LoRa signal anti-interference decoding method for LPWAN star-ground fusion
By identifying ground station interference signals and combining dynamic CFO estimation and rectangular window functions, the problems of weak low-orbit satellite signals and dynamic frequency offset were solved, achieving stable decoding of LoRa signals in satellite-ground integrated communication and improving data packet reception rate.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- YANSHAN UNIV
- Filing Date
- 2026-04-28
- Publication Date
- 2026-08-04
AI Technical Summary
Low-Earth orbit satellite signals are weak, dynamic carrier frequency offset is significant, and data packet collisions are severe. Existing technologies cannot effectively decode LoRa signals, especially in satellite-ground integrated communication, where there are decoding challenges due to interference and high dynamic channels.
By identifying ground station interference signals, using a multi-data packet aggregation method and dynamic CFO estimation, combined with a rectangular window function to erase the interference spectrum, and using a satellite orbit model to predict Doppler frequency shift, anti-interference decoding of LoRa signals is performed.
Stable decoding of LoRa satellite signals under strong interference and high dynamic channel conditions was achieved, improving data packet reception rate and decoding success rate.
Smart Images
Figure CN122512980A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of satellite-ground converged communication technology, and in particular relates to a LoRa signal anti-interference decoding method for LPWAN satellite-ground convergence. Background Technology
[0002] Low Earth Orbit (LEO) satellite constellations, with their advantages of low latency and global coverage, have become a key solution for IoT communication in remote areas. LoRa technology, due to its long-range and low-power characteristics, is widely used in satellite-ground converged communication, but it faces three core challenges: 1. Weak satellite signal: Low-Earth orbit satellites are far from the ground, and the received power of the signal is extremely low after path loss (RSSI is usually between -90dBm and -130dBm), making it easy to be masked by strong ground interference; 2. Significant dynamic carrier frequency offset (CFO): The high-speed movement of the satellite generates a Doppler frequency shift of up to 20kHz and a change rate of 270Hz / s, which cannot be adapted to the existing terrestrial LoRa static CFO estimation methods; 3. Severe data packet collisions: Concurrent transmission from multiple satellites and interference from ground gateways on the same frequency cause data packet collisions. Existing collision decoding solutions are unable to cope with high dynamic and low signal-to-noise ratio scenarios between satellite and ground.
[0003] In existing technologies, TinyGS can only decode the strongest signal, SatNOGS does not integrate a dedicated LoRa decoder, and land collision decoding methods such as CIC and Mc-LoRa cannot adapt to satellite-to-ground Doppler shift and low signal-to-noise ratio environments. There is an urgent need for an anti-interference decoding solution for satellite-to-ground fusion scenarios. Summary of the Invention
[0004] In view of this, the purpose of this application is to provide a LoRa signal anti-interference decoding method for LPWAN satellite-ground integration, which can achieve stable decoding of LoRa satellite signals under strong interference and high dynamic channel conditions.
[0005] This application provides a method for anti-interference decoding of LoRa signals for LPWAN satellite-ground fusion, including: Based on the fact that the RSSI of the ground station signal is much higher than that of the satellite signal, the interference signal from the ground station in the mixed signal is identified. The Preamble, SFD and Payload parts of the interference signal are demodulated and the corresponding frequency range is erased to filter out the interference signal. A multi-data packet aggregation method is used to accumulate the spectra of detected Preambles that have undergone dechirp operation and FFT transformation. Based on the characteristic that the peak superposition result is many times greater than the peak value of a single chirp, satellite data packets are detected. A rough estimate of CFO is made based on the predicted Doppler frequency shift value of the Preamble part of the satellite data packet. Then, a fine-grained estimate of dynamic CFO is made based on the last two symbols of the Preamble part of the satellite data packet and the first two symbols of the SFD part, combined with the data packet aggregation approach. The dynamic CFO fine-grained estimation result and the time offset value of the satellite data packet are used together as the features of the target satellite data packet to decode the payload part of the satellite data packet.
[0006] Furthermore, the preamble portion of the interference signal is demodulated using the following method: Employ a signal with a chirp slope consistent with satellite data packets and larger SF and BW parameters. Matching with the preamble portion of the interference signal to detect whether the interference signal is inter-channel interference; If a match is successful, it indicates cross-channel interference. Then, based on the larger SF and BW parameters of the successful match, a dechirp operation and FFT transform are performed to obtain the spectrum of the Preamble part. If the match fails, it means that it is not cross-channel interference. In this case, the original SF and BW parameters of the satellite data packet are used to perform the dechirp operation and FFT transformation. The interference spectrum erasure range of the Preamble section is determined using the following method: Assume the RSSI of the interfering data packet is The corresponding amplitude is The satellite data packet RSSI is To ensure that the power of the interference signal is lower than that of the satellite signal, it is necessary to follow... The power of the interference signal is attenuated, and the corresponding amplitude attenuation factor is... To meet the need to eliminate strong interference; The spectrum function of the rectangular window is as follows: ; Among them, the peak value of the main lobe The side lobe amplitude is ; Then the sidelobe amplitude attenuation is the main lobe peak value. The position of the side lobe is ; The frequency distance between the sidelobes and the main lobe was further obtained as follows: ; Set the center frequency of the Preamble portion of the interference signal to be... The design frequency erasure range is... To reduce the amplitude of the side lobes of the covered rectangular window to a minimum. This part ensures that the amplitude of the interference signal is weakened to the target multiple.
[0007] Furthermore, the SFD portion of the interference signal is demodulated and the corresponding spectral range is erased using the following method: Because the SDF portion of the interference signal consists of and 2.25 conjugates Therefore, the dechirp operation on the SDF part is multiplied by... Then, the spectrum of the SDF part is obtained through FFT transformation; Within each decoding window, peaks exceeding a threshold are erased. In the last 0.25 chirs of the SDF section, since it has only accumulated the energy of 1 / 4 chirp, its spectrum needs to be quadrupled before being removed by the threshold.
[0008] Furthermore, the payload portion of the interference signal is demodulated and the corresponding spectral range is erased using the following method: Locating the two peaks in the payload portion To prevent missed detections due to peak values below a threshold, the design incorporates the spectrum of the payload portion after the dechirp operation and FFT transformation. Partially shifted to the right by one bandwidth and Partial addition is performed to re-superimpose the originally dispersed peaks, thereby obtaining... Partial peak Then, by using bandwidth relationships, the location is determined. peak ; calculate Corresponding frequency erasure range: To prevent useful information from being erased, based on two peak values... The frequency erasure range of the payload portion is recalculated based on the proportion of the full peak value. The specific method is as follows: Assume that the amplitude of the first peak value is equal to the amplitude of the full peak value. times The amplitude of the second peak is equal to the amplitude of the complete peak. If the amplitudes of the two peaks are multiplied by a factor of 1, then the amplitudes after the two peaks decay are and Then, based on the rectangular window function, the erasure range of the first peak is calculated as follows: ,in The erasure range of the second peak ,in .
[0009] Furthermore, the Doppler shift of the Preamble portion of the satellite data packets is predicted using the following method: The SGP4 orbital model is used to model the satellite's orbital state, and two lines of orbital data (TLE) are used as input to calculate the orbital state at any given time. The position vector of the satellite in the TEE coordinate system and speed Vector; Introducing Greenwich Mean Time (GMT) to calculate the rotation angle caused by Earth's rotation. And construct its corresponding rotation matrix as follows: ; Based on the rotation matrix, the satellite state vector in the TEE coordinate system is transformed to obtain the satellite position vector in the ECEF coordinate system. With velocity vector They are respectively: ; ; In the formula, This is the vector of Earth's rotational angular velocity; The latitude, longitude, and altitude of the ground receiver are converted into a position vector in the ECEF coordinate system based on the WGS-84 standard. ; The ellipsoid parameters are set as follows: Equatorial radius: Flatness: First eccentricity squared: Radius of the Mao-You circle: ; The position vector of the receiver in the ECEF coordinate system can be calculated using the following formula: ; In the formula, For the longitude of the ground receiver, Latitude of the ground receiver The altitude of the ground receiver; Then, the predicted Doppler shift value of the Preamble portion of the satellite signal is calculated using the following formula. : ; In the formula, The wavelength of the satellite signal. For the clock frequency offset between the satellite and the receiver, express In the ECEF coordinate system The amount, express In the ECEF coordinate system The amount, express In the ECEF coordinate system The amount.
[0010] Furthermore, a rough estimate of the CFO can be made using the following methods: The Doppler frequency shift caused by the relative motion between the satellite and the ground station, and the frequency shift caused by receiver hardware non-ideals, are taken as the total offset and defined as follows: Will to The frequency shift caused by the signal in the same direction, and the time shift caused by the misalignment of the demodulation window with the LoRa data packet are defined as follows: The signal causes a frequency shift in the opposite direction; but The resulting frequency shift is expressed as: ; In the formula, To represent the Nyquist correlation value, ; as well as The resulting frequency shift is expressed as: ; Further obtain The rough estimates are as follows: ; .
[0011] Furthermore, dynamic CFO fine-grained estimation is performed in the following ways: Based on the predicted Doppler shift of the satellite signal, the Doppler change per second of the satellite is calculated, and then the Doppler change rate of each symbol of the satellite signal is calculated as follows: ; In the formula, The duration of each symbol, ; Define the initial CFO value of the satellite signal. The search scope is ,in, This is a rough estimate of the error. Based on the data packet aggregation approach, each search frequency is used The last two symbols of the preamble portion of the satellite signal and the first two symbols of the SFD section Perform frequency compensation and obtain the compensated FFT peak accumulation value. : ; in, ; ; ; ; If the current With If there is a perfect match, the peak value will reach its maximum, and therefore, its corresponding frequency is... : ; At this point, the CFO value of each Chirp in the satellite signal payload is expressed as: ; In the formula, Indicates the first A symbol.
[0012] Furthermore, when there is no cross-channel interference, the satellite data packet payload is decoded using single-channel matching in the following manner: The peak value of the last symbol in the Preamble portion of the target satellite data packet is used as the matching benchmark for the peak height of the Payload portion. The specific method is as follows: Align the decoding window with the last Preamble portion of the target satellite data packet; use The target satellite data packets are compensated so that the compensated Preamble starting frequency returns to 0; the first peak height in the decoding result is extracted and defined as the matching reference. Payload matching and decoding: utilizing Align the target satellite data packet payload with the decoding window; use Carrier offset compensation is performed on each payload symbol; dechirp operation and FFT transform are performed on each decoding window to obtain the decoding peak values of the satellite signal and the interference signal; the similarity of the decoding peak value with the matching reference is compared, and the peak value that is closest to the matching reference is selected as the conflict decoding result of the current window; when there are multiple peak results that are close to the matching reference, the decoding window is moved to the left and right by the same amount of time, and the conflict decoding results that show the same trend are identified as the target decoding peak value.
[0013] Furthermore, when cross-channel interference exists, the payload portion is subjected to cross-channel matching decoding in the following manner: Use of standards The decoder performs initial signal processing; it uses a chirp slope consistent with satellite data packets and larger SF and BW parameters. The decoder performs secondary processing; by comparing the two decoding results, the target signal is identified based on the signal characteristic that the peak position shifts regularly but the amplitude remains stable.
