A quasi-coherent demodulation method for multi-mode narrowband wireless communication

Abstract

The application relates to the technical field of wireless communication, and particularly discloses a quasi-coherent demodulation method for multi-mode narrowband wireless communication. By introducing a narrowband digital phase-locked loop to separate a slowly-varying carrier component and using a secondary PLL to eliminate residual frequency offset, the demodulation obstacle caused by phase jump is successfully overcome under a full-digital architecture without a complex analog mixing circuit, the suppression capability of the system on low-frequency noise and frequency drift is significantly improved, high-sensitivity quasi-coherent demodulation is realized on the premise of maintaining ultra-low power consumption and low hardware resource occupation, and the problem that performance and power consumption are difficult to be considered together in multi-mode narrowband communication is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, specifically to a quasi-coherent demodulation method for multimode narrowband wireless communication. Background Technology

[0002] With the rapid development of wireless communication technology, multimode narrowband wireless communication systems such as Bluetooth Low Energy (BLE), Sub-GHz, and walkie-talkies place higher demands on receiver performance. In these systems, GFSK (Gaussian Frequency Shift Keying) and FSK modulation methods are widely used due to their high spectral efficiency and relatively simple implementation. However, existing demodulation schemes mainly fall into two categories:

[0003] One approach is the commonly used incoherent demodulation (such as phase differential detection). This method has a simple structure and does not require complex carrier recovery circuits, but its core principle is to use the phase difference between adjacent symbols for decision-making. The fatal flaw of this method is its low sensitivity; the 1e-3 bit error rate threshold is usually high (Eb / N0 > 12dB), and it is sensitive to carrier disturbances (such as local oscillator noise and crystal frequency drift), making it difficult to meet the high sensitivity requirements of modern low-power chips.

[0004] Second, there is traditional coherent demodulation. To overcome the performance bottleneck of noncoherent demodulation, traditional solutions typically employ a Costa ring in conjunction with a mixer for quadrature down-conversion, or use complex digital signal processing algorithms (such as Maximum Likelihood Sequence Estimation, MLSE). While these methods can significantly improve demodulation sensitivity, they often require larger front-end circuitry or higher digital computing resources, leading to increased power consumption and area, making them unsuitable for some scenarios. Furthermore, in dynamic channels, the phase recovered from the GFSK signal is easily lost or corrupted. Loop filters used to recover the reference phase that are too wide or too narrow are unsuitable. Too narrow a filter will fail to keep up, resulting in severe reference phase distortion and rendering traditional carrier tracking schemes ineffective. Too wide a filter will lead to excessive reference phase noise.

[0005] Therefore, how to effectively suppress slow-varying carrier disturbances (such as frequency offset and phase noise) while maintaining a fully digital structure, low hardware complexity, and low power consumption, and accurately demodulate GFSK signals that may have cycle skipping, has become a pressing technical problem in the field of multimode narrowband wireless communication. Summary of the Invention

[0006] This invention aims to provide a quasi-coherent demodulation method for multimode narrowband wireless communication, which effectively solves the problem of balancing performance and power consumption in multimode narrowband communication.

[0007] To achieve the above objectives, the technical solution adopted by this invention is: a quasi-coherent demodulation method for multimode narrowband wireless communication, comprising:

[0008] S1: Receive the I / Q quadrature components of the wireless communication signal, and calculate the instantaneous total phase φ_total[n] of the wireless communication signal through CORDIC coordinate transformation. The instantaneous total phase φ_total[n] includes the slowly varying carrier component and the GFSK modulation phase.

[0009] S2: Input the instantaneous total phase φ_total[n] into the narrowband digital PLL carrier tracking module, and use a second-order narrowband digital phase-locked loop to track only the slowly varying carrier component in the instantaneous total phase φ_total[n], and output the carrier phase estimate φ_pll[n].

[0010] S3: Subtract the instantaneous total phase φ_total[n] from the carrier phase estimate φ_pll[n] to obtain the pure GFSK modulation phase φ_demod[n] that eliminates slow-varying carrier disturbances;

[0011] S4: Perform phase differential processing on the pure GFSK modulation phase φ_demod[n] to obtain the frequency modulation baseband signal f_demod[n];

[0012] S5: Input the frequency modulation baseband signal f_demod[n] to the residual frequency difference and frequency drift removal module, and use the pure digital PLL again to remove the residual frequency difference and frequency drift signal to obtain the frequency signal after frequency offset correction;

[0013] S6: Perform low-pass filtering on the frequency signal after frequency offset correction to remove high-frequency noise remaining in the signal, and obtain the filtered frequency signal.

[0014] S7: Use the CDR module to recover clock data from the filtered frequency signal and obtain a sample of the moment when the eye diagram is at its maximum opening;

[0015] S8: Perform a sign determination on the sample and output demodulation bits according to the positive or negative sign of the sample.

[0016] Preferably, in S1, the instantaneous total phase φ_total[n] output by the CORDIC coordinate transformation takes a value in the range of [-π,π], and when the value overflows, it directly loops back.

[0017] Preferably, in S2, the loop bandwidth of the narrowband digital PLL carrier tracking module is set to 3kHz-10kHz, with an optimal value of 5kHz, and the damping ratio ζ=0.707, to ensure that the loop only tracks the slowly varying carrier components and does not respond to the transitions of the high-speed modulation phase.

[0018] Preferably, in S2, the slow-varying carrier component includes the phase component corresponding to the local oscillator low-frequency phase noise, crystal oscillator drift, or slow-varying frequency offset.

[0019] Preferably, in S3, the formula for calculating the pure GFSK modulation phase φ_demod[n] is: φ_demod[n]=φ_total[n]-φ_pll[n], and the pure GFSK modulation phase φ_demod[n] has eliminated all slow-varying carrier disturbances and only retains GFSK modulation information.

[0020] Preferably, in S4, the specific operation of the phase differential processing is: f_demod[n]=φ_demod[n]-φ_demod[n-1], thereby obtaining the frequency modulation baseband signal.

[0021] Preferably, in S5, the residual frequency difference and frequency drift removal module uses a digital PLL to dynamically suppress the residual frequency difference and low-frequency drift signal, so as to avoid interfering with the subsequent synchronization and decision module.

[0022] Preferably, in S6, the low-pass filtering process includes FIR filtering and a bit-interval integrator, so that subsequent modules can operate reliably and stably.

[0023] Preferably, in S7, the CDR module is used to ensure that the sampling sample at the optimal moment is obtained, that is, the sample at the moment when the eye diagram is at its maximum opening.

[0024] Preferably, in S8, the specific rule for the symbol decision is: when the frequency value corresponding to the sample is greater than or equal to 0, it is decided as bit 1; when the frequency value corresponding to the sample is less than 0, it is decided as bit 0.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0026] This invention introduces a narrowband digital phase-locked loop to separate slowly varying carrier components and utilizes a secondary PLL to eliminate residual frequency offset. Under a fully digital architecture that eliminates the need for complex analog mixing circuits, it successfully overcomes the demodulation obstacles caused by phase jumps, significantly improving the system's ability to suppress low-frequency noise and frequency drift. Thus, while maintaining ultra-low power consumption and low hardware resource usage, it achieves highly sensitive quasi-coherent demodulation, effectively solving the problem of balancing performance and power consumption in multimode narrowband communication. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating the quasi-coherent demodulation system for multimode narrowband wireless communication according to the present invention.

[0028] Figure 2 This is a flowchart illustrating the phase and low-frequency noise removal process of the present invention.

[0029] Figure 3 This is a flowchart of the frequency offset and frequency drift removal process of the present invention. Detailed Implementation

[0030] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0031] Example 1

[0032] This invention proposes a quasi-coherent demodulation method for multimode narrowband wireless communication, comprising the following steps:

[0033] S1: Receive the I / Q quadrature components of the wireless communication signal, and calculate the instantaneous total phase φ_total[n] of the wireless communication signal through CORDIC coordinate transformation. The instantaneous total phase φ_total[n] includes the slowly varying carrier component and the GFSK modulation phase. The value range of the instantaneous total phase φ_total[n] output by CORDIC coordinate transformation is [-π, π]. When the value overflows, it directly loops back.

[0034] By utilizing the shift-addition characteristics of the CORDIC algorithm, the risk of numerical overflow is effectively avoided in the range of [-π, π] through a direct loopback mechanism, achieving low-power and high-stability instantaneous total phase extraction. This ensures accurate separation of the carrier component and the GFSK modulation phase in environments with strong noise or large frequency offset, significantly improving the synchronization accuracy and demodulation reliability of wireless communication systems.

[0035] S2: The instantaneous total phase φ_total[n] is input to the narrowband digital PLL carrier tracking module. A second-order narrowband digital phase-locked loop tracks only the slowly varying carrier components in the instantaneous total phase φ_total[n], and outputs the carrier phase estimate φ_pll[n]. The loop bandwidth of the narrowband digital PLL carrier tracking module is set to 3kHz-10kHz, with an optimal value of 5kHz, and a damping ratio ζ=0.707 to ensure that the loop tracks only the slowly varying carrier components and does not respond to high-speed modulation phase transitions. The slowly varying carrier components include the phase components corresponding to local oscillator low-frequency phase noise, crystal oscillator drift, or slowly varying frequency offset.

[0036] By configuring an optimal bandwidth of 5kHz and a damping ratio of 0.707, the second-order narrowband PLL accurately filters out high-speed GFSK modulation transitions and tracks only slowly varying carrier components (such as crystal drift and local oscillator noise), effectively suppressing high-frequency noise interference, significantly improving the stability of phase estimation and signal-to-noise ratio, and providing a clean carrier reference for subsequent demodulation.

[0037] S3: Subtract the instantaneous total phase φ_total[n] from the carrier phase estimate φ_pll[n] to obtain the pure GFSK modulation phase φ_demod[n] which eliminates slow-varying carrier disturbances; wherein, the calculation formula of the pure GFSK modulation phase φ_demod[n] is: φ_demod[n]=φ_total[n]-φ_pll[n], and the pure GFSK modulation phase φ_demod[n] has eliminated all slow-varying carrier disturbances and retains only GFSK modulation information.

[0038] By precisely removing slowly varying carrier disturbances such as crystal oscillator drift and local oscillator noise through differential operations, the total phase is restored to a pure GFSK modulation phase, effectively eliminating the impact of low-frequency drift on demodulation, significantly improving the signal-to-noise ratio and demodulation accuracy, and ensuring highly reliable data recovery even in complex channel environments.

[0039] S4: Perform phase difference processing on the pure GFSK modulation phase φ_demod[n] to obtain the frequency modulation baseband signal f_demod[n]; specifically, the specific operation of phase difference processing is: f_demod[n]=φ_demod[n]-φ_demod[n-1], thereby obtaining the frequency modulation baseband signal.

[0040] By using phase difference to convert the pure GFSK modulation phase into a baseband frequency signal, the FM demodulation function is directly realized. This effectively avoids the dependence of traditional frequency discriminators on carrier synchronization accuracy, while further suppressing residual low-frequency noise, significantly improving the linearity and anti-interference capability of the baseband signal, and providing high-quality frequency information for subsequent decision-making.

[0041] S5: Input the frequency modulation baseband signal f_demod[n] to the residual frequency difference and frequency drift removal module, and use a pure digital PLL to remove the residual frequency difference and frequency drift signal again to obtain the frequency signal after frequency offset correction; furthermore, the residual frequency difference and frequency drift removal module uses a digital PLL to dynamically suppress the residual frequency difference and low frequency drift signal to avoid interfering with the subsequent synchronization and decision module.

[0042] S6: Perform low-pass filtering on the frequency signal after frequency offset correction to remove high-frequency noise residue in the signal and obtain the filtered frequency signal; the low-pass filtering process includes FIR filtering and bit interval integrator so that subsequent modules can work reliably and stably.

[0043] S7: Use the CDR module to recover clock data from the filtered frequency signal and obtain the sample at the moment when the eye diagram is at its maximum opening; the CDR module is used to ensure that the sample at the best moment is obtained, that is, the sample at the moment when the eye diagram is at its maximum opening.

[0044] S8: Perform a sign determination on the sample and output demodulation bits according to the sign of the sample. Specifically, the sign determination rule is as follows: when the frequency value corresponding to the sample is greater than or equal to 0, the determination is bit 1; when the frequency value corresponding to the sample is less than 0, the determination is bit 0.

[0045] By precisely locking the moment of maximum eye opening through the CDR module for sampling, the effects of channel noise, jitter, and inter-symbol interference on signal amplitude are effectively overcome, significantly improving the signal-to-noise ratio and decision accuracy of the sampling. Combined with the sign decision rule based on the positive and negative of the frequency, high-reliability demodulation in the frequency shift keying (FSK) scenario is achieved, reducing the bit error rate and ensuring the stable transmission of the communication system.

[0046] Example 2

[0047] like Figures 1-3 As shown, this embodiment proposes a quasi-coherent demodulation system for multimode narrowband wireless communication, including a CORDIC phase extraction module, a narrowband digital PLL carrier tracking module, a phase subtraction module, and a symbol decision module connected in sequence; the functions of each module are as follows:

[0048] The CORDIC phase extraction module takes the I / Q quadrature components of the BLE GFSK received signal as input and calculates the instantaneous total phase φ_total[n] of the received signal using the CORDIC algorithm. This instantaneous total phase φ_total[n] includes the slowly varying carrier component φ_carrier[n] and the GFSK modulation phase φ_mod[n], satisfying φ_total[n] = φ_carrier[n] + φ_mod[n]; its value range is [-π, π], and it directly loops back when overflow occurs.

[0049] Narrowband digital PLL carrier tracking module: Input instantaneous total phase φ_total[n], employing a second-order narrowband digital phase-locked loop, tracking only the slowly varying carrier component φ_carrier[n] (covering the phase components corresponding to local oscillator low-frequency phase noise, crystal drift, and slowly varying frequency offset), outputting the carrier phase estimate φ_pll[n]. The loop bandwidth of this narrowband digital PLL is set to 3kHz~10kHz, with an optimal value of 5kHz, and a damping ratio ζ=0.707, ensuring that the loop only tracks the slowly varying component and does not respond to high-speed modulation phase transitions.

[0050] Phase subtraction module: Subtracts the instantaneous total phase φ_total[n] output by the CORDIC phase extraction module from the carrier phase estimate φ_pll[n] output by the narrowband digital PLL carrier tracking module to obtain the pure GFSK modulation phase φ_demod[n], i.e., φ_demod[n] = φ_total[n] - φ_pll[n]. This φ_demod[n] has eliminated all slowly varying carrier disturbances and only retains the GFSK modulation information (the phase jump per symbol is ±π / 2).

[0051] Phase differential module: The frequency modulation baseband signal is obtained by the operation f_demod[n] = φ_demod[n] - φ_demod[n-1].

[0052] Residual frequency difference and frequency drift removal module: After phase differential, the residual frequency difference and frequency drift signal may interfere with the subsequent synchronization and decision modules. The relevant effects can be easily eliminated by digital PLL.

[0053] Low-pass filter module: A large amount of high-frequency noise remaining in the signal can be removed by the FIR filter and the bit interval integrator, ensuring the reliable and stable operation of the subsequent synchronization decision module.

[0054] CDR module: It again uses a pure digital PLL to achieve symbol synchronization and obtains the sampled sample at the moment when the eye diagram is at its maximum opening, that is, the sampled sample at the optimal moment.

[0055] Symbol decision module: Directly performs symbol decision on the pure modulation frequency f_demod[n] output by the CDR module, and outputs demodulation bits according to the sign of the modulation frequency; the specific decision rule is: when f_demod[n] ≥ 0, the decision is bit "1"; when f_demod[n] < 0, the decision is bit "0" (which can be adjusted according to the actual modulation mapping relationship).

[0056] This invention eliminates the need for complex circuits such as mixers, multipliers, and multiple low-pass filters. It consists only of a CORDIC module, a narrowband PLL, a phase subtractor, and a sign decision unit. It has low hardware complexity, a small number of gate circuits, and power consumption far lower than traditional quasi-coherent demodulation schemes, making it fully compatible with the ultra-low power requirements of BLE.

[0057] Narrowband PLLs can dynamically track and cancel slow-varying carrier disturbances such as local oscillator low-frequency phase noise, crystal drift, and slow frequency offset, solving the problem of sensitivity to carrier disturbances in existing non-quasi-coherent demodulation, and are more robust in real chip applications.

[0058] By utilizing the phase detection characteristics of the PLL itself, phase unwinding is naturally achieved, avoiding the impact of the ±2π phase jump on demodulation, and no additional unwinding circuit is required. At the same time, noise suppression is achieved through a narrowband loop, eliminating the need for direct filtering in the phase domain, thus solving the technical pain point that the phase domain cannot be directly filtered in the existing technology.

[0059] Employing a fully digital architecture with no complex analog circuitry, it can be directly integrated into BLE SoC chips, resulting in a small footprint, low cost, and ease of mass production. Achieving an optimal balance between performance and complexity, it delivers near-ideal MSK quasi-coherent demodulation performance without the need for complex equalization or sequence detection algorithms, making it suitable for various low-power, high-sensitivity BLE devices such as IoT sensors and wearable devices.

[0060] In summary, this embodiment employs a two-stage cascaded digital phase-locked loop (PLL) collaborative demodulation architecture: a narrowband PLL precisely separates the slowly varying carrier component and filters out the GFSK modulation component, achieving carrier disturbance removal; a residual frequency offset suppression PLL further suppresses the residual frequency offset and frequency drift of the baseband signal after phase differentiation, