Low-power-consumption cross-channel communication conflict separation method based on opposite non-uniform scaling
Through opposing non-uniform scaling technology, interfering signals across channels in the LoRa network are identified and removed, which solves the problem of low decoding efficiency in the prior art and improves communication throughput.
Patent Information
- Application Number
- CN202510377536.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-28
- Publication Date
- 2025-06-27
AI Technical Summary
The existing LoRa collision decoding method cannot accurately identify interference between different BW and different SF in the cross-channel situation, resulting in a reduced decoding efficiency.
Using a low-power cross-channel communication collision separation method based on opposite non-uniform scaling, interference signals across orthogonal channels are identified and removed through preamble detection, peak extraction and symbol operation steps.
It improves communication throughput, solves the problem that existing decoding methods cannot accurately identify cross-channel interference, and improves decoding efficiency.
Smart Images
Figure CN120223494A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical fields of LoRa collision decoding and Internet of Things applications, and specifically to a low-power cross-channel communication conflict separation method based on opposite non-uniform scaling. Background Art
[0002] In a general LoRa network, there are often hundreds or thousands of devices connected to the same LoRa gateway. And due to the requirement of the LoRa network based on the Aloha protocol not to monitor the channel state, there is a possibility of signal conflicts at the gateway. The latest collision decoding method is to use the unique time domain or frequency domain of each Chirp segment of the LoRa data packet to distinguish the target signal and interference signal in a segmented window. For example, some methods find the frequency domain bin value of each Chirp segment by comparing the FFT peaks between adjacent windows for decoding, that is, by performing a dechirp operation on the signal in each demodulation window and performing a fast Fourier transform to obtain the frequency domain peak, and classifying according to the peak information in various states. However, these existing collision decoding methods cannot correctly distinguish the interference of different BW and different SF in the case of cross-channel, resulting in a decrease in decoding efficiency. Summary of the Invention
[0003] In order to overcome the problem that the existing decoding method cannot accurately identify the existence of cross-channel interference, the present invention proposes a low-power cross-channel communication conflict separation method based on opposite non-uniform scaling, which improves the communication throughput.
[0004] The present invention is implemented by adopting the following technical solutions: A low-power cross-channel communication conflict separation method based on opposite non-uniform scaling, the method at least includes the following steps:
[0005] Step 1: Preamble detection: When the received signal fails to be decoded by the LoRa standard decoder, enter the conflict recognition mode; configure the correlation standard up-chirp according to the parameters of the target signal, and move it in the received signal to generate multiple peaks, and identify the correlation of each preamble segment by comparing these peaks, so as to detect the starting position of the target signal preamble; after detecting the starting position of the target signal preamble, move a certain number of positions to accurately find the starting point of the target signal data part;
[0006] Step 2: Peak extraction: Divide windows according to the starting point of the target signal data part, perform opposite non-uniform scaling operations on the signals of each window, and utilize the characteristic that the peak positions corresponding to the FFT results are the same after two opposite non-uniform scaling operations, and also combine the phase rotation of the Fourier transform to identify the continuous symbol duration with unique characteristics, that is, identify the Chirp signal of the target signal;
[0007] Step 3: Symbol operation: First, the center frequency offset and related operations are eliminated through the characteristics of the preamble and SFD in the LoRa signal. After that, the target signal is extracted based on the Chirp signal of the target signal within the window, and after each received signal detection, the separated target signal is excluded and the next operation is performed, so as to classify the symbols into multiple clusters corresponding to each conflicting data packet.
[0008] For the above method for separating conflicts in low-power cross-channel communication based on opposing non-uniform scaling, the specific steps of preamble detection in Step 1 are as follows:
[0009] Step 11: Data preprocessing, including but not limited to removing noise and other interferences outside the standard signal segment in the received signal;
[0010] Step 12: Detect the starting position of the target signal preamble using the signal amplitude difference;
[0011] Step 13: Design an up-chirp with configuration parameters equal to those of the target signal, move it in the received signal and perform a correlation comparison with the preamble of the received signal to generate correlation peaks. The detailed steps of the correlation comparison include:
[0012] 1) Define the signal: Assume a standard up-chirp, which is used for standard comparison;
[0013] 2) Correlation comparison. When the standard up-chirp is aligned with n up-chirps in the target signal, N c correlation peaks will be generated. Since the distance between each chirp in the LoRa signal is equal, the interval between two adjacent correlation peaks is equal, which is equal to the number of samples of each Chirp in the data part of the target signal, that is, N. Therefore, the correlation recognition work can be expressed as: where s is the starting position of the preamble of the target signal, and δ represents the minimum correlation requirement.