[0014] The LoRa signal anti-interference decoding method for LPWAN satellite-ground integration provided in this application can achieve stable decoding of LoRa satellite signals under strong interference and high dynamic channel conditions. Attached Figure Description
[0015] Figure 1 This paper illustrates the Lora packet reception rate under different power differences according to an embodiment of this application. Figure 2 This illustration shows the LoRa packet reception rate under different delays provided in the embodiments of this application. Figure 3 This paper presents a schematic diagram illustrating the spectral range of a strong interference signal provided in an embodiment of this application. Figure 4 A schematic diagram showing the correspondence between the peak values of the interference signal Payload spectrum provided in the embodiments of this application is shown; Figure 5 This illustration shows a schematic diagram of satellite data packet preamble aggregation detection provided in an embodiment of this application; Figure 6 This illustration shows a schematic diagram of the satellite data packet preamble aggregation detection results provided in an embodiment of this application; Figure 7 A schematic diagram of the TEE coordinate system provided in an embodiment of this application is shown; Figure 8 A schematic diagram of the ECEF coordinate system provided in an embodiment of this application is shown; Figure 9 This illustration shows a schematic diagram of single-channel payload matching decoding provided in an embodiment of this application; Figure 10 This illustration shows another single-channel payload matching decoding schematic provided in an embodiment of this application; Figure 11 This illustration shows a cross-channel payload matching decoding diagram provided in an embodiment of this application; Figure 12 A flowchart of a LoRa signal anti-interference decoding method for LPWAN satellite-ground fusion provided in an embodiment of this application is shown. Detailed Implementation
[0016] To make the objectives, technical solutions, and advantages of this technical solution clearer, the following detailed description, in conjunction with specific embodiments, further illustrates this technical solution. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of this technical solution.
[0017] To address the technical challenges of weak satellite signals, significant dynamic CFO, and severe packet collisions in space-ground converged communications, and to achieve stable decoding of LoRa satellite signals under strong interference and highly dynamic channels, this application proposes an anti-interference decoding method for LoRa signals in LPWAN space-ground converged communications. Please refer to [reference needed]. Figure 12 The flowchart shown illustrates a LoRa signal anti-interference decoding method for LPWAN satellite-to-ground fusion. The method includes: S101. Based on the fact that the RSSI of the ground station signal is much higher than that of the satellite signal, the interference signal from the ground station in the mixed signal is identified, and the Preamble, SFD and Payload parts of the interference signal are demodulated and the corresponding frequency range is erased to filter out the interference signal.
[0018] S102. Using a multi-data packet aggregation method, the spectrum of the detected Preamble after performing the dechirp operation and the FFT transformation is accumulated. Based on the characteristic that the peak superposition result is many times greater than the peak value of a single chirp, satellite data packets are detected.
[0019] S103. Based on the predicted Doppler frequency shift value of the Preamble part of the satellite data packet, a rough estimate of CFO is made. Then, based on the last two symbols of the Preamble part of the satellite data packet and the first two symbols of the SFD part, a dynamic fine-grained estimate of CFO is made by combining the data packet aggregation approach.
[0020] S104. The dynamic CFO fine-grained estimation result and the time offset value of the satellite data packet are used together as the features of the target satellite data packet to decode the payload part of the satellite data packet.
[0021] The following section will detail the LoRa signal anti-interference decoding method for LPWAN satellite-ground integration proposed in this application.
[0022] Example 1: Detecting ground station data packets: This embodiment utilizes Received Signal Strength Indicator (RSSI) to distinguish between strong satellite and ground-based interference signals. Since the receiver simultaneously receives mixed signals from both satellites and ground stations, and ground stations are closer and experience less path loss, their RSSI is typically much higher than that of satellite signals. This creates a capture effect, where strong ground-based interference signals capture the receiver's demodulation resources, while satellite signals are suppressed and difficult to decode correctly. Therefore, this embodiment leverages this power difference to identify and filter ground station data packets with higher RSSI at the receiver, thereby improving the detection and decoding capabilities for satellite data packets with lower RSSI.
[0023] To verify the feasibility of this approach, this embodiment deploys two commercial LoRa wireless nodes (model SX1262) as transmitters, with a transmission interval of 1 second and a data packet length of 0.2 seconds. The receiver uses a UERPN210 and runs the gr-lora receive decoding program. The RSSI of the two nodes is controlled by adjusting their spatial layout. The two nodes transmit different content to distinguish the source, and content decoding is used to determine whether each node's data packet has been completely received. A transmission delay is added during the transmission process of one node to prevent two data packets from arriving at the receiver simultaneously, and the center frequency of that node is modified to simulate the Doppler shift caused by the high-speed movement of a satellite. Unless otherwise specified, the parameters of all nodes used are configured with a spreading factor SF=8 and a bandwidth BW=125kHz.
[0024] Experimental results show that most successfully received data packets originate from the first node, which is closer and has a higher RSSI. When the frequency offset of the second node is set to 10kHz and the transmission delay is 1 / 10 of the chirp duration of a linear frequency modulated signal, the following results are obtained: Figure 1 The diagram shows the LoRa packet reception rate under different power differences. Figure 1 In the diagram, red indicates data packets sent normally by the first node, and blue indicates data packets sent with a delay by the second node. For example... Figure 1 As shown, when the power difference between two data packets is greater than 0.4dB, the data packet with higher power will be received stably. However, as shown in the gray area, when the power of the two packets is in the range of 0~0.4dB, both data packets may be received, and there may be cases where some data packets cannot be completely and correctly decoded.
[0025] For further details, please refer to... Figure 2 The diagram shows the LoRa packet reception rate under different delays. Figure 2 As shown, when the RSSI of the first node's data packet is 0.5 dB higher than that of the second node, and the second node has a 10 kHz frequency offset, the first node's data packet can achieve a stable reception rate of over 90% as long as the transmission interval between the two data packets is within the range of 0.06 to 0.94 chirp lengths. If the two data packets are too close in time, their chirp signals will overlap in the time domain, making it difficult for the receiver to distinguish their respective start and end positions.
[0026] Example 2: Filtering interference signals: (1) Method for erasing the preamble of the interference signal: For the preamble section, the received signal is aligned with the decoding window, and a signal with a chirp slope consistent with the satellite data packets and larger SF and BW parameters is used. The signal is matched with the received signal to detect whether there is cross-channel interference in the received signal; if the match is successful, it means that cross-channel interference (i.e., strong ground interference) has been detected, and the standard LoRa decoding process is performed based on the larger SF and BW parameters (the purpose is to accurately locate the location and characteristics of the interference in preparation for subsequent erasure); if the match fails, it means that no cross-channel interference has been detected, and the standard LoRa decoding is performed directly using the original SF and BW parameters of the satellite data packet.
[0027] Typically, in the decoded spectrum, any spectral peak exceeding a set threshold within each window is identified as a strong interference peak (i.e., the main lobe energy concentration area of the interference). Erasing this peak can eliminate strong ground interference. However, due to the inherent spectral leakage problem in FFT operations, simply erasing the main peak cannot completely suppress the interference.
[0028] Please see as follows Figure 3 The diagram shows the spectral range of a strong interference signal. The spectral leakage peak of the interference signal may exceed the amplitude of the target signal. If the main lobe and side lobes are not erased simultaneously, the target signal may not be accurately identified. To address this, this application proposes a range-adjustable preamble interference spectral erasure method based on the RSSI of strong ground-based interference signals. This method balances interference suppression with target signal preservation.
[0029] Assume the RSSI of the interfering data packet is The corresponding amplitude is ( Figure 3 middle (as shown in the peak value), the satellite data packet RSSI is To ensure that the power of the interference signal is lower than that of the satellite signal, it is necessary to follow... The power of the interference signal is attenuated, and the corresponding amplitude needs to be reduced to [a certain value]. This is to double the amount needed to eliminate strong interference. Figure 3 According to The range for eliminating strong interference frequencies was determined to be as follows: Based on the above analysis, this application uses a rectangular window to determine the spectrum erasure range of strong interference signals. The reason is that the rectangular window has a narrow roll-off bandwidth, which can effectively reduce the waste of spectrum resources.
[0030] The spectrum function of the rectangular window is as follows: (1) Among them, the peak value of the main lobe The side lobe amplitude is ; Then the sidelobe amplitude attenuation is the main lobe peak value. The positions of the side lobes are: (2) The frequency distance between the sidelobes and the main lobe is further obtained as follows: (3) Set the center frequency of the Preamble portion of the interference signal to be... The design frequency erasure range is... To reduce the amplitude of the side lobes of the covered rectangular window to a minimum. This part ensures that the amplitude of the interference signal is weakened to the target multiple.
[0031] (2) Method for erasing the start delimiter (SFD) of the interference signal frame: The SFD part consists of and 2.25 conjugates The dechirp operation for the SFD part needs to be adjusted to multiply by... The spectrum is then obtained through FFT transformation. In each decoding window, decoding peaks exceeding a threshold are also erased. It's important to note that when processing the last 0.25 chirs of the SFD section, since this part only accumulates the energy of 1 / 4 of a chirp, the decoding result needs to be quadrupled before being removed using a threshold.
[0032] (3) Method for erasing the interference signal payload: Please see as follows Figure 4 The diagram shows the correspondence between the peak values of the interference signal payload spectrum. 4(a) is the modulated upper-chirped waveform, and 4(b) is the demodulated upper-chirped signal. Each chirp in the payload is typically divided into two segments, such as... Figure 4 As shown in (a), there are two corresponding peaks in the decoding result, as follows: Figure 4 As shown in (b), since the two sub-chirps are formed by cyclic shifting of the upchirp, their corresponding frequency difference is equal to the signal bandwidth. If only one peak is erased, the remaining part may still interfere with the detection and decoding of satellite data packets. Therefore, both peaks must be erased simultaneously.
[0033] Figure 4 (b) after segmentation Peak values may be below the threshold and thus missed. Therefore, we will consider the dechirp+FFT results... Partially shifted to the right by one bandwidth and By adding the parts together, the originally scattered peaks are re-superimposed, thus finding the one... Partial peak Then, by using bandwidth relationships, the location was determined. peak .
[0034] Next calculation Regarding the corresponding frequency erasure range, it's important to note that if the same erasure range as the preamble is used, erasing the two peak segments would reduce useful information by half. To address this issue, this application recalculates the frequency erasure range based on the proportion of the two peaks to the complete peak. It is assumed that the amplitude of the first peak is [missing information - likely a percentage] of the complete peak amplitude. The amplitude of the second peak is equal to the amplitude of the complete peak. If the attenuation values are times, then the attenuated amplitudes of the two are respectively Based on this, we calculated the erasure range of the first peak using formulas (1) to (3) respectively. ,in The erasure range of the second peak. ,in .
[0035] (4) Method for setting the critical erase ratio: As the power difference between the satellite and the interference increases, the probability of the satellite signal being erased increases accordingly. This embodiment then quantitatively analyzes the relationship between this power difference and the signal erasure range to set a critical erasure ratio. The specific method is as follows: The parameters of the satellite data packet are set as SF and BW. Let be the sampling rate. Then the number of FFT points corresponding to a single chirp signal in the satellite data packet is... The number of FFT points erased can be calculated as follows: Therefore, the proportion of erased points to the total number of signal points is: (4) When the power difference between the interference signal and the satellite signal As the frequency band increases, the bandwidth masked by the interference signal also expands, causing the number of FFT points to be erased to increase monotonically. It also increases monotonically. In other words, An increase means that the frequency range being erased and the proportion of satellite energy being erased simultaneously rise, reducing the effective components available for decoding and thus causing SER degradation.
[0036] To determine when the algorithm will fail, a critical erase ratio needs to be set. When the erasure ratio exceeds this threshold, the remaining valid signal is insufficient to support reliable decoding, the bit error rate will rapidly deteriorate, and the algorithm will lose its decoding capability. Typically, the failure criterion is that half of the valid components are erased. (At this point, the system has entered a "random guessing" state, and decoding is essentially ineffective.) Taking a typical parameter SF=8 as an example, Substitute the inverse function: (5) That is, if SF is 8 and the power difference between the interference and the satellite reaches 40dB, the algorithm will fail.
[0037] Example 2: Fine-grained estimation of dynamic carrier frequency offset (CFO): Existing terrestrial LoRa packet collision decoding can only provide a fixed CFO estimate, which is clearly insufficient to meet the large-scale CFO estimation requirements caused by the high-speed movement of the aforementioned low-Earth orbit LoRa satellite communication system. To address this issue, this embodiment cleverly utilizes the unique Doppler frequency shift characteristics of low-Earth orbit satellites to propose a fine-grained CFO estimation method. This method achieves symbol-level CFO estimation granularity, significantly improving the accuracy of large-scale CFO estimation in dynamic scenarios of low-Earth orbit LoRa satellite communication systems.
[0038] (1) Detect satellite data packets: In environments with poor signal-to-noise ratios, traditional correlation detection methods struggle to effectively detect satellite data packets. This embodiment employs a multi-data packet aggregation method to achieve effective detection of satellite data packets. For its principle, please refer to [reference needed]. Figure 5 The diagram illustrates satellite data packet preamble aggregation detection. The spectra of detected preambles belonging to the same preamble after dechirp operation and FFT transformation are accumulated. The resulting peak summation is many times greater than the peak value of a single chirp. Because noise has random phase, positive and negative noise peaks cancel each other out during the multi-chirp summation process. Therefore, even when multiple preamble spectrum peaks are accumulated, the accumulated noise peak value will be much lower than the accumulated peak value of the satellite data packet. Please refer to [example diagram]. Figure 6 The diagram shows the Preamble aggregation detection results of satellite data packets. Figure 6 The results are presented as a random sample of the detection results. The left side shows the Preamble spectrum of a single chirp, and the right side shows the Preamble spectrum of multiple chirps aggregated together. Figure 6 The peak value after aggregation is 6 times that of a single Preamble peak. Although the noise amplitude is improved, the amplitude difference between the peak and noise increases from 0.5 to 10, which is obviously very helpful for packet detection.
[0039] However, the aggregation result may still show no obvious peak. This phenomenon usually stems from the receiver hardware failing to adequately compensate for the signal strength of the data packets, especially when the RSSI of the received data packets at the ground station remains around -130dBm, this problem occurs frequently. To address the above issue, this embodiment is equipped with a 15dB antenna gain and a 30dB LNA in the RF front-end to ensure that the RSSI of the received data packets is not lower than -110dBm.
[0040] (2) Predicting the Doppler shift of satellite signals: The frequency offset of satellite data packets mainly consists of two parts: one is the Doppler frequency shift caused by the relative motion between the satellite and the ground station, which is the main component and can reach up to 20kHz; the other is the frequency offset caused by the non-ideal nature of the receiver hardware, which is the minor component and is usually below 100Hz.
[0041] First, the satellite orbital state is modeled using the SGP4 orbital model, and the TLE (Time Limit Expiration) data from two lines of orbital data is used as input to calculate the orbital state at any given time. Position of the next satellite in the True Equator, Mean Equinox (TEME) coordinate system and speed Vector, TEE coordinate system as follows Figure 7 As shown.
[0042] Since the location of ground stations is usually described using the Earth-Centered, Earth-Fixed (ECEF) coordinate system, the ECEF coordinate system is as follows: Figure 8 As shown. Therefore, a coordinate system transformation is required for the satellite state vector in the TEE coordinate system. During the transformation, Greenwich Apparent Sidereal Time (GAST) is introduced to calculate the rotation angle caused by the Earth's rotation. The corresponding rotation matrix is constructed as follows: (6) Based on the above transformations, the satellite's position and velocity vectors in the ECEF coordinate system are obtained as follows: (7) (8) In the formula, This is the Earth's rotational angular velocity vector.
[0043] Then, the latitude, longitude, and altitude of the ground receiver are converted into position vectors in the ECEF coordinate system based on the WGS-84 standard. The parameters of the ellipsoid used are as follows: Equatorial radius: Flatness: First eccentricity squared: Radius of the Mao-You circle: ; The position vector of the receiver in the ECEF coordinate system can be calculated using the following formula: (9) In the formula, This refers to the longitude (in radians) of the ground receiver. This refers to the latitude (in radians) of the ground receiver. The altitude of the ground receiver is in meters.
[0044] Finally, after obtaining the ECEF position of the ground receiver... ECEF position of the satellite Velocity vector Then, the predicted Doppler shift of the satellite signal is calculated using the following formula: (10) In the formula, The wavelength of the satellite signal. For the clock frequency offset between the satellite and the receiver, express In the ECEF coordinate system Components of the axis, express In the ECEF coordinate system Components of the axis, express In the ECEF coordinate system axis, shaft and The components of the axis.
[0045] (3) Roughly estimate the CFO of satellite signals: This embodiment uses the preamble and SFD of satellite data packets to obtain a rough estimate of CFO. The Doppler frequency shift caused by the relative motion between the satellite and the ground station, and the frequency offset caused by receiver hardware non-ideals, are taken as the overall offset and defined as follows: Will be opposite The signal causes a frequency shift in the same direction; in addition, there is a time shift caused by the misalignment of the demodulation window and the satellite data packets, defined as... The signal causes a frequency shift in the opposite direction.
[0046] but The resulting frequency shift is expressed as: (11) In the formula, To represent the Nyquist correlation value, ; as well as The resulting frequency shift is expressed as: (12) From formulas (9) and (10), we can obtain Rough estimate: (13) (14) (4) Dynamic fine-grained estimation of satellite signal CFO: However, the above estimate is for a fixed frequency. In a real LEO communication scenario, the maximum Doppler variation rate of satellite motion is... This caused This is a variable value. This dynamic change in Doppler rate is particularly sensitive at SF=12, where the frequency difference between adjacent LoRa symbols is only 30Hz, and a frequency shift of 2Hz / symbol will result in a positional shift of 1 symbol every 15 symbols. Furthermore, even a small CFO error can affect the height of the decoded peak. Therefore, dynamic fine-grained estimation of the satellite CFO in LEO scenarios is necessary.
[0047] Achieving fine-grained CFO estimation requires meeting two core requirements: first, accurately calculating the Doppler frequency shift rate of each symbol of the satellite signal, defined as the Doppler rate of change (Hz / chirp). This indicates two main objectives: first, to track the dynamic changes in CFO; and second, to accurately obtain the initial CFO value from satellite signals. This provides a benchmark for CFO calculation in the Payload portion. At this point, the CFO value for each chirp in the satellite signal Payload portion is expressed as: (15) In the formula, Indicates the first A symbol.
[0048] Specifically, the satellite Doppler change rate is calculated using the following method. : Based on the predicted Doppler shift of the satellite signal, the Doppler change per second of the satellite is calculated, and then the Doppler change rate of each symbol of the satellite signal is calculated as follows: (16) In the formula, The duration of each symbol, ; It should be noted that the numerator in formula (16) is obtained by subtracting the Doppler frequency shifts of adjacent times, and the subtraction of the two can cancel out the Doppler frequency shift error term.
[0049] And, the initial CFO value is calculated in the following way. : In this embodiment, the CFO value of the last symbol of the preamble is selected as the initial value for fine-grained estimation. The main reason is that the rough estimate is based on this symbol, and this symbol is closest to the payload compared to other preamble symbols, thus having the smallest cumulative error caused by Doppler variations. Theoretically, it can... The preamble is frequency-compensated by iterating through all frequency values across the entire bandwidth, and then the highest FFT peak is found to determine the frequency. However, this method has a high computational cost, making it difficult to meet real-time processing requirements. Therefore, this embodiment introduces a coarse estimation result to narrow the search range, thereby reducing computational complexity.
[0050] Specifically, existing technology only points out There is a certain margin of error, small Errors can also affect the height of the decoding peak. To obtain a more accurate CFO value, this embodiment defines... The search scope is ,in, This is a rough estimate of the error. Using a step size of 1 Hz, in... Perform a search within the search area, using each search frequency. Perform frequency compensation on the last preamble of the data packet and calculate the peak value after compensation. If the current... With If there is a perfect match, the peak value will reach its maximum, and therefore, its corresponding frequency is... .
[0051] Using only a single symbol in the preamble The estimation is susceptible to random noise. Therefore, to combat the impact of random noise on the estimation, this embodiment also incorporates a data packet aggregation approach, combining the last two symbols of the Preamble part... and the first two symbols of the SFD section The FFT peaks of the four symbols are summed to obtain: (17) in, (18) (19) (20) ;(twenty one) ;(twenty two) .(twenty three) Example 3: Multi-feature joint conflict decoding: This addresses the issue of weak ground interference (whose intensity does not exceed a set threshold) after interference cancellation, as well as potential satellite data packet collisions. Based on estimated parameters... In conjunction with the RSSI of received satellite data packets, this embodiment proposes a signal collision decoding method based on multi-feature joint processing.
[0052] (1) Acquisition of target satellite data packet features: Due to the differences in satellite trajectories, data packets from different satellites carry their unique Doppler frequency offset information, and the CFO of satellite data packets mainly comes from the aforementioned Doppler frequency offset. Based on this, this embodiment first extracts the predicted Doppler frequency shift value of the Preamble portion of the target satellite according to formula (10). Then calculate the Preamble portion for all data packets within the receive window. Will with Highest matching Value and corresponding These values together serve as characteristics of the target satellite data packet.
[0053] (2) Single-channel payload matching decoding: Please see as follows Figure 9 The diagram shows a single-channel payload matching and decoding process. Figure 9 (a) is a schematic diagram of collision data packet glass offset compensation, and (b) is a schematic diagram of decoding matching. Since the Preamble portion has consecutive identical upchirps, this embodiment uses the decoding peak value of the last upchirp of the target satellite data packet preamble as the matching benchmark for the peak height of the Payload portion. First, the decoding window is aligned with the last Preamble of the target satellite data packet (red); then, using... Compensate the data packets, from Figure 9 (a) Decoding window As can be seen, after compensation, the Preamble's initial frequency returns to 0; finally, the height of the first peak in the decoding result is extracted and defined as the matching reference, using... Indicates, such as Figure 9 (b) The red peak is shown in the middle.
[0054] Payload matching and decoding: First, using Align the target satellite data packet payload with the decoding window, and then use... Carrier offset compensation is performed for each payload symbol. Figure 9 (a) This represents the compensated decoding window. Subsequently, a dechirp operation and FFT transform are performed on each decoding window to obtain the decoded peak values of the satellite signal and the interference signal, such as... Figure 9 (b) As shown. Finally, the decoded peak value and... Perform a similarity comparison and select the closest match. The peak value is used as the collision decoding result for the current window, such as Figure 9 (b) As shown.
