Ultra-narrowband internet of things unified demodulation system and method

Abstract

The application discloses a kind of ultra-narrow-band Internet of Things unified demodulation system and method, set up pre-RSSI power grading pre-judgment link before demodulation, first carry out SNR or RSSI detection to signal after pre-processing, signal is divided into time domain direct solution mode, time-frequency holographic layered demodulation mode and ultra-weak signal spectrum noise reduction mode three grades of working mode according to detection result;In the tree-shaped subchannel decomposition process, after completing each layer decomposition, each subchannel can be judged according to the characteristics of pure noise subchannel, and the subsequent tree-shaped subchannel decomposition operation of the subchannel judged as pure noise is stopped;At the same time, based on the general bottom physical principle of spread spectrum despreading equivalent ultra-narrow-band MFSK, the general baseband architecture of ultra-narrow-band MFSK and all kinds of Chirp spread spectrum integration is constructed, the use area and power consumption of FFT IP core are avoided, the expensive crystal oscillator commonly used in ultra-narrow-band communication system is avoided, and the tree-shaped subchannel decomposition technology used by the application is also easy to combine with DPLL, to form the comprehensive effect of coherent demodulation.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, specifically to an ultra-narrowband Internet of Things unified demodulation system and method. Background Technology

[0002] In ultra-narrowband IoT communication systems, terminal nodes typically use low-power modulation methods such as MFSK modulation or Chirp spread spectrum to transmit data. The receiving end needs to demodulate the received signal to recover the original data. In existing receiver solutions, regardless of the received signal strength, a complete demodulation link is initiated uniformly, sequentially executing all processing steps such as down-conversion, analog-to-digital conversion, spectrum analysis, correlation operations, and decoding. In the sub-channel decomposition stage, regardless of whether the sub-channel carries a signal, complex mixing, filtering, and decimation operations are performed on all sub-channels level by level until all levels of decomposition and frequency point determination are completed. At the same time, existing FFT frequency point detection schemes are fixed-block operations, requiring the buffering of entire frames of multi-symbol data. The computing power cannot be pruned as needed, and full-load operation is maintained regardless of signal quality. In traditional solutions, the hardware architectures of ultra-narrowband MFSK baseband and Chirp spread spectrum baseband are fragmented, and the equipment cannot be compatible with both systems simultaneously, resulting in poor networking flexibility and low hardware reuse rate.

[0003] However, in real-world ultra-narrowband IoT deployment scenarios, a large number of terminal nodes are in a state of extremely weak or even no signal, and only a few sub-channels carry valid signals within the same receiving window. Existing solutions still perform tree-structured sub-channel decomposition operations at all levels for purely noisy sub-channels with no signal, and still initiate the same full-link demodulation process as the time-domain direct demodulation mode for weak signals, without a mechanism to cut off computing power in advance, causing the receiver to continuously consume computing power on invalid data; at the same time, the baseband architectures of the two systems are independent and cannot be used interchangeably, resulting in high chip costs and overall power consumption that cannot meet the battery life requirements of ultra-low power IoT terminals, which seriously restricts the large-scale civilian and industrial adoption of ultra-high sensitivity ultra-narrowband communication. Summary of the Invention

[0004] The present invention aims to provide an ultra-narrowband Internet of Things unified demodulation system and method, which realizes unified processing of multiple modes, thereby reducing the computational overhead of the receiver on invalid signals, and reducing the power consumption and hardware cost of the whole machine.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a unified demodulation method for ultra-narrowband Internet of Things, comprising:

[0006] The received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal is preprocessed, wherein the Chirp spread spectrum signal is compressed into a single-frequency sinusoidal carrier signal by Dechirp processing, so that the effective signal frequency point after compression uniquely corresponds to one of multiple discrete sub-bands;

[0007] The preprocessed signal is subjected to SNR or RSSI received signal strength detection, and three working modes are divided according to the detection results.

[0008] According to the three working modes, the corresponding demodulation mode is selected for demodulation processing. The three working modes include time-domain direct demodulation mode, time-frequency holographic layered demodulation mode, and ultra-weak signal spectrum noise reduction mode.

[0009] In the time-frequency holographic layered demodulation mode or the ultra-weak signal spectrum noise reduction mode, the binary, quartic, N-part or hybrid tree-shaped sub-channel decomposition module is activated. The sub-channels are decomposed at each layer through complex mixing, filtering and decimation. The pre-processed signal is then divided into layers of spectrum to obtain the sub-channels with signals.

[0010] After each layer of tree-shaped sub-channel decomposition, each sub-channel is judged based on the characteristics of pure noise sub-channels, and the sub-channels judged as pure noise are stopped from subsequent tree-shaped sub-channel decomposition operations.

[0011] Based on the sub-channel with signal, complete the ultra-narrowband frequency point decision and output the demodulated data.

[0012] Preferably, the step of performing SNR or RSSI received signal strength detection includes:

[0013] The amplitude of the preprocessed signal is calculated using a simplified algorithm for calculating amplitude using Cordic.

