small, and within this window, the carrier phase continuity can be reliably maintained; Sensitive window Corresponding to The value dropped to the threshold The following periods may see sharp fluctuations; within this window, the signal is prone to significant attenuation and phase abrupt changes.

[0010] Preferably, based on the availability window and the sensitivity window, constructing a two-layer rotating spectrum model includes: The fundamental part of constructing the two-layer rotating spectrum model is the periodic layer. This periodic layer aims to describe the stable and predictable repetitive modulation structure of the signal energy within one rotation cycle using a set of core parameters. Based on the output of step S1, the parameter vector of the periodic layer is initialized. ; Secondly, the dynamic part of the two-layer rotating spectrum model—the perturbation layer—is constructed. The perturbation layer is mediated by a state vector. At discrete time Characterization: ; in, Indicates at time The state vector of the perturbation layer; Indicates at time The actual starting phase of the sensitive window relative to the reference phase of the periodic layer The instantaneous offset; Indicates at time The mean measured power within the availability window relative to the baseline mean of the periodic layer. The instantaneous offset; Indicates at time The measured attenuation slope within the sensitive window relative to the periodic layer reference parameter The instantaneous offset.

[0011] Preferably, the two-layer rotating spectrum model is calibrated to obtain a calibrated two-layer rotating spectrum model, including: Kalman filtering algorithm is used to analyze the state of the perturbation layer. Online recursive estimation is performed, and the calibration process is carried out at each processing time. The core of execution is to construct and derive the state prediction equation and the observation equation; The Kalman filter recursively performs prediction and update steps, ultimately outputting the optimal unbiased estimate of the state of the perturbation layer at the current moment. And its estimated error covariance matrix.

[0012] Preferably, based on a calibrated rotating spectrum model, the abnormal signals detected by the missile-borne navigation receiver are converted into observation vectors, including: The missile-borne navigation receiver compares the instantaneous total power in real time. Compared with the periodic layer reference power in step S2 The difference is used to continuously monitor signal quality, and when the conditions are met... When this occurs, a "significant degradation" is determined and the exception handling process is triggered. The preset power drop threshold; Subsequently, the anomalous time window was determined and the observation vector was constructed. The duration of the anomalous period was defined as the period from when the power drop fell below a threshold. Start, until power recovers to the threshold. Within a stable, continuous time window, the receiver performs parallel computations and extracts five key parameters, which together form a structured observation vector. This vector is calculated within the time window of the abnormal event, and its specific form is as follows: ; Where O represents the observation vector; Represents the current instantaneous total power Relative to the periodic layer reference power The magnitude of the decline; This represents the instantaneous change in total power within the time window of the abnormal event. The variance; This represents the difference between the actual starting phase of the current sensitive window estimated based on real-time signal analysis and the current phase estimate predicted by the two-layer rotating spectrum model. The difference between them; This represents the average signal power measured within the availability window at the current moment, compared to the current power reference value given by the two-layer rotated spectrum model. The difference between them; The relative entropy between the measured spectral energy distribution during the anomalous signal period and the unperturbed ideal spectral energy distribution predicted by the two-layer rotating spectral model at the current phase, where T is the transpose.

[0013] Preferably, the calculation of rotational consistency score and disturbance probability score includes: Calculate rotational consistency score : ; in, This represents the rotational consistency score, and its range is... ; Represents the natural exponential function; It is the phase tolerance, set based on three times the standard deviation of the historical window position estimation error; This is the expected power attenuation value; It is the tolerance parameter for power matching; It is a relative entropy saturation threshold; Next, the interference probability scores are calculated in parallel. : ; in, The interference probability score is represented by the range of values. ; Represents the hyperbolic tangent function; It is the normalized benchmark for power fluctuation variance, determined based on experiments of typical fluctuation levels under rotational modulation; It is a normalized benchmark for the power mean offset.

[0014] Preferably, the determination of the abnormal signal discrimination result based on the rotational consistency score and the interference probability score includes: Based on the calculated rotational consistency score And interference probability score The score is used for collaborative discrimination based on preset decision rules, and a definite category result R is output: ; in, Indicates the judgment result of the abnormal signal; It is a preset decision threshold. Mapping a continuous score space to a discrete, well-defined decision category.

[0015] Preferably, based on the discrimination result of the abnormal signal, rotation compensation is performed on the missile-borne navigation receiver to generate continuous navigation measurement information, including: Based on the discrimination result R of the output abnormal signal, corresponding signal processing strategies are adopted for the missile-borne navigation receiver to compensate for the rotation modulation effect and suppress interference, ultimately generating continuous navigation measurement information including: Carrier phase observation The tracking loop, after strategy adjustment, outputs directly, and this quantity remains continuous throughout the entire rotation cycle without interruption. Pseudodistance observation Calculated by the following formula: ; in, At the speed of light, For the local time of the missile-borne navigation receiver, The satellite launch time is calculated from the navigation message; Measurement status indicator This is used to attach a metadata indicator to each output observation, indicating the processing status at the time of its generation.

[0016] This invention provides a rotation compensation method for missile-borne navigation receivers, which involves machine learning and deep learning technologies, and has the following beneficial effects: (1) By converting the one-dimensional time-domain signal into a two-dimensional time-frequency domain representation using the short-time Fourier transform method, the system achieves accurate and visualized extraction of the periodic modulation texture of signal energy for the first time under the airborne rotation environment. Compared with simple time-domain or frequency-domain analysis, it can clearly reveal the signal energy migration law synchronized with the rotation period, thus providing an irreplaceable time-frequency joint feature basis for reliably dividing the availability window and the sensitivity window, thereby improving the accuracy and robustness of feature analysis from the source.