[0055] Please see as follows Figure 10 This is another schematic diagram of single-channel payload matching decoding. It addresses the possibility of multiple... Similar peak results, such as Figure 10 As shown in (a) and (b), this embodiment proposes a sliding window demodulation method. The basic idea is that regardless of whether the decoding window moves to the left or right, it will no longer be aligned with the satellite chirp, thus reducing the peak value of the satellite chirp decoding. Figure 10 The red decoding peaks are shown in (c) and (d). In contrast, other satellite interference or ground interference shows an inverse trend as the decoding window moves left or right. As shown in the gray interference, when the decoding window moves to the left, more signals enter the decoding window, and the decoding peak increases; conversely, the decoding peak decreases. Therefore, the matching decoding scheme given in this embodiment involves moving the decoding window left and right by the same amount of time. Conflicting decoding results showing the same trend will be identified as the target decoding peak.
[0056] The above decoding method can be represented using mathematical functions as follows: use The function calculates the matching benchmark of satellite data packets. : ;(twenty four) use The function extracts peak values from the decoding results: (25) use Function detection and Closest peak: (26) use Function to detect if there are multiple peaks and similar: (27) If so, then use The function moves the decoding window to the left: (28) Reuse The function moves the decoding window to the right: (29) use The function detects peak values whose peak height decreases after two window moves. (30) (3) Cross-channel payload matching and decoding: Please see as follows Figure 11 The diagram shown illustrates cross-channel payload matching and decoding.
[0057] When cross-channel interference exists, this embodiment adopts a solution based on two-parameter decoding: using standard The decoder performs initial signal processing (such as...) Figure 11 (As shown by the red curve); using a curve with the same slope but larger SF and BW parameters. The decoder performs secondary processing (such as...) Figure 11 (As shown by the blue curve); By comparing the two decoding results, it was found that the interference signal always appeared stably. The frequency points and peak heights are proportionally amplified, while the target signal exhibits a unique "frequency shift characteristic"—its peak position shifts regularly but its amplitude remains stable. This differentiated response characteristic allows us to establish reliable identification criteria: when a signal that simultaneously satisfies both "peak amplitude stability" and "fixed frequency shift pattern" is detected, it can be accurately identified as the target signal.
[0058] The above content is only a preferred embodiment of the present invention. For those skilled in the art, many changes can be made in the specific implementation and application scope based on the ideas of the present invention. As long as these changes do not depart from the concept of the present invention, they all fall within the protection scope of the present invention.
Claims
1. A method for anti-interference decoding of LoRa signals for LPWAN satellite-ground fusion, characterized in that, The method includes: Based on the fact that the RSSI of the ground station signal is much higher than that of the satellite signal, the interference signal from the ground station in the mixed signal is identified. The Preamble, SFD and Payload parts of the interference signal are demodulated and the corresponding frequency range is erased to filter out the interference signal. A multi-data packet aggregation method is used to accumulate the spectra of detected Preambles that have undergone dechirp operation and FFT transformation. Based on the characteristic that the peak superposition result is many times greater than the peak value of a single chirp, satellite data packets are detected. A rough estimate of CFO is made based on the predicted Doppler frequency shift value of the Preamble part of the satellite data packet. Then, a fine-grained estimate of dynamic CFO is made based on the last two symbols of the Preamble part of the satellite data packet and the first two symbols of the SFD part, combined with the data packet aggregation approach. The dynamic CFO fine-grained estimation result and the time offset value of the satellite data packet are used together as the features of the target satellite data packet to decode the payload part of the satellite data packet.
2. The method as described in claim 1, characterized in that, The preamble portion of the interference signal is demodulated using the following method: Employ a signal with a chirp slope consistent with satellite data packets and larger SF and BW parameters. Matching with the preamble portion of the interference signal to detect whether the interference signal is inter-channel interference; If a match is successful, it indicates cross-channel interference. Then, based on the larger SF and BW parameters of the successful match, a dechirp operation and FFT transform are performed to obtain the spectrum of the Preamble part. If the match fails, it means that it is not cross-channel interference. In this case, the original SF and BW parameters of the satellite data packet are used to perform the dechirp operation and FFT transformation. The interference spectrum erasure range of the Preamble section is determined using the following method: Assume the RSSI of the interfering data packet is The corresponding amplitude is The satellite data packet RSSI is To ensure that the power of the interference signal is lower than that of the satellite signal, it is necessary to follow... The power of the interference signal is attenuated, and the corresponding amplitude attenuation factor is... To meet the need to eliminate strong interference; The spectrum function of the rectangular window is as follows: ； Among them, the peak value of the main lobe The side lobe amplitude is ; Then the sidelobe amplitude attenuation is the main lobe peak value. The position of the side lobe is ; The frequency distance between the sidelobes and the main lobe was further obtained as follows: ; Set the center frequency of the Preamble portion of the interference signal to be... The design frequency erasure range is... To reduce the amplitude of the side lobes of the covered rectangular window to a minimum. This part ensures that the amplitude of the interference signal is weakened to the target multiple.
3. The method as described in claim 1, characterized in that, The SFD portion of the interference signal is demodulated and the corresponding spectral range is erased using the following method: Because the SDF portion of the interference signal consists of and 2.25 conjugates Therefore, the dechirp operation on the SDF part is multiplied by... Then, the spectrum of the SDF part is obtained through FFT transformation; Within each decoding window, peaks exceeding a threshold are erased. In the last 0.25 chirs of the SDF section, since it has only accumulated the energy of 1 / 4 chirp, its spectrum needs to be quadrupled before being removed by the threshold.