[0014] The calculation formula for the simplified algorithm is as follows: Where |I| is the absolute value of the in-phase branch and |Q| is the absolute value of the orthogonal branch;

[0015] Based on the calculated amplitude Amp, calculate the RSSI value, where RSSI = Amp * reference - link gain, and the reference is the power value corresponding to ADC Vref;

[0016] The SNR value is calculated based on the ratio of the average value to the variance of the calculated amplitude Amp.

[0017] Based on the calculated RSSI and SNR values, the preprocessed signal is compared with the preset power threshold and signal-to-noise ratio threshold, and then divided into three working modes.

[0018] Preferably, the time-domain direct solution mode includes:

[0019] When the RSSI and SNR meet the first preset condition, the time domain direct decoding mode is selected, all the tree sub-channel decomposition modules are bypassed, and clock-gated sleep is performed on the tree module;

[0020] The on-chip Cordic arithmetic unit is invoked to solve the instantaneous phase of the preprocessed signal point by point.

[0021] Based on the solved instantaneous phase, instantaneous frequency discrimination is achieved through phase difference between adjacent sampling points, where... , The sampling interval is... The phase difference between adjacent sample points;

[0022] Based on the frequency value obtained from instantaneous frequency discrimination, the discrete frequency point position is located, and decoding is completed.

[0023] Preferably, the time-frequency holographic layered demodulation mode includes:

[0024] When the RSSI and SNR meet the second preset condition, the time-frequency holographic hierarchical demodulation mode is selected, and the binary, quartic, N-part, or hybrid sub-tree subchannel decomposition module is started.

[0025] For the initiated tree-shaped sub-channel decomposition module, complex mixing, filtering and decimation operations are performed level by level to decompose each sub-channel at each level;

[0026] For each sub-channel after decomposition at each layer, the characteristics of the pure noise sub-channel are used for determination. The characteristics of the pure noise sub-channel include: the time-domain PAPR is greater than that of the sub-channel with signal, the average power is less than the noise power spectral density multiplied by the bandwidth, the pure noise sub-channel shows a 3dB decrease per layer when decomposed into a bi-tree sub-channel or a 6dB decrease per layer when decomposed into a quaternary tree sub-channel, and the SNR is less than 0dB.

[0027] Based on the judgment results, the subsequent tree-shaped subchannel decomposition operation is stopped for subchannels that are judged as pure noise; the subchannels that are not judged as pure noise are saved and the tree-shaped subchannel decomposition operation is continued when the judgment conditions are met in the future, or the N subchannels with the highest average power ranking are saved and the operation is continued.

[0028] Preferably, the ultra-weak signal spectral noise reduction mode includes:

[0029] When RSSI and SNR meet the third preset condition, select the ultra-weak signal spectrum noise reduction mode and turn off all phase frequency discrimination modules.

[0030] For signals with phase discrimination disabled, perform binary, quartic, N-part, or hybrid tree-structured channel decomposition starting from the beginning;

[0031] In the hierarchical tree-shaped channel decomposition process, branch selection is performed based on branch energy characteristics and time-domain waveform fluctuation characteristics. The branch energy characteristics include average power, which is reduced by 3dB per layer for binary decomposition, 6dB per layer for quartic decomposition, and 10*log10(N)dB per layer for N-decomposition. The time-domain waveform fluctuation characteristics include SNR and PAPR.

[0032] Based on the result of the branch selection, a sub-channel with signal is obtained; or it is determined that there is no signal in the current frame.

[0033] Preferably, the activation of the binary, quartic, N-part, or hybrid tree-structured subchannel decomposition module decomposes each subchannel at each level through complex mixing, filtering, and decimation, including:

[0034] The digital frequency circle of the preprocessed signal is decomposed into four parts at the first level, and four mutually orthogonal sub-bands without spectral overlap are formed by coarse frequency shifting of ±π / 4.

[0035] Each sub-band after the first-level quartic decomposition is further decomposed into a second-level quartic decomposition. The sub-bands are further subdivided by ±π / 8 fine frequency shifting so that the effective signal and the image interference frequency distance is 3π / 4.

[0036] In each decomposition stage, the signal after complex mixing is filtered by a half-band low-pass filter, the number of taps of which is configurable.

[0037] The filtered signal is decimated using a streaming point-by-point processing method, repeating complex mixing, filtering, and decimation operations step by step until all sub-channels are decomposed.

[0038] Preferably, the number of taps in the half-band low-pass filter can be configured to include:

[0039] Ultra-low power consumption and ultra-high sensitivity settings: The last stage uses Haar (i.e., 1 tap), the second to last HBF uses 3 taps, the third to last HBF uses 7 taps, the fourth to last HBF uses 15 taps, and all other HBFs use 15 taps.

[0040] General balance settings: The last level uses Haar (i.e., 1 tap), the second to last HBF uses 3 taps, the third to last HBF uses 7 taps, the fourth to last HBF uses 15 taps, and the remaining HBFs all use 31 taps.

[0041] Ultra-high anti-interference networking level: The last level uses Haar (i.e., 1 tap), the second to last HBF uses 3 taps, the third to last HBF uses 7 taps, the fourth to last HBF uses 15 taps, the fifth to last HBF uses 31 taps, and the remaining HBFs all use 63 taps.

[0042] Preferably, the step of initiating a binary, quartic, N-part, or hybrid tree-structured subchannel decomposition module, which decomposes each subchannel at each level through complex mixing, filtering, and decimation, further includes:

[0043] For the data generated by multi-rate interpolation in each sub-channel at each level through complex mixing, filtering and decimation, an in-situ multiplexing operation mechanism with frequency domain read-only synchronization and time domain sliding coverage is adopted, eliminating the independent ping-pong buffer and FIFO array.

[0044] Preferably, the preprocessing of the received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal includes:

[0045] The received radio frequency signal is down-converted and analog-to-digital converted through the radio frequency receiving link to obtain a digital I / Q signal;

[0046] Perform digital I / Q imbalance correction on the obtained digital I / Q signals;

[0047] The corrected signal is digitally I / Q mixed to shift the signal to the baseband.

[0048] For the mixed signal, if it is a chirp spread spectrum signal, then the Decirp process is performed, multiplying the received signal with the basic chirped conjugate signal to compress the wideband swept frequency signal into a single-frequency sinusoidal carrier signal; if it is an MFSK modulated signal, then the MFSK modulated signal is a narrowband single-tone signal for a certain time period.

[0049] The processed signal is subjected to Cordic coordinate transformation to obtain amplitude and phase information, and low-frequency phase noise, residual frequency offset and low-frequency frequency drift noise are removed by pure digital PLL tracking.

[0050] On the other hand, this invention proposes an ultra-narrowband IoT unified demodulation system, comprising:

[0051] The signal preprocessing module is used to preprocess the received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal. The Chirp spread spectrum signal is compressed into a single-frequency sinusoidal carrier signal by Dechirp processing, so that the effective signal frequency point after compression uniquely corresponds to one of multiple discrete sub-bands.

[0052] The pre-processing RSSI power classification prediction module is used to detect the SNR or RSSI received signal strength of the signal output by the signal preprocessing module and divide it into three working modes according to the detection results.

[0053] The three-level adaptive demodulation mode switching module is used to select the corresponding demodulation mode for demodulation processing according to the three working modes divided by the front-end RSSI power classification prediction module. The three working modes include time-domain direct demodulation mode, time-frequency holographic layered demodulation mode and ultra-weak signal spectrum noise reduction mode.

[0054] The tree-shaped subchannel decomposition module is used to initiate binary, quartic, N-part, or hybrid tree-shaped subchannel decomposition when the three-level adaptive demodulation mode switching module selects the time-frequency holographic hierarchical demodulation mode or the ultra-weak signal spectrum noise reduction mode. The subchannels are decomposed at each level through complex mixing, filtering, and decimation to obtain subchannels with signals.

[0055] The pure noise subchannel determination and computing power shutdown module is used to determine each subchannel based on the characteristics of the pure noise subchannel after each layer of the tree subchannel decomposition module, and to stop the subsequent tree subchannel decomposition operation for subchannels determined to be pure noise.

[0056] The frequency point decision and decoding output module is used to complete the ultra-narrowband frequency point decision based on the sub-channels with signals obtained by the tree-shaped sub-channel decomposition module and the pure noise sub-channel decision and computing power shutdown module, and output the demodulated data.

[0057] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0058] This invention incorporates a pre-demodulation RSSI power classification prediction stage. The pre-processed signal undergoes SNR or RSSI detection, and based on the results, the signal is divided into three operating modes: time-domain direct demodulation, time-frequency holographic layered demodulation, and ultra-weak signal spectral noise reduction. This ensures that only signals of different intensities trigger demodulation operations matching their strengths, preventing unnecessary blind computations from the outset. During the tree-like subchannel decomposition process, after each layer is decomposed, each subchannel is judged based on its pure noise characteristics (including time-domain PAPR, average power, power drop characteristics at each layer, and SNR). Subchannels judged as pure noise are decomposed further, preventing complex mixing, filtering, and decimation operations from being performed on subchannels with no signal. Simultaneously, based on the universal underlying physical principle of spread spectrum despreading equivalent ultra-narrowband MFSK, an integrated universal baseband architecture combining ultra-narrowband MFSK and all types of Chirp spread spectrum is constructed, enabling unified processing of multiple systems. This reduces the receiver's computational overhead on invalid signals, lowering overall power consumption and hardware costs. Attached Figure Description

[0059] Figure 1 This is a flowchart of the ultra-narrowband Internet of Things unified demodulation method of the present invention;

[0060] Figure 2 This is a block diagram of the pipeline architecture for subchannel decomposition in this invention;

[0061] Figure 3 This is a block diagram of the first-level unit architecture of the tree-shaped four-molecule channel decomposition of the present invention;

[0062] Figure 4 This is a block diagram of the second-level first-path architecture of the tree-shaped four-molecule channel decomposition of the present invention;

[0063] Figure 5 This is a block diagram of the second-level, second-path architecture of the tree-shaped four-molecule channel decomposition of the present invention;

[0064] Figure 6 This is a block diagram of the second-level third path structure of the tree-shaped four-molecule channel decomposition of the present invention;

[0065] Figure 7 This is a block diagram of the second-level fourth path structure of the tree-shaped four-molecule channel decomposition of the present invention;

[0066] Figure 8 This is a block diagram of the ultra-narrowband IoT unified demodulation system of the present invention. Detailed Implementation

[0067] 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 a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0068] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a specific posture. If the specific posture changes, the directional indicators will also change accordingly.

[0069] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the use of "and / or" or "and / or" throughout the text includes three parallel solutions. For example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0070] Please see Figures 1-7 As shown, the present invention provides a unified demodulation method for ultra-narrowband Internet of Things, comprising the following steps:

[0071] The received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal is preprocessed, wherein the Chirp spread spectrum signal is compressed into a single-frequency sinusoidal carrier signal by Dechirp processing, so that the effective signal frequency point after compression uniquely corresponds to one of multiple discrete sub-bands;

[0072] Specifically, the received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal is preprocessed, including: down-converting and analog-to-digital converting the received radio frequency signal through the radio frequency receiving link to obtain a digital I / Q signal; performing digital I / Q imbalance correction on the obtained digital I / Q signal; and performing digital I / Q mixing on the corrected signal to shift the signal to the baseband.

[0073] For the mixed signal, if it is a chirp spread spectrum signal, then Dechirp processing is performed, multiplying the received signal with the basic chirped conjugate signal to compress the wideband swept frequency signal into a single-frequency sinusoidal carrier signal; if it is an MFSK modulated signal, then the MFSK modulated signal is a narrowband single-tone signal within a certain time period; the processed signal is subjected to Cordic coordinate transformation to obtain amplitude and phase information, and low-frequency phase noise, residual frequency offset and low-frequency frequency drift noise are removed by pure digital PLL tracking.

[0074] In practical applications, many MFSK-type modulated signals are narrowband single-tone signals within a certain time period; and many spread spectrum signals, after despreading, compress the broadband swept frequency signal into a single-frequency sinusoidal carrier signal. After compression, the effective signal frequency point uniquely corresponds to one of multiple (e.g., 4096) discrete subbands, laying the foundation for subsequent N-division of the spectrum (commonly divided into two, four, N, or mixed divisions).

[0075] Taking Lora as an example, its instantaneous frequency expression is: ;

[0076] The expression for the instantaneous phase after integration is: ;

[0077] The corresponding complex signal is: ;

[0078] The Lora transmit baseband signal for transmitting symbol s is: the equivalent superposition of symbols s has an initial frequency offset, i.e. Therefore: This is the broadband sweep frequency signal.

[0079] The Dechirp process is as follows: using the basic chirped conjugate signal (i.e., inverse frequency sweep), that is... Multiply by the received signal:

[0080] Right now Ignore noise terms Then there is ;

[0081] This is the single-tone signal after Decirp.

[0082] The preprocessed signal is subjected to SNR or RSSI received signal strength detection, and three working modes are divided according to the detection results.

[0083] The process of detecting the received signal strength of SNR or RSSI includes: calculating the amplitude of the preprocessed signal using a simplified algorithm for calculating amplitude using Cordic.

[0084] The simplified algorithm's calculation formula is as follows: Where |I| is the absolute value of the in-phase branch and |Q| is the absolute value of the orthogonal branch;

[0085] Based on the calculated amplitude Amp, the RSSI value is calculated, where RSSI = Amp * reference - link gain, and the reference is the power value corresponding to the ADC Vref. The SNR value is calculated based on the ratio of the average to the variance of the calculated amplitude Amp. The calculated RSSI and SNR values ​​are then compared with preset power and signal-to-noise ratio (SNR) thresholds to classify the pre-processed signal into three operating modes. For example, for signals within a 125kHz bandwidth, two key power thresholds can be set (this is just an example): -110dBm and -123dBm, corresponding to industry-tested demodulation SNR thresholds of 13dB and 0dB.

[0086] According to the three working modes, the corresponding demodulation mode is selected for demodulation processing. The three working modes include time-domain direct demodulation mode, time-frequency holographic layered demodulation mode, and ultra-weak signal spectrum noise reduction mode.

[0087] In this embodiment, the time-domain direct decoding mode includes:

[0088] When the RSSI and SNR meet the first preset conditions (RSSI≥-110dBm, SNR≥13dB), the time-domain direct solution mode is selected, all the tree sub-channel decomposition modules are bypassed, and clock-gated sleep is performed on the tree module; the on-chip Cordic arithmetic unit is called to solve the instantaneous phase of the preprocessed signal point by point;

[0089] Based on the solved instantaneous phase, instantaneous frequency discrimination is achieved through phase difference between adjacent sampling points, where... , The sampling interval is... The phase difference between adjacent samples is used; based on the frequency value obtained from instantaneous frequency discrimination, the discrete frequency point is located, and decoding is completed. This mode does not involve any tree-structured filtering decomposition operations, has the lowest power consumption, and covers most everyday near-field communication scenarios.