[0014] For the above method for separating conflicts in low-power cross-channel communication based on opposing non-uniform scaling, the specific steps of Step 2 are as follows:
[0015] Step 21: Window division and signal processing. First, divide the window according to the Payload start point and perform opposing non-stationary scaling operations on the signals in each window;
[0016] Step 22: Window signal demodulation. After using two opposing non-stationary scaling operations, by using the characteristic that the corresponding peak positions in the FFT result are the same, combined with the phase rotation in the Fourier transform, identify the continuous symbol duration with unique characteristics.
[0017] The above-mentioned low-power cross-channel communication conflict separation method based on opposite non-uniform scaling, step 2 further includes step 23: patching operation, based on the unique features of adjacent channels, identifying the target signal from multiple signals similar to the target symbol and filling the window. The signals across orthogonal channels, that is, interference signals, due to different parameter configurations from the target signal, so a part of the interference signal will appear on the adjacent channels. Therefore, demodulation operation is performed on the positions on the adjacent channels. If a peak appears at the same position, it indicates an interference signal, and at this time, the interference signal can be excluded.
[0018] The above-mentioned low-power cross-channel communication conflict separation method based on opposite non-uniform scaling, the opposite non-uniform scaling is to perform opposite non-uniform transformation on the standard downchirp. The opposite non-uniform transformation is to multiply the standard downchirp and two linear functions with opposite slopes respectively. During demodulation, the signal in the window is multiplied by the downchirp after opposite non-uniform transformation, and then FFT is performed to obtain a peak at the center frequency.
[0019] The specific steps of step 3 of the above-mentioned low-power cross-channel communication conflict separation method based on opposite non-uniform scaling are as follows:
[0020] Step 31: Target signal marking and symbol classification, extracting the target signal according to the Chirp signal of the target signal in the window. After each received signal detection, the separated target signal is excluded, and the next operation is performed. The Chirp of the target signal is classified into multiple target signals corresponding to each conflicting data packet;
[0021] Step 32: Central frequency offset elimination, before decoding the data packet, eliminating the central frequency offset through the preamble and SFD characteristics and performing relevant operations f offset represents the central frequency offset, up freq represents the frequency of up-chirp, down freq represents the frequency of down-chirp, nsamp is the number of sampling points of a Chirp, t offset is the time offset, round is a function;
[0022] Step 33: Decoding process, inputting the target signal into the standard decoder for decoding.
[0023] The method of the present invention adopts opposite non-uniform scaling during demodulation, which can remove the interference signals across orthogonal channels and solves the problem that the existing decoding methods cannot accurately identify the existence of cross-channel interference. Brief Description of the Drawings
[0024] Figure 1It is a flow block diagram of a cross-orthogonal channel conflict method based on opposite non-uniform scaling separation.
[0025] Figure 2 It is a schematic diagram of the opposite non-uniform scaling processing process of a cross-orthogonal channel conflict method based on opposite non-uniform scaling separation.
[0026] Figure 3 It is a schematic diagram of the patch design of a cross-orthogonal channel conflict method based on opposite non-uniform scaling separation. Specific implementation manners
[0027] A low-power cross-channel communication conflict separation method based on opposite non-uniform scaling, the specific steps are as follows:
[0028] Step 1: Preamble detection: When the received signal fails to be decoded in the LoRa standard decoder, it indicates that a conflict occurs and enters the conflict recognition mode. Configure the correlation standard up-chirp according to the parameters of the target signal, move it in the received signal to generate multiple peaks, and identify the correlation of each preamble by comparing these peaks, so as to detect the starting position of the target signal preamble. After detecting the starting position of the target signal preamble, according to the LoRa physical layer coding rule, move several positions to accurately find the starting point of the target signal data part (Payload starting point). The specific steps are as follows:
[0029] Step 11: Data preprocessing, including but not limited to removing noise and other interferences outside the standard signal segment in the received signal.
[0030] Step 12: Detect the starting position of the target signal preamble by using the signal amplitude difference.
[0031] Step 13: Design an up-chirp with the same configuration parameters as the target signal, move it in the received signal and compare its correlation with the preamble of the received signal to generate correlation peaks. The detailed steps of the correlation comparison include:
[0032] 1) Define the signal: Set a standard uplink linear frequency modulation pulse signal (up-chirp), which is used for standard comparison.
[0033] 2) Correlation comparison, when the standard up-chirp is aligned with n up-chirps in the target signal, N c correlation peaks will be generated. The distance between each chirp of the LoRa signal is equal, so the interval between two adjacent correlation peaks is equal, equal to the number of samples of each Chirp in the Payload, that is, N. Therefore, the correlation recognition work can be expressed as:
[0034]
[0035] where s is the starting position of the target signal preamble, and δ represents the minimum correlation requirement.
[0036] Step 2: Peak extraction: Window division is performed according to the Payload start point, and an anti-uniform scaling operation is carried out on the signals in each window. By utilizing the characteristic that the corresponding bin values (peaks) in the FFT results are in the same position after two anti-uniform scalings, and also combining the phase rotation of the Fourier transform, the continuous symbol duration with unique characteristics is identified. The specific steps are as follows:
[0037] Step 21: Window division and signal processing. The present invention first performs window division according to the Payload start point, and an anti-stationary scaling operation is carried out on the signals in each window.
[0038]
[0039] The above two formulas are respectively the time-domain representations of the target signal Chirp and the interference signal Chirp, and C T (t) represents the time-domain representation of the target signal Chirp, which is obtained by simplification, where f t is the initial frequency of the Chirp, t is the time variable, and C(t) is a standard up-chirp, whose value is equal to where |k| = BW / N, f0 = -BW / 2, and BW represents the signal bandwidth; C I (t) represents the time-domain representation of the interference signal Chirp. Since it is not aligned with the window, it appears as a delay Δt with respect to the signal within the window. Therefore, it needs to be multiplied by C(t + Δt), and after simplification, C I (t) is obtained.
[0040] Step 22: Window signal demodulation. By using the characteristic that the corresponding bin values (peaks) in the FFT results are in the same position after two anti-uniform scaling operations, and combining the phase rotation in the Fourier transform, the continuous symbol duration with unique characteristics is identified, that is, the Chirp signal of the target signal is identified. Anti-uniform scaling is an anti-uniform transformation of the standard downchirp. The anti-uniform transformation is to multiply the standard downchirp by two linear functions with opposite slopes respectively. During demodulation, the signal within the window is multiplied by the downchirp after the anti-uniform transformation, and then an FFT is performed to obtain a peak at the center frequency.
[0041] Step 23: Patching operation. Based on the unique characteristics of adjacent channels, the present invention identifies the target signal from multiple signals that are similar to the target symbol and fill the window, thereby improving the fault tolerance of the system.
[0042] Step 3: Symbol operation: First, the center frequency offset (CFO) and related operations are eliminated through the characteristics of the preamble and SFD. After that, the target signal is extracted according to the Chirp of the target signal within the window, and after each received signal detection, the separated target signal is excluded and the next operation is performed, so as to classify the Chirp of the target signal into multiple target signals corresponding to each conflicting data packet.
[0043] Step 31: Target signal marking and symbol classification. The target signal is extracted according to the Chirp of the target signal within the window. After each target data packet detection, the separated target signal is excluded and the next operation is performed to classify the Chirp of the target signal into multiple target functions corresponding to each conflicting data packet.
[0044] Step 32: Center frequency offset elimination. Before decoding the data packet, the present invention eliminates the center frequency offset (CFO) through the characteristics of the preamble and SFD and performs related operations.
[0045]
[0046] f offset represents the center frequency offset, up freq represents the frequency of the up-chirp, down freq represents the frequency of the down-chirp, nsamp is the number of sampling points of a Chirp, t offset is the time offset, round is a function.
[0047] Step 33: Decoding process. The target signal is input into a standard decoder for decoding.
[0048] The above embodiments are only used to illustrate rather than limit the technical solutions described in the present invention; therefore, although this specification has described the present invention in detail with reference to the above embodiments, those of ordinary skill in the art should understand that the present invention can still be modified or equivalently replaced; and all technical solutions and their improvements that do not depart from the spirit and scope of the present invention should be covered by the scope of the claims of the present invention.
Claims
1. A low-power cross-channel communication conflict separation method based on opposing non-uniform scaling, characterized by: The method comprises the following steps: Step 1: Preamble detection: When the received signal fails to be decoded in the LoRa standard decoder, it enters the conflict identification mode; configure the correlation standard up-chirp according to the parameters of the target signal, and move it in the received signal to generate multiple peaks. By comparing these peaks, the correlation of each preamble segment is identified, thereby detecting the starting position of the target signal preamble; After detecting the starting position of the target signal preamble code, move a number of positions to accurately find the starting point of the target signal data portion; Step 2: Peak extraction: Divide the window according to the starting point of the target signal data part, perform an opposing non-uniform scaling operation on the signal in each window, and use the characteristic that the FFT results after two opposing non-uniform scaling operations have the same corresponding peak position. At the same time, combined with the phase rotation of the Fourier transform, identify the continuous symbol duration with unique characteristics, that is, identify the Chirp signal of the target signal. Step 3: Symbol operation: First, the center frequency offset and related operations are eliminated through the characteristics of the preamble code and SFD in the LoRa signal. Then, the target signal is extracted according to the Chirp signal of the target signal in the window, and the next operation is performed after excluding the separated target signal after each received signal detection, so as to classify the symbol into multiple target signals corresponding to each conflicting data packet.
2. The low-power cross-channel communication conflict separation method based on opposing non-uniform scaling according to claim 1, characterized in that: The specific steps of preamble detection in step 1 are as follows: Step 11: Data preprocessing, including but not limited to removing noise and other interference outside the standard signal segment in the received signal; Step 12: Detect the starting position of the target signal preamble code by using the signal amplitude difference; Step 13: Design an up-chirp with the same configuration parameters as the target signal, move it in the received signal and compare the correlation with the preamble of the received signal to generate a correlation peak. The detailed steps of correlation comparison include: 1) Define the signal: Set a standard up-chirp, which is used for standard comparison; 2) Correlation comparison: When the standard up-chirp is aligned with the n up-chirps in the target signal, N c There are correlation peaks, and the distance between each chirp of the LoRa signal is equal, so the interval between two adjacent correlation peaks is equal, which is equal to the number of samples of each Chirp in the target signal data part, that is, N, so the correlation identification work can be expressed as: Where s is the starting position of the target signal preamble, and δ represents the minimum correlation requirement.
3. The low-power cross-channel communication conflict separation method based on opposing non-uniform scaling according to claim 2, characterized in that: The specific steps of step 2 are as follows: Step 21: Window division and signal processing, firstly, the window is divided according to the starting point of the target signal data part, and the signal of each window is subjected to an opposing non-stationary scaling operation; Step 22: Window signal demodulation, using the same characteristics of corresponding peak positions in the FFT results after two opposing non-stationary scaling operations, combined with the phase rotation in the Fourier transform, to identify continuous symbol durations with unique characteristics.
4. The low-power cross-channel communication conflict separation method based on opposing non-uniform scaling according to claim 3, characterized in that: Step 2 also includes step 23: a patch operation, which identifies a target signal from a plurality of signals that are similar to the target symbol and fill the window based on unique features of adjacent channels.
5. A low-power cross-channel communication conflict separation method based on opposing non-uniform scaling according to claim 3 or 4, characterized in that: Opposing non-uniform scaling is to perform an opposing non-uniform transformation on the standard downchirp. The opposing non-uniform transformation is to multiply the standard downchirp and two linear functions with opposite slopes respectively. During demodulation, the signal in the window is multiplied by the downchirp after the opposing non-uniform transformation, and then FFT is performed to obtain a peak value at the center frequency.
6. A low-power cross-channel communication conflict separation method based on opposing non-uniform scaling according to claim 3 or 4, characterized in that: The specific steps of step 3 are as follows: Step 31: target signal marking and symbol classification, extracting the target signal according to the Chirp signal of the target signal in the window, excluding the separated target signal after each received signal detection, and proceeding to the next operation, classifying the Chirp signal of the target signal into multiple target signals corresponding to each conflicting data packet; Step 32: Center frequency offset elimination: Before decoding the data packet, the center frequency offset is eliminated and related operations are performed through the preamble and SFD characteristics. f offset It indicates the center frequency deviation, up freq Indicates the frequency of up-chirp, down freq represents the frequency of down-chirp, nsamp is the number of sampling points of a chirp, t offset is the time offset, round is a function; Step 33: Decoding process: input the target signal into a standard decoder for decoding.