4. The method as described in claim 1, characterized in that, The payload portion of the interference signal is demodulated and the corresponding spectral range is erased using the following method: Locating the two peaks in the payload portion To prevent missed detections due to peak values below a threshold, the design incorporates the spectrum of the payload portion after the dechirp operation and FFT transformation. Partially shifted to the right by one bandwidth and Partial addition is performed to re-superimpose the originally dispersed peaks, thereby obtaining... Partial peak Then, by using bandwidth relationships, the location is determined. peak ; calculate Corresponding frequency erasure range: To prevent useful information from being erased, based on two peak values... The frequency erasure range of the payload portion is recalculated based on the proportion of the full peak value. The specific method is as follows: Assume that the amplitude of the first peak value is equal to the amplitude of the full peak value. times The amplitude of the second peak is equal to the amplitude of the complete peak. If the amplitudes of the two peaks are multiplied by a factor of 1, then the amplitudes after the two peaks decay are and Then, based on the rectangular window function, the erasure range of the first peak is calculated as follows: ,in The erasure range of the second peak ,in .
5. The method as described in claim 1, characterized in that, The Doppler shift of the Preamble portion of satellite data packets can be predicted using the following method: The SGP4 orbital model is used to model the satellite's orbital state, and two lines of orbital data (TLE) are used as input to calculate the orbital state at any given time. The position vector of the satellite in the TEE coordinate system and speed Vector; Introducing Greenwich Mean Time (GMT) to calculate the rotation angle caused by Earth's rotation. And construct its corresponding rotation matrix as follows: ； Based on the rotation matrix, the satellite state vector in the TEE coordinate system is transformed to obtain the satellite position vector in the ECEF coordinate system. With velocity vector They are respectively: ； ； In the formula, This is the vector of Earth's rotational angular velocity; The latitude, longitude, and altitude of the ground receiver are converted into a position vector in the ECEF coordinate system based on the WGS-84 standard. ; The ellipsoid parameters are set as follows: Equatorial radius: Flatness: First eccentricity squared: Radius of the Mao-You circle: ; The position vector of the receiver in the ECEF coordinate system can be calculated using the following formula: ； In the formula, For the longitude of the ground receiver, Latitude of the ground receiver The altitude of the ground receiver; Then, the predicted Doppler shift value of the Preamble portion of the satellite signal is calculated using the following formula. : ； In the formula, The wavelength of the satellite signal. For the clock frequency offset between the satellite and the receiver, express In the ECEF coordinate system axis, shaft and Components of the axis, express In the ECEF coordinate system axis, shaft and Components of the axis, express In the ECEF coordinate system axis, shaft and The components of the axis.
6. The method as described in claim 1, characterized in that, A rough estimate of CFO can be made using the following methods: The Doppler frequency shift caused by the relative motion between the satellite and the ground station, and the frequency shift caused by receiver hardware non-ideals, are taken as the total offset and defined as follows: Will to The frequency shift caused by the signal in the same direction, and the time shift caused by the misalignment of the demodulation window with the LoRa data packet are defined as follows: The signal causes a frequency shift in the opposite direction; but The resulting frequency shift is expressed as: ； In the formula, To represent the Nyquist correlation value, ; as well as The resulting frequency shift is expressed as: ； Further obtain The rough estimates are as follows: ； 。 7. The method as described in claim 1, characterized in that, Dynamic CFO fine-grained estimation is performed using the following methods: Based on the predicted Doppler shift of the satellite signal, the Doppler change per second of the satellite is calculated, and then the Doppler change rate of each symbol of the satellite signal is calculated as follows: ； In the formula, The duration of each symbol, ; Define the initial CFO value of the satellite signal. The search scope is ,in, This is a rough estimate of the error. Based on the data packet aggregation approach, each search frequency is used The last two symbols of the preamble portion of the satellite signal and the first two symbols of the SFD section Perform frequency compensation and obtain the compensated FFT peak accumulation value. : ； in, ; ； ； ； If the current With If there is a perfect match, the peak value will reach its maximum, and therefore, its corresponding frequency is... : ； At this point, the CFO value of each Chirp in the satellite signal payload is expressed as: ； In the formula, Indicates the first A symbol.
8. The method as described in claim 1, characterized in that, When there is no cross-channel interference, the satellite data packet payload is decoded using single-channel matching in the following manner: The peak value of the last symbol in the Preamble portion of the target satellite data packet is used as the matching benchmark for the peak height of the Payload portion. The specific method is as follows: Align the decoding window with the last Preamble portion of the target satellite data packet; use The target satellite data packets are compensated so that the compensated Preamble starting frequency returns to 0; the first peak height in the decoding result is extracted and defined as the matching reference. Payload matching and decoding: utilizing Align the target satellite data packet payload with the decoding window; use Carrier offset compensation is performed on each payload symbol; dechirp operation and FFT transformation are performed on each decoding window to obtain the decoded peak values of the satellite signal and the interference signal; The similarity between the decoded peak and the matching benchmark is compared, and the peak closest to the matching benchmark is selected as the conflict decoding result of the current window. When there are multiple peak results that are close to the matching benchmark, the decoding window is moved to the left and right by the same amount of time. Conflict decoding results that show the same trend will be identified as the target decoded peak.
9. The method as described in claim 1, characterized in that, When cross-channel interference exists, the payload portion is cross-channel matched and decoded in the following way: Use of standards The decoder performs initial signal processing; it uses a chirp slope consistent with satellite data packets and larger SF and BW parameters. The decoder performs secondary processing; by comparing the two decoding results, the target signal is identified based on the signal characteristic that the peak position shifts regularly but the amplitude remains stable.