[0090] In this embodiment, the time-frequency holographic layered demodulation mode includes:

[0091] When the RSSI and SNR meet the second preset conditions (-123dBm < RSSI < -110dBm, 0dB ≤ SNR < 13dB), the time-frequency holographic layered demodulation mode is selected, and the binary, quartic, N-part, or hybrid tree-shaped sub-channel decomposition module is started; for the started tree-shaped sub-channel decomposition module, complex mixing, filtering, and decimation operations are performed step by step to decompose each sub-channel at each layer;

[0092] For each sub-channel after decomposition at each layer, the characteristics of the pure noise sub-channel are used for determination. The characteristics of the pure noise sub-channel include: the time-domain PAPR is greater than that of the sub-channel with signal, the average power is less than the noise power spectral density multiplied by the bandwidth, the pure noise sub-channel shows a 3dB decrease per layer when decomposed into a bi-tree sub-channel or a 6dB decrease per layer when decomposed into a quaternary tree sub-channel, and the SNR is less than 0dB.

[0093] Based on the judgment results, the subsequent tree-shaped sub-channel decomposition operation is stopped for sub-channels that are judged as pure noise; the sub-channels that are not judged as pure noise are saved and the tree-shaped sub-channel decomposition operation continues when the judgment conditions are met in the future, or the top N (e.g. 8) sub-channels are saved and the operation continues.

[0094] In this embodiment, the ultra-weak signal spectrum noise reduction mode includes:

[0095] When the RSSI and SNR meet the third preset condition (RSSI≤-123dBm, SNR<0dB), the ultra-weak signal spectrum noise reduction mode is selected, and all phase discrimination related modules are turned off; for the signal after the phase discrimination is turned off, the binary, quartic, N-part or hybrid tree-shaped channel decomposition is performed step by step from the beginning.

[0096] During the hierarchical tree-shaped channel decomposition process, branch selection is performed based on branch energy characteristics and time-domain waveform fluctuation characteristics. The branch energy characteristics include average power, which is reduced by 3dB per layer for binary decomposition, 6dB per layer for quartic decomposition, and 10*log10(N)dB per layer for N-decomposition. The time-domain waveform fluctuation characteristics include SNR and PAPR. Based on the branch selection results, a sub-channel with signal is obtained; or it is determined that there is no signal in the current frame.

[0097] In the time-frequency holographic layered demodulation mode or the ultra-weak signal spectrum noise reduction mode, the binary, quartic, N-part or hybrid tree-shaped sub-channel decomposition module is activated. The sub-channels are decomposed at each layer through complex mixing, filtering and decimation. The pre-processed signal is then divided into layers of spectrum to obtain the sub-channels with signals.

[0098] This involves initiating a binary, quartic, N-part, or hybrid tree-structured subchannel decomposition module, which progressively decomposes each subchannel at each level through complex mixing, filtering, and decimation, including:

[0099] The digital frequency circle of the preprocessed signal is decomposed into four parts at the first level, and four mutually orthogonal sub-bands without spectral overlap are formed by coarse frequency shifting of ±π / 4.

[0100] Each sub-band after the first-level quartic decomposition is further decomposed into a second-level quartic decomposition. The sub-bands are further subdivided by ±π / 8 fine frequency shifting so that the effective signal and the image interference frequency distance is 3π / 4.

[0101] In each decomposition stage, the signal after complex mixing is filtered by a half-band low-pass filter, the number of taps of which is configurable; the filtered signal is then decimated, and a streaming point-by-point processing method is used to repeat the complex mixing, filtering and decimation operations stage by stage until the decomposition of all sub-channels is completed.

[0102] Specifically, the number of taps for the half-band low-pass filter can be configured to include:

[0103] Ultra-low power consumption and ultra-high sensitivity settings: The last stage uses Haar (i.e., 1 tap), the second to last HBF uses 3 taps, the third to last HBF uses 7 taps, the fourth to last HBF uses 15 taps, and all other HBFs use 15 taps.

[0104] General balance settings: The last level uses Haar (i.e., 1 tap), the second to last HBF uses 3 taps, the third to last HBF uses 7 taps, the fourth to last HBF uses 15 taps, and the remaining HBFs all use 31 taps.

[0105] Ultra-high anti-interference networking level: The last level uses Haar (i.e., 1 tap), the second to last HBF uses 3 taps, the third to last HBF uses 7 taps, the fourth to last HBF uses 15 taps, the fifth to last HBF uses 31 taps, and the remaining HBFs all use 63 taps.

[0106] Preferably, the module initiates a binary, quartic, N-part, or hybrid tree-structured subchannel decomposition module, decomposing each subchannel at each level through complex mixing, filtering, and decimation. It also includes: for the data generated by multi-rate interpolation in each subchannel decomposition at each level through complex mixing, filtering, and decimation, an in-situ multiplexing operation mechanism with frequency-domain read-only synchronization and time-domain sliding coverage is adopted, eliminating the independent ping-pong buffer and FIFO array.

[0107] After each layer of tree-shaped sub-channel decomposition, each sub-channel is judged based on the characteristics of the pure noise sub-channel. Sub-channels judged as pure noise are stopped from subsequent tree-shaped sub-channel decomposition operations. Based on the sub-channels with signals, the ultra-narrowband frequency point decision is completed, and the demodulated data is output.

[0108] On the other hand, this invention proposes an ultra-narrowband IoT unified demodulation system, such as... Figure 8 As shown, it includes:

[0109] The signal preprocessing module is used to preprocess the received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal. The Chirp spread spectrum signal is compressed into a single-frequency sinusoidal carrier signal by Dechirp processing, so that the effective signal frequency point after compression uniquely corresponds to one of multiple discrete sub-bands.

[0110] The pre-processing RSSI power classification prediction module is used to detect the SNR or RSSI received signal strength of the signal output by the signal preprocessing module and divide it into three working modes according to the detection results.

[0111] The three-level adaptive demodulation mode switching module is used to select the corresponding demodulation mode for demodulation processing according to the three working modes divided by the front-end RSSI power classification prediction module. The three working modes include time-domain direct demodulation mode, time-frequency holographic layered demodulation mode and ultra-weak signal spectrum noise reduction mode.

[0112] The tree-shaped subchannel decomposition module is used to initiate binary, quartic, N-part, or hybrid tree-shaped subchannel decomposition when the three-level adaptive demodulation mode switching module selects the time-frequency holographic hierarchical demodulation mode or the ultra-weak signal spectrum noise reduction mode. The subchannels are decomposed at each level through complex mixing, filtering, and decimation to obtain subchannels with signals.

[0113] The pure noise subchannel determination and computing power shutdown module is used to determine each subchannel based on the characteristics of the pure noise subchannel after each layer of the tree subchannel decomposition module, and to stop the subsequent tree subchannel decomposition operation for subchannels determined to be pure noise.

[0114] The frequency point decision and decoding output module is used to complete the ultra-narrowband frequency point decision based on the sub-channels with signals obtained by the tree-shaped sub-channel decomposition module and the pure noise sub-channel decision and computing power shutdown module, and output the demodulated data.

[0115] Furthermore, the modules in the above system are also used to implement other steps of the above-mentioned unified demodulation method for ultra-narrowband Internet of Things, which will not be elaborated here.

[0116] In summary, based on the general physical principles of spread spectrum despreading equivalent ultra-narrowband MFSK, this invention constructs an integrated universal baseband architecture combining ultra-narrowband MFSK and all types of Chirp spread spectrum, achieving the following technical effects:

[0117] At the principle level, this invention is the first to uniformly model the despread signal after Chirp spread spectrum as an ultra-narrowband MFSK signal, incorporating traditional spread spectrum systems such as LoRa into the scope of this general model as special cases, thus breaking down the barriers between narrowband and spread spectrum communication systems from the ground up. In traditional solutions, the two types of baseband architectures are independent and cannot be reused. This invention achieves full system compatibility with a single baseband through a unified underlying physical model. A single IP can replace multiple dedicated basebands, significantly reducing chip design costs and development complexity.

[0118] In terms of receiving performance, leveraging the physical characteristic of constant channel noise power spectral density, the intra-band noise integral bandwidth is compressed step-by-step through tree-shaped sub-channel decomposition, causing the effective signal frequency to converge to an ultra-narrowband range, thereby approaching the upper limit of the theoretical receiving sensitivity of wireless communication. Compared with the traditional FFT block operation architecture, the signal-to-noise ratio is significantly improved, enabling longer transmission distances at the same transmit power, meeting the coverage requirements of ultra-long-distance IoT.

[0119] At the level of independent control, it completely abandons the technical paths such as chirped scanning, FFT block transformation, and fixed frame length constraints relied upon by existing spread spectrum patents, and adopts a brand-new tree-shaped multi-rate ultra-narrowband frequency point decision architecture. The underlying principles, hardware topology and operation logic are all independent of the existing patent system, and there is no risk of infringement, thus achieving complete independent control of the core communication baseband.

[0120] In terms of system efficiency, the system avoids blind computation on invalid signals by using three working modes for adaptive matching and a dynamic shutdown mechanism for pure noise sub-channels; at the same time, it can complete frequency point decision without buffering the entire frame of data, achieving sub-symbol level low latency.

[0121] In terms of compatibility and versatility, it natively adapts to all ultra-narrowband MFSK, linear and nonlinear Chirp spread spectrum, and custom spread spectrum waveforms, without being constrained by frame length and spreading factor, and can flexibly adapt to diverse IoT application scenarios and networking configuration requirements.