[0017] (2) By introducing the Kalman filter, the optimal estimation algorithm, into the online calibration of the rotating spectrum model, real-time and optimal estimation of the model's perturbation layer state is achieved. Its core effect is that it can dynamically track and "absorb" non-ideal disturbances caused by slight changes in rotational speed and structural vibrations, while effectively filtering out observation noise, enabling the model to continuously and accurately match the current real rotational state. This significantly improves the model's adaptability and long-term prediction accuracy in actual dynamic environments, and avoids model failure due to environmental changes.

[0018] (3) A dual judgment mechanism including signal anomaly source discrimination and model careful update is introduced. This mechanism intelligently distinguishes the root cause of anomalies by evaluating rotation consistency score and interference probability score in parallel, and robustly triggers the update of the core parameters of the model based on continuous perturbation trend and low confidence evidence, thereby avoiding misjudgment at the root cause and ensuring the high sensitivity and high stability of the model in complex environments. Attached Figure Description

[0019] Figure 1 This is a flowchart of a rotation compensation method for a missile-borne navigation receiver proposed in this invention.

[0020] Figure 2 This is a hierarchical diagram of a calibrated two-layer rotating spectrum model obtained by the rotation compensation method for an airborne navigation receiver proposed in this invention.

[0021] Figure 3 This invention provides a method for obtaining continuous navigation measurement information in a missile-borne navigation receiver rotation compensation system. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figures 1-3 This invention provides a technical solution: a rotation compensation method for an airborne navigation receiver. Specifically, the following rotation compensation method for an airborne navigation receiver is provided; please refer to [link / reference]. Figure 1 The method includes the following steps: Step S1: Obtain the complex intermediate frequency satellite signal sequence of the missile-borne navigation receiver, analyze the complex intermediate frequency satellite signal sequence using the short-time Fourier transform method, calculate the instantaneous total power sequence, identify the rotation period of the missile-borne navigation receiver based on the instantaneous total power sequence, and divide the availability window and the sensitive window.

[0024] After the missile-borne navigation receiver enters operational mode, it begins acquiring satellite signals. Specifically, the receiver performs down-conversion and analog-to-digital conversion on the radio frequency satellite signals received by the antenna, outputting a complex intermediate frequency (IF) satellite signal sequence. This sequence is mathematically represented as... Among them, characters This represents a complex intermediate frequency signal; the character 'n' represents the discrete-time index, with a value range of... N, where N is the total number of discrete time points in the complex intermediate frequency satellite signal sequence; character It is an index used to distinguish different signal processing channels (such as different satellites or different polarization channels). Indicates in the channel The complex signal sampled at time point n contains two components: in-phase (I) and quadrature (Q). This sequence fully preserves all the characteristics of the signal modulated by the projectile rotation under a specific channel, and serves as the data foundation for all subsequent analyses.

[0025] To observe the joint distribution of signal energy over time and frequency, a complex intermediate frequency satellite signal sequence was analyzed. Time-frequency analysis is performed using the short-time Fourier transform method: the long-time sequence is divided into overlapping short time segments (time frames), and a discrete Fourier transform is performed on each time frame after windowing. For the first... Each time frame, and its corresponding complex spectrum It is obtained by calculation using the following formula: ; in, Indicates channel The complex spectral value at the m-th time frame and the k-th frequency unit; the character m identifies the sequence number of the time frame; k identifies the sequence number of the frequency unit, k=0,1,…,N-1; This indicates a summation operation performed on L time points within the window; It is a passage In the complex intermediate frequency signal sequence, the sample value corresponds to the r-th point of the m-th frame; the character H is the frame shift length, that is, the time interval between the start points of adjacent time frames (in units of sampling points). It is a window function of length L (Hamming window function); character is the kernel function of the Discrete Fourier Transform, j is the imaginary unit; and N is the length of the Discrete Fourier Transform. The purpose of the formula is to transform a one-dimensional time-domain signal into a two-dimensional time-frequency domain complex representation through windowing and Fourier transform, thereby revealing the evolution of the signal frequency components with time m.

[0026] Next, by calculating the square of the modulus of each complex spectral value, a time-frequency energy distribution matrix representing the signal power distribution in the two-dimensional time-frequency plane is obtained. Its elements are .

[0027] To more clearly observe the overall trend of signal energy change over time, the time-frequency energy distribution matrix of each channel is... Summing is performed along the k-dimensional frequency index to generate a one-dimensional instantaneous total power sequence. The value of each element in this sequence is calculated using the following formula: ; Among them, characters Indicates channel Total power of the signal in the m-th time frame; character This indicates a summation operation on the frequency index k from 0 to N-1. The purpose of the formula is to compress two-dimensional time-frequency information into a one-dimensional time-domain power profile by aggregating the energy of all frequency components, thereby highlighting the periodic amplitude modulation across multiple frequency units caused by the rotation of the projectile. This is more robust to reflecting the rotation effect than observing a single frequency line.

[0028] In the generated instantaneous total power sequence In this study, by detecting the peak value of the autocorrelation function of the sequence, the basic rotation period with regular fluctuations in energy distribution was identified. Its unit is time frames.

[0029] Specifically, the process of periodic identification using the autocorrelation function method is as follows: First, consider the instantaneous total power sequence of length M. Calculate its autocorrelation function : ; in, It is the mean of the instantaneous total power sequence, used to eliminate the influence of the DC component; For time delay (number of frames); Set the maximum search delay to the preset value; Then normalization is performed. Divide by the autocorrelation value at zero delay The normalized autocorrelation coefficient was obtained. Its value range is [-1, 1], which facilitates the standardization of threshold judgment; Next, peak detection is performed, with a time delay. Within the interval, find the normalized autocorrelation coefficient. The local maximum points (peaks). The delays corresponding to these peak points. This represents the potential period length of the sequence. The first significant peak (required) The delay corresponding to a value greater than a preset threshold (e.g., 0.5) Determined as the fundamental rotation period (Unit: frame).

[0030] To ensure robustness in cycle identification, multiple processing windows can be observed continuously to confirm the basic rotation cycle. The stability. If the fluctuation of the periodic value identified by adjacent windows is within the allowable range (e.g., If the period is valid, then the search is considered valid; otherwise, the search is repeated.

[0031] After identifying the rotation period Based on this, within a complete rotation cycle, the power sequence of the main channel is... The morphology is analyzed to divide feature windows. To ensure the robustness of the analysis, the basis for this is... Periodic samples should satisfy the following: the periodicity of power fluctuations and the identified fundamental rotation period. The data met the criteria and showed no obvious sudden distortions, thus dividing the data into two distinct sub-intervals. The division was based on a fixed power threshold. This threshold is determined through experimental statistics, for example, by setting it to 0.7 times the long-term average power of the channel.

[0032] Availability Window Corresponding to Values ​​consistently above the threshold Small fluctuations (e.g., variance below a threshold) This threshold is determined through experimental statistics, for example, set as a time period (1.5 times the power variance of the availability window). Within this window, carrier phase continuity can be reliably maintained.

[0033] Sensitive window Corresponding to The value dropped to the threshold The following period may see sharp fluctuations. Within this window, the signal is prone to significant attenuation and phase abrupt changes.

[0034] Window boundaries The following steps are used to determine: Specifically, the detection sequence. Crossing the threshold All points are selected as initial candidate locations for the boundary; in the neighborhood of each initial candidate location... Inside (e.g.) (Sampling points), calculate the first difference of the sequence. By checking the neighborhood Are all absolute values ​​less than a preset smoothness threshold? To determine if all difference values ​​within the neighborhood This indicates that the power change trend at that point is stable, and it is a true boundary point (such as a flat drop from high power to below the threshold, or a flat rise from low power back to above the threshold). Therefore, it is confirmed as... or Conversely, if the difference value changes drastically ( If a point is found to have a noise spike or abrupt change, it indicates that such a candidate point should be eliminated, and the search should continue in the neighboring region for points that satisfy the smoothness condition as the actual boundary. The points that ultimately satisfy the smoothness condition are confirmed as the actual window boundaries. or .

[0035] Subsequently, the stability of the window partitioning was verified. This was done by observing the window over N consecutive rotation cycles (e.g., N=10). and Starting point location, calculate its positional change trend. (In frames). Simultaneously, a confidence factor C is calculated to quantify the reliability of this window partitioning result. C is specifically calculated by: calculating the variance of the window starting point sequence over N consecutive periods. Then through After normalization, the closer the value of C is to 1, the more stable the window position and the higher the confidence level.

[0036] At this point, step S1 is complete, and its output window time parameters are... Basic rotation period Trends The confidence factor C provides the necessary prior information and input conditions for the modeling and calibration of the subsequent step S2.

[0037] Step S2: Based on the availability window and sensitivity window, construct a two-layer rotating spectrum model; calibrate the two-layer rotating spectrum model to obtain a calibrated two-layer rotating spectrum model.

[0038] Following step S1, this step utilizes the extracted window time parameters. Basic rotation period Trends A two-layer rotation spectrum model is constructed using a confidence factor C, and then calibrated online using a Kalman filter algorithm to accurately characterize the dynamic characteristics of the signal modulated by rotation.

[0039] First, the foundational part of the two-layer rotating spectrum model—the periodic layer—is constructed. This periodic layer aims to describe a stable and predictable repetitive modulation structure of the signal energy within one rotation cycle using a set of core parameters. Based on the output of step S1, the parameter vector of the periodic layer is initialized. It contains the following elements: ; in, Represents the parameter vector of the periodic layer; Indicates a sensitive window In a standard rotation cycle The normalized phase angle (in radians) of the inner starting point is determined by... The start time is mapped to obtain; Indicates a sensitive window Phase width; Availability window Mean internal signal power; character Availability window Standard deviation of internal signal power; It is a quantization sensitive window The parameter representing the power attenuation trend within the internal signal is obtained through statistical analysis of the power attenuation curves observed in step S1 over multiple consecutive rotation cycles. It represents the overall rate and shape pattern of power decrease as the phase increases within the window. The purpose of the formula is to transform the periodic window texture observed in step S1 into a set of mathematical parameters with clear physical meaning, thereby establishing a benchmark framework for describing the ideal periodic modulation behavior of the entire two-layer rotating spectrum model.

[0040] Secondly, the dynamic part of the two-layer rotating spectrum model—the perturbation layer—is constructed. This layer describes the time-varying, non-ideal perturbations superimposed on the regular structure described by the periodic layer, such as spectral distortions caused by uneven rotation and changes in occlusion. The perturbation layer is mediated by a state vector. At discrete time The vector is represented and then estimated and updated in real time: ; in, Indicates at time The state vector of the perturbation layer; Indicates at time The actual starting phase of the sensitive window relative to the reference phase of the periodic layer The instantaneous offset; Indicates at time Mean measured power within the availability window relative to the baseline mean of the periodic layer The instantaneous offset; Indicates at time Measured attenuation slope within the sensitive window relative to the periodic layer reference parameter The instantaneous offset. The purpose of defining the perturbation layer is to enable the model to accommodate and express small deviations from the real situation in short-term fluctuations without immediately changing the fundamental description of periodic patterns.

[0041] Next, to endow the two-layer rotating spectrum model with the ability to dynamically adapt to environmental changes, a Kalman filter algorithm is used to analyze the state of the perturbation layer. Perform online recursive estimation. The calibration process occurs at each processing time. The core of execution is to construct and derive the state prediction equation and the observation equation.

[0042] State prediction equation (describes the evolution of the state from the previous time step to the current time step): ; in, Indicates based on Time information The prior prediction of the state of the perturbation layer at any given time; F is the state transition matrix, which is set here as the identity matrix, assuming that the perturbation changes slowly over a short period of time; character yes The optimal posterior estimate of the state of the perturbation layer at time t; w This is process noise, modeled as a Gaussian white noise sequence with zero mean and covariance matrix Q, used to characterize the uncertainty in the evolution of the perturbation state. The purpose of the formula is to predict the perturbation state at the current moment based on the optimal estimate from the previous moment.

[0043] Observation equations (establishing the correlation between actual observations and system state): ; Among them, characters At any moment The observation vector, whose elements are derived from features obtained from real-time signals, such as... (Measured window phase offset) (Measured power mean offset); Character This is the observation matrix, set here as an identity matrix, indicating that the observations directly correspond to state variables; characters It is observation noise, modeled as having a mean of zero and a covariance matrix of... The Gaussian white noise vector is used to reflect the measurement error in the feature extraction process. The purpose of the formula is to establish the true hidden state. To actual measurable The relationship between them.

[0044] Kalman filtering recursively performs the following steps: First, make a prediction using the state prediction equation, based on the optimal estimate from the previous time step. and its error covariance Calculate the prior state prediction at the current time. and its prediction error covariance ; Update again to obtain the actual observation vector at the current time. Then, calculate the Kalman gain. Subsequently, using the observation equation and , prior prediction With new observations By fusing the data, we can obtain the optimal posterior state estimate at the current time step. and its estimation error covariance ; The filtering process involves output and iteration at each time step. The final output is the optimal unbiased estimate of the current perturbation layer state. and its estimated error covariance matrix . This characterizes the uncertainty of the estimation results. It will be used to correct the rotating neck spectrum model in real time, and This feedback is then used in the prediction step at the next moment, and so on iteratively. The purpose of this process is to track the dynamics of the disturbance in real time with the optimal estimation method and effectively filter out observation noise.

[0045] The calibration of the perturbation layer continues, and the periodic layer parameters... The update requires more caution and must be triggered only when two conditions are met simultaneously: First, the perturbation state estimated by the Kalman filter (such as...) In continuous (For example Within a certain number of cycles, a clear systematic trend is observed, and this trend corresponds to the window drift trend output in step S1. First, the trend is consistent in direction; second, the statistical significance of this trend is high (e.g., the mean offset within the sliding window exceeds a threshold). (degrees / week), and the confidence factor provided in step S1 Below the stability threshold This supports the judgment that "the rotational dynamics have truly changed." When the conditions are met, the statistical mean of the long-term trend of the disturbance state (such as...) is used. Feedback is sent to the corresponding parameters of the periodic layer (e.g., This mechanism ensures that the model baseline can robustly adapt to real-world changes, while avoiding model instability caused by transient noise or isolated disturbances.

[0046] Step S2 is now complete. Its core output is an online-calibrated two-layer rotating spectrum model, which contains stable periodic layer parameters. and real-time estimated state of the perturbation layer This provides a dynamic and reliable benchmark for the collaborative discrimination in step S3.

[0047] Step S3: Using the calibrated rotating spectrum model as a reference, the abnormal signals detected by the missile-borne navigation receiver are converted into observation vectors, and the rotational consistency score and interference probability score are calculated; based on the rotational consistency score and interference probability score, the discrimination result of the abnormal signal is determined.

[0048] Following step S2, this step uses the online-calibrated dual-layer rotating spectrum model as a benchmark to perform quantitative analysis and collaborative discrimination on the abnormal signals monitored in real time by the missile-borne navigation receiver, in order to distinguish between the rotation effect and real external interference. First, the missile-borne navigation receiver compares the instantaneous total power in real time. Compared with the periodic layer reference power in step S2 The difference is used to continuously monitor signal quality. When the conditions are met... When this occurs, a "significant degradation" is determined, and the exception handling process is triggered. Among these, The preset power drop threshold (e.g., 3dB) is determined through experimental statistics to effectively distinguish between periodic decay caused by rotation and random fluctuations.

[0049] Subsequently, the anomalous time window was determined and the observation vector was constructed. The duration of the anomalous period was defined as the period from when the power drop fell below a threshold. Start, until power recovers to the threshold. The above-mentioned period of time remains stable. Within this time window, the receiver performs parallel computation and extracts five key parameters, which are designed as a set of observations that simultaneously and independently reflect the rotation modulation characteristics and potential interference characteristics, together forming a structured observation vector. This vector is calculated within the time window of the abnormal event, and its specific form is as follows: ; Where O represents the observation vector; Represents the current instantaneous total power Relative to the periodic layer reference power The decrease (in dB); This represents the instantaneous change in total power within the time window of the abnormal event. The variance of the power fluctuation is used to measure the severity of the power fluctuation. This represents the difference between the actual starting phase of the current sensitive window estimated based on real-time signal analysis and the current phase estimate predicted by the two-layer rotating spectrum model. The difference between them (unit: radians); This represents the average signal power measured within the availability window at the current moment, compared to the current power reference value given by the two-layer rotated spectrum model. The difference between them; The relative entropy between the measured spectral energy distribution during an anomalous signal period and the undisturbed ideal spectral energy distribution predicted by the two-layer rotating spectral model at the current phase is used to measure the degree of distortion in the spectral structure. T is the transpose. The purpose of the formula is to transform specific anomalous signal data into a structured data object containing multidimensional, quantifiable evidence, providing input for subsequent quantitative scoring.

[0050] Secondly, calculate the rotational consistency score. This score aims to quantify the degree of agreement between the characteristics of the current anomalous signal and the rotation modulation pattern predicted by the two-layer rotation spectrum model: ; in, This represents the rotational consistency score, and its range is... A higher value indicates a better match with the rotation model; Represents the natural exponential function; It is the phase tolerance, set based on three times the standard deviation of the historical window position estimation error; Indicates based on periodic layer parameters The described linear attenuation model, in phase shift The expected power attenuation value calculated at the location; It is the tolerance parameter for power matching; This is a relative entropy saturation threshold used to prevent a single indicator from excessively influencing the response. The purpose of the formula is to evaluate the observation vector from three dimensions: spatiotemporal consistency (first term), process consistency (second term), and structural consistency (third term), through a product-like comprehensive function. The degree of matching with the model prediction.

[0051] Next, the interference probability scores are calculated in parallel. This score aims to quantify the degree to which anomalous signals, which cannot be explained by rotational models, but conform to typical interference characteristics.

[0052] ; in, The interference probability score is represented by the range of values. The higher the score, the less the observed abnormal features conform to the rotation modulation law and the more they conform to the characteristics of sudden interference, thus the greater the possibility of external interference. This represents the hyperbolic tangent function, used to compress variance measures to a smaller value. interval; It is the normalized benchmark for power fluctuation variance, determined based on experiments of typical fluctuation levels under rotational modulation; It is a normalized benchmark for the power mean offset. The purpose of the formula is to generate a score characterizing the risk of external interference by assessing the suddenness / randomness of the anomaly (first term), the destructiveness of the power benchmark (second term), and the destructiveness of the spectral structure (third term).

[0053] Finally, based on the calculated rotational consistency score And interference probability score The score is determined collaboratively according to preset decision rules, and a definite output is given. R: ; in, Indicates the judgment result of the abnormal signal; The preset decision threshold is determined through ROC curve analysis of a large amount of simulation and measured data, and can be taken as... , , , A high confidence threshold is used to determine whether rotational or interference features are extremely significant and absolutely dominant. With moderate confidence, it is used to identify "superposition" states where both rotation and interference features are obvious and coexist. The low confidence threshold is used to exclude the weak influence of another factor, ensuring the purity of the judgment. and All with The significance of the comparison lies in the fact that only when both scores exceed the moderate confidence level are the rotation effect and interference effect considered significant, thus classifying it as a superposition scenario. If either score is low, the influence of that factor is considered weak and does not constitute "superposition." The purpose of the formula (decision logic) is to map the continuous score space to a discrete, explicit decision category, providing clear and actionable instructions for step S4. At this point, step S3 is complete, and its output is the decision result. This serves as the direct basis for subsequent targeted compensation or suppression.

[0054] Step S4: Based on the discrimination result of the abnormal signal, perform rotation compensation on the missile-borne navigation receiver to generate continuous navigation measurement information.

[0055] Following step S3, this step adopts corresponding signal processing strategies for the missile-borne navigation receiver based on the discrimination result R of the output abnormal signal, in order to compensate for the rotation modulation effect and suppress interference, and finally generate continuous navigation measurement information. The input for this step includes two key pieces of information: the first is the result of the abnormal signal discrimination output from step S3. Its value belongs to the set ("rotation-dominated", "interference-dominated", "rotation and interference superposition"); the second term is the current signal processed in real time by the missile-borne navigation receiver. .according to Depending on the value of , subsequent operations will proceed along one of three preset paths.

[0056] When the result of abnormal signal discrimination When the mode is "rotation-dominated" or "rotation and interference superimposed", amplitude recovery compensation and dynamic reset of the tracking loop strategy driven by a two-layer rotation spectrum model are performed.

[0057] Amplitude recovery compensation: within the sensitive window identified in step S1 and calibrated by the model in step S2. Internally, regarding the current signal Apply a model-predicted gain for inverse correction. The compensated signal... It is given by the following formula: ; in, This indicates the signal after amplitude recovery compensation; This indicates the current input signal sample value, which comes directly from the receiver's RF front-end and analog-to-digital converter; the character... This represents the compensation gain function calculated in real time, the shape of which is determined by the attenuation slope reference of the periodic layer of the two-layer rotating spectrum model. Real-time migration estimation of its perturbation layer Both parameters are derived from the model calibration output in step S2. This function is an adaptive gain that varies with time and is inversely proportional to the signal attenuation slope within the sensitive window, thus achieving inverse compensation for the amplitude of the attenuated signal.

[0058] The tracking loop strategy is dynamically reset, and the carrier tracking loop parameters of the missile-borne navigation receiver are determined and predicted in real time based on the window state determined in steps S1 and S2. or The system adaptively switches between two operating modes, dynamically resetting between them: a continuity-priority mode and a precision-priority mode. For the continuity-first mode, in the sensitive window Internal activation. This mode maintains lock-in by relaxing the loop filtering characteristics; the specific adjustments are as follows: ; in, Indicates the carrier loop equivalent noise bandwidth within the sensitive window; letters This represents the nominal loop bandwidth design value of the receiver under static or quasi-static conditions, and is an inherent parameter of the system. This represents the loop integral time within the sensitive window; This indicates the nominal integration time design value of the receiver, which is an inherent parameter of the system; the letter... and It is an adjustment coefficient greater than 1, and its specific value is determined through experiments to ensure that the loop does not lose lock under the maximum expected attenuation. These experiments are conducted based on the worst-case signal attenuation scenario predicted by the rotating modulation model established in steps S1 and S2. Simultaneously, the phase estimate of the loop at the end of the previous availability window... As the phase prediction benchmark within this window, and to help maintain stability, this value is derived from the output of the receiver tracking loop at the end of the previous availability window.

[0059] For precision-first mode, in the availability window Internal activation. In this mode, loop parameters are restored to high-precision settings:

[0060] This setting is designed to optimize tracking accuracy for stable signals and output low-noise carrier phase observations.

[0061] The suppression strategy under interference-dominant conditions is as follows: When the result of abnormal signal discrimination When the operation is "interference-dominated," the core principle is to suppress external interference without disrupting the normal operation of the two-layer rotating spectrum model. This means that window identification in step S1, model calibration in step S2, and discrimination logic in step S3 all continue to operate in the background, activating a front-end frequency domain interference detection and suppression module. This module performs frequency domain analysis on the current signal, identifies narrowband interference lines outside the common band of the availability window and sensitive window predicted by the rotating spectrum model, and applies adaptive notch filtering only to these lines. This design ensures the protection of the core frequency band of the rotating modulation. The design constraint for all suppression operations is to keep the periodic modulation structure caused by rotation in the signal from being distorted, ensuring the integrity of the rotation compensation baseline. Once the interference weakens or disappears, the discrimination result of the abnormal signal... If the situation changes, the processing flow can be immediately and seamlessly switched back to the aforementioned rotation effect compensation and reconstruction mode.

[0062] Regardless of which processing path is executed, the final operation of this step is to generate and output continuous navigation measurement information. This information is the fundamental observation data necessary for the high-precision calculation of the missile-borne integrated navigation system, and its continuity is crucial to ensuring navigation accuracy. Specifically, it includes: Carrier phase observation The value, directly output by the tracking loop after the above strategy adjustment, reflects the change of satellite signal carrier phase over time and is a core observation for high-precision velocity measurement and positioning. After compensation and strategy reconstruction, this quantity remains continuous and uninterrupted throughout the entire rotation cycle.

[0063] Pseudodistance observation Calculated by the following formula: ; in, At the speed of light, For the local time of the missile-borne navigation receiver, The satellite launch time, calculated from the navigation message, is corrected based on the stable code phase measurement obtained after compensation through the above process. Measurement status indicator This is used to attach a metadata indicator to each output observation, indicating its processing status at the time of generation. This identifier represents the carrier phase observation at the time of generation. and pseudo-distance observation It is generated and bound synchronously. Its value is determined in real time by the window state in step S1. (Values) or The anomaly detection results output in step S3 The following deterministic mapping rules jointly determine the following: ; in, Indicates "in available windows" "Generated in precision-first mode"; Indicates "in sensitive window" "Generated after rotation compensation"; This indicates that it was generated during the external interference suppression process.

[0064] At this point, step S4 is complete. Based on the intelligent discrimination results of abnormal signals, it dynamically selects and executes precise compensation, reconstruction, or suppression strategies, ultimately transforming signals affected by rotation and interference into stable and continuous navigation observations. This achieves a closed loop from "identifying the problem" to "solving the problem," ensuring the continuity and reliability of missile-borne navigation in complex dynamic environments.