[0122] The above description is merely an exemplary embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformations made using the contents of the present invention specification and drawings under the technical concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A unified demodulation method for ultra-narrowband Internet of Things, characterized in that, include: The received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal is preprocessed, wherein the Chirp spread spectrum signal is compressed into a single-frequency sinusoidal carrier signal by Dechirp processing, so that the effective signal frequency point after compression uniquely corresponds to one of multiple discrete sub-bands; The preprocessed signal is subjected to SNR or RSSI received signal strength detection, and three working modes are divided according to the detection results. According to the three working modes, the corresponding demodulation mode is selected for demodulation processing. The three working modes include time-domain direct demodulation mode, time-frequency holographic layered demodulation mode, and ultra-weak signal spectrum noise reduction mode. In the time-frequency holographic layered demodulation mode or the ultra-weak signal spectrum noise reduction mode, the binary, quartic, N-part or hybrid tree-shaped sub-channel decomposition module is activated. The sub-channels are decomposed at each layer through complex mixing, filtering and decimation. The pre-processed signal is then divided into layers of spectrum to obtain the sub-channels with signals. After each layer of tree-shaped sub-channel decomposition, each sub-channel is judged based on the characteristics of pure noise sub-channels, and the sub-channels judged as pure noise are stopped from subsequent tree-shaped sub-channel decomposition operations. Based on the sub-channel with signal, complete the ultra-narrowband frequency point decision and output the demodulated data.

2. The unified demodulation method for ultra-narrowband Internet of Things according to claim 1, characterized in that, The process of detecting the received signal strength (SNR or RSSI) includes: The amplitude of the preprocessed signal is calculated using a simplified algorithm for calculating amplitude using Cordic. The calculation formula for the simplified algorithm is as follows: Where |I| is the absolute value of the in-phase branch and |Q| is the absolute value of the orthogonal branch; Based on the calculated amplitude Amp, calculate the RSSI value, where RSSI = Amp * reference - link gain, and the reference is the power value corresponding to ADCVref; The SNR value is calculated based on the ratio of the average value to the variance of the calculated amplitude Amp. Based on the calculated RSSI and SNR values, the preprocessed signal is compared with the preset power threshold and signal-to-noise ratio threshold, and then divided into three working modes.

3. The unified demodulation method for ultra-narrowband Internet of Things according to claim 1, characterized in that, The time-domain direct solution mode includes: When the RSSI and SNR meet the first preset condition, the time domain direct decoding mode is selected, all the tree sub-channel decomposition modules are bypassed, and clock-gated sleep is performed on the tree module; The on-chip Cordic arithmetic unit is invoked to solve the instantaneous phase of the preprocessed signal point by point. Based on the solved instantaneous phase, instantaneous frequency discrimination is achieved through phase difference between adjacent sampling points, where... ,in, The sampling interval is... The phase difference between adjacent sample points; Based on the frequency value obtained from instantaneous frequency discrimination, the discrete frequency point position is located, and decoding is completed.

4. The unified demodulation method for ultra-narrowband Internet of Things according to claim 1, characterized in that, The time-frequency holographic layered demodulation mode includes: When the RSSI and SNR meet the second preset condition, the time-frequency holographic hierarchical demodulation mode is selected, and the binary, quartic, N-part, or hybrid sub-tree subchannel decomposition module is started. For the initiated tree-shaped sub-channel decomposition module, complex mixing, filtering and decimation operations are performed level by level to decompose each sub-channel at each level; For each sub-channel after decomposition at each layer, the characteristics of the pure noise sub-channel are used for determination. The characteristics of the pure noise sub-channel include: the time-domain PAPR is greater than that of the sub-channel with signal, the average power is less than the noise power spectral density multiplied by the bandwidth, the pure noise sub-channel shows a 3dB decrease per layer when decomposed into a bi-tree sub-channel or a 6dB decrease per layer when decomposed into a quaternary tree sub-channel, and the SNR is less than 0dB. Based on the judgment results, the subsequent tree-shaped subchannel decomposition operation is stopped for subchannels that are judged as pure noise; the subchannels that are not judged as pure noise are saved and the tree-shaped subchannel decomposition operation is continued when the judgment conditions are met in the future, or the N subchannels with the highest average power ranking are saved and the operation is continued.

5. The unified demodulation method for ultra-narrowband Internet of Things according to claim 1, characterized in that, The ultra-weak signal spectral noise reduction mode includes: When RSSI and SNR meet the third preset condition, select the ultra-weak signal spectrum noise reduction mode and turn off all phase frequency discrimination modules. For signals with phase discrimination disabled, perform binary, quartic, N-part, or hybrid tree-structured channel decomposition starting from the beginning; In the hierarchical tree-shaped channel decomposition process, branch selection is performed based on branch energy characteristics and time-domain waveform fluctuation characteristics. The branch energy characteristics include average power, which is reduced by 3dB per layer for binary decomposition, 6dB per layer for quartic decomposition, and 10*log10(N)dB per layer for N-decomposition. The time-domain waveform fluctuation characteristics include SNR and PAPR. Based on the result of the branch selection, a sub-channel with signal is obtained; or it is determined that there is no signal in the current frame.