[0065] This invention provides a systematic rotation compensation method for missile-borne navigation receivers, aiming to solve the problem of signal periodic attenuation caused by high-speed missile rotation, susceptibility to external interference, and subsequent navigation interruption. This method constructs a complete technical closed loop from perception and cognition to decision-making and execution. First, time-frequency analysis of the complex intermediate frequency satellite signal sequence is performed using short-time Fourier transform to generate an instantaneous total power sequence. This allows for the first clear extraction of the periodic modulation texture of the signal energy synchronized with rotation, and based on this, accurate division of availability windows and sensitivity windows representing stable quality and easy attenuation lays the foundation for accurate perception. Second, a two-layer rotation spectrum model containing a periodic layer and a disturbance layer is constructed based on the window structure, and an innovative Kalman filter algorithm is applied for online calibration. This enables the model to dynamically track and absorb non-ideal disturbances such as slight changes in rotation speed, effectively filtering out noise and achieving adaptive and accurate cognition of the model's true rotation state. Finally, the core lies in the introduction of a dual-judgment update mechanism. This mechanism intelligently identifies anomalous signals by calculating rotation consistency scores and interference probability scores in parallel and making collaborative decisions, fundamentally distinguishing between rotation effects and real external interference. Furthermore, it combines the identification results with model confidence levels, acting as a dual gate to prudently trigger updates to core model parameters, ensuring the model's long-term stability during the adaptation process. Finally, based on the intelligent identification results, the model is driven to perform amplitude recovery compensation for the signal within a sensitive window and dynamically reconstruct the tracking loop strategy. This allows for the continuous generation of highly usable carrier phase and pseudorange navigation measurement information even in environments with both rotation and interference.

[0066] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the statement "including a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0067] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their likenesses.

Claims

1. A method for rotation compensation of an airborne navigation receiver, characterized in that, Includes the following steps: Step S1: Obtain the complex intermediate frequency satellite signal sequence of the missile-borne navigation receiver, analyze the complex intermediate frequency satellite signal sequence using the short time Fourier transform method, calculate the instantaneous total power sequence, identify the rotation period of the missile-borne navigation receiver based on the instantaneous total power sequence, and divide the availability window and the sensitive window. Step S2: Based on the availability window and sensitivity window, construct a two-layer rotating spectrum model; calibrate the two-layer rotating spectrum model to obtain a calibrated two-layer rotating spectrum model; Step S3: Using the calibrated rotating spectrum model as a reference, the abnormal signals detected by the missile-borne navigation receiver are converted into observation vectors, and the rotational consistency score and interference probability score are calculated; based on the rotational consistency score and interference probability score, the discrimination result of the abnormal signal is determined. Step S4: Based on the discrimination result of the abnormal signal, perform rotation compensation on the missile-borne navigation receiver to generate continuous navigation measurement information.

2. The method for rotation compensation of a missile-borne navigation receiver according to claim 1, characterized in that, Obtain the complex intermediate frequency satellite signal sequence from the missile-borne navigation receiver, including: After the missile-borne navigation receiver enters operational mode, it begins acquiring satellite signals. The receiver performs down-conversion and analog-to-digital conversion on the radio frequency satellite signals received by the antenna, outputting a complex intermediate frequency (IF) satellite signal sequence. This sequence is mathematically represented as... , where characters This represents a complex intermediate frequency signal; the character 'n' represents the discrete-time index, with a value range of... N, where N is the total number of discrete time points in the complex intermediate frequency satellite signal sequence; character It is an index used to distinguish different signal processing channels.

3. The method for rotation compensation of a missile-borne navigation receiver according to claim 2, characterized in that, The complex intermediate frequency satellite signal sequence is analyzed using the short-time Fourier transform method to calculate the instantaneous total power sequence, including: For complex intermediate frequency satellite signal sequences Time-frequency analysis was performed using the short-time Fourier transform method: the long-time sequence was divided into overlapping short time segments, and a discrete Fourier transform was performed on each time frame after windowing. For the _th_ time segment... Each time frame has its corresponding complex spectrum. It is obtained by calculation using the following formula: ； in, Indicates channel The complex spectral value at the m-th time frame and the k-th frequency unit; the character m identifies the sequence number of the time frame; k identifies the sequence number of the frequency unit, k=0,1,…,N-1; This indicates a summation operation performed on L time points within the window; It is a passage In the complex intermediate frequency signal sequence, the sample value corresponds to the r-th point of the m-th frame; the character H is the frame shift length, that is, the time interval between the start points of adjacent time frames; It is a window function of length L; character is the kernel function of the Discrete Fourier Transform, j is the imaginary unit; character N is the length of the Discrete Fourier Transform; Next, by calculating the square of the modulus of each complex spectral value, a time-frequency energy distribution matrix representing the signal power distribution in the two-dimensional time-frequency plane is obtained. Its elements are ; The time-frequency energy distribution matrix of each channel Summing along the frequency dimension k generates a one-dimensional instantaneous total power sequence. .

4. The method for rotation compensation of a missile-borne navigation receiver according to claim 3, characterized in that, Based on the instantaneous total power sequence, the rotation period of the missile-borne navigation receiver is identified, and an availability window and a sensitive window are defined, including: In the generated instantaneous total power sequence In this study, by performing periodogram analysis or autocorrelation function peak detection on the sequence, the basic rotation period with regular fluctuations in energy distribution is identified. Its unit is time frames; After identifying the rotation period Based on this, within a complete rotation cycle, the power sequence of the main channel is... The morphology is analyzed to divide feature windows. To ensure the robustness of the analysis, the basis for this is... Periodic samples should satisfy the following: the periodicity of power fluctuations and the identified fundamental rotation period. The data meets the criteria and shows no obvious sudden distortion, thus dividing the data into two distinct sub-intervals based on a fixed power threshold. This threshold was determined through experimental statistics; Availability Window Corresponding to Values ​​consistently above the threshold The fluctuations are relatively small, and within this window, the carrier phase continuity can be reliably maintained; Sensitive window Corresponding to The value dropped to the threshold The following periods may see sharp fluctuations; within this window, the signal is prone to significant attenuation and phase abrupt changes.

5. The method for rotation compensation of a missile-borne navigation receiver according to claim 4, characterized in that, Based on the availability window and sensitivity window, a two-layer rotating spectrum model is constructed, including: The fundamental part of constructing the two-layer rotating spectrum model is the periodic layer. This periodic layer aims to describe the stable and predictable repetitive modulation structure of the signal energy within one rotation cycle using a set of core parameters. Based on the output of step S1, the parameter vector of the periodic layer is initialized. ; Secondly, the dynamic part of the two-layer rotating spectrum model—the perturbation layer—is constructed. The perturbation layer is mediated by a state vector. At discrete time Characterization: ； in, Indicates at time The state vector of the perturbation layer; Indicates at time The actual starting phase of the sensitive window relative to the reference phase of the periodic layer The instantaneous offset; Indicates at time The mean measured power within the availability window relative to the baseline mean of the periodic layer. The instantaneous offset; Indicates at time The measured attenuation slope within the sensitive window relative to the periodic layer reference parameter The instantaneous offset.

6. The method for rotation compensation of a missile-borne navigation receiver according to claim 5, characterized in that, The two-layer rotating spectrum model is calibrated to obtain a calibrated two-layer rotating spectrum model, including: Kalman filtering algorithm is used to analyze the state of the perturbation layer. Online recursive estimation is performed, and the calibration process is carried out at each processing time. The core of execution is to construct and derive the state prediction equation and the observation equation; The Kalman filter recursively performs prediction and update steps, ultimately outputting the optimal unbiased estimate of the state of the perturbation layer at the current moment. And its estimated error covariance matrix.

7. The method for rotation compensation of a missile-borne navigation receiver according to claim 6, characterized in that, Based on the calibrated rotating spectrum model, the abnormal signals detected by the missile-borne navigation receiver are transformed into observation vectors, including: The missile-borne navigation receiver compares the instantaneous total power in real time. Compared with the periodic layer reference power in step S2 The difference is used to continuously monitor signal quality, and when the conditions are met... When this occurs, a "significant degradation" is determined and the exception handling process is triggered. The preset power drop threshold; Subsequently, the anomalous time window was determined and the observation vector was constructed. The duration of the anomalous period was defined as the period from when the power drop fell below a threshold. Start, until power recovers to the threshold. Within a stable, continuous time window, the receiver performs parallel computations and extracts five key parameters, which together form a structured observation vector. This vector is calculated within the time window of the abnormal event, and its specific form is as follows: ； Where O represents the observation vector; Represents the current instantaneous total power Relative to the periodic layer reference power The magnitude of the decline; This represents the instantaneous change in total power within the time window of the abnormal event. The variance; This represents the difference between the actual starting phase of the current sensitive window estimated based on real-time signal analysis and the current phase estimate predicted by the two-layer rotating spectrum model. The difference between them; This represents the average signal power measured within the availability window at the current moment, compared to the current power reference value given by the two-layer rotated spectrum model. The difference between them; T represents the relative entropy between the measured spectral energy distribution during the abnormal signal period and the unperturbed ideal spectral energy distribution predicted by the two-layer rotating spectral model at the current phase, where T is the transpose.

8. The method for rotation compensation of a missile-borne navigation receiver according to claim 7, characterized in that, The rotational consistency score and disturbance probability score are calculated, including: Calculate rotational consistency score : ； in, This represents the rotational consistency score, and its range is... ; Represents the natural exponential function; It is the phase tolerance, set based on three times the standard deviation of the historical window position estimation error; This is the expected power attenuation value; It is the tolerance parameter for power matching; It is a relative entropy saturation threshold; Next, the interference probability scores are calculated in parallel. : ； in, The interference probability score is represented by the range of values. ; Represents the hyperbolic tangent function; It is the normalized benchmark for power fluctuation variance, determined based on experiments of typical fluctuation levels under rotational modulation; It is a normalized benchmark for the power mean offset.

9. A method for rotation compensation of a missile-borne navigation receiver according to claim 8, characterized in that, Based on the rotational consistency score and the interference probability score, the discrimination result of the abnormal signal is determined, including: Based on the calculated rotational consistency score And interference probability score The score is used for collaborative discrimination based on preset decision rules, and a definite category result R is output: ； in, Indicates the judgment result of the abnormal signal; It is a preset decision threshold that maps the continuous score space to a discrete, explicit decision category.

10. A method for rotation compensation of a missile-borne navigation receiver according to claim 9, characterized in that, Based on the discrimination result of the abnormal signal, rotation compensation is performed on the missile-borne navigation receiver to generate continuous navigation measurement information, including: Based on the discrimination result R of the output abnormal signal, corresponding signal processing strategies are adopted for the missile-borne navigation receiver to compensate for the rotation modulation effect and suppress interference, ultimately generating continuous navigation measurement information, including: Carrier phase observation The tracking loop, after strategy adjustment, outputs directly, and this quantity remains continuous throughout the entire rotation cycle without interruption. Pseudodistance observation Calculated by the following formula: ； in, At the speed of light, For the local time of the missile-borne navigation receiver, The satellite launch time is calculated from the navigation message; Measurement status indicator This is used to attach a metadata indicator to each output observation, indicating the processing status at the time of its generation.

Images