6. The ultra-narrowband IoT unified demodulation method according to claim 1, characterized in that, The initiated binary, quartic, N-part, or hybrid tree-structured subchannel decomposition module decomposes each subchannel at each level through complex mixing, filtering, and decimation, including: The digital frequency circle of the preprocessed signal is decomposed into four parts at the first level, and four mutually orthogonal sub-bands without spectral overlap are formed by coarse frequency shifting of ±π / 4. Each sub-band after the first-level quartic decomposition is further decomposed into a second-level quartic decomposition. The sub-bands are further subdivided by ±π / 8 fine frequency shifting so that the effective signal and the image interference frequency distance is 3π / 4. In each decomposition stage, the signal after complex mixing is filtered by a half-band low-pass filter, the number of taps of which is configurable. The filtered signal is decimated using a streaming point-by-point processing method, repeating complex mixing, filtering, and decimation operations step by step until all sub-channels are decomposed.

7. The ultra-narrowband IoT unified demodulation method according to claim 6, characterized in that, The number of taps of the half-band low-pass filter can be configured specifically including: Ultra-low power consumption and ultra-high sensitivity settings: The last stage uses Haar, i.e., 1 tap; the second to last HBF stage uses 3 taps; the third to last HBF stage uses 7 taps; the fourth to last HBF stage uses 15 taps; and all other HBF stages use 15 taps. General balance settings: The last level uses Haar, i.e., 1 tap; the second to last HBF level uses 3 taps; the third to last HBF level uses 7 taps; the fourth to last HBF level uses 15 taps; and all other HBF levels use 31 taps. Ultra-high anti-interference networking level: The last level uses Haar, i.e., 1 tap; the second to last HBF uses 3 taps; the third to last HBF uses 7 taps; the fourth to last HBF uses 15 taps; the fifth to last HBF uses 31 taps; and the remaining HBFs all use 63 taps.

8. The unified demodulation method for ultra-narrowband Internet of Things according to claim 1, characterized in that, The module for initiating binary, quartic, N-part, or hybrid sub-channel decomposition, which decomposes each sub-channel at each level through complex mixing, filtering, and decimation, further includes: For the data generated by multi-rate interpolation in each sub-channel at each level through complex mixing, filtering and decimation, an in-situ multiplexing operation mechanism with frequency domain read-only synchronization and time domain sliding coverage is adopted, eliminating the independent ping-pong buffer and FIFO array.

9. The unified demodulation method for ultra-narrowband Internet of Things according to claim 1, characterized in that, The preprocessing of the received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal includes: The received radio frequency signal is down-converted and analog-to-digital converted through the radio frequency receiving link to obtain a digital I / Q signal; Perform digital I / Q imbalance correction on the obtained digital I / Q signals; The corrected signal is digitally I / Q mixed to shift the signal to the baseband. For the mixed signal, if it is a chirp spread spectrum signal, then the Decirp process is performed, multiplying the received signal with the basic chirped conjugate signal to compress the wideband swept frequency signal into a single-frequency sinusoidal carrier signal; if it is an MFSK modulated signal, then the MFSK modulated signal is a narrowband single-tone signal for a certain time period. The processed signal is subjected to Cordic coordinate transformation to obtain amplitude and phase information, and low-frequency phase noise, residual frequency offset and low-frequency frequency drift noise are removed by pure digital PLL tracking.

10. An ultra-narrowband IoT unified demodulation system, used to implement the ultra-narrowband IoT unified demodulation method as described in any one of claims 1-9, characterized in that, include: The signal preprocessing module is used to preprocess the received ultra-narrowband MFSK modulated signal or Chirp spread spectrum signal. The Chirp spread spectrum signal is compressed into a single-frequency sinusoidal carrier signal by Dechirp processing, so that the effective signal frequency point after compression uniquely corresponds to one of multiple discrete sub-bands. The pre-processing RSSI power classification prediction module is used to detect the SNR or RSSI received signal strength of the signal output by the signal preprocessing module and divide it into three working modes according to the detection results. The three-level adaptive demodulation mode switching module is used to select the corresponding demodulation mode for demodulation processing according to the three working modes divided by the front-end RSSI power classification prediction module. The three working modes include time-domain direct demodulation mode, time-frequency holographic layered demodulation mode and ultra-weak signal spectrum noise reduction mode. The tree-shaped subchannel decomposition module is used to initiate binary, quartic, N-part, or hybrid tree-shaped subchannel decomposition when the three-level adaptive demodulation mode switching module selects the time-frequency holographic hierarchical demodulation mode or the ultra-weak signal spectrum noise reduction mode. The subchannels are decomposed at each level through complex mixing, filtering, and decimation to obtain subchannels with signals. The pure noise subchannel determination and computing power shutdown module is used to determine each subchannel based on the characteristics of the pure noise subchannel after each layer of the tree subchannel decomposition module, and to stop the subsequent tree subchannel decomposition operation for subchannels determined to be pure noise. The frequency point decision and decoding output module is used to complete the ultra-narrowband frequency point decision based on the sub-channels with signals obtained by the tree-shaped sub-channel decomposition module and the pure noise sub-channel decision and computing power shutdown module, and output the demodulated data.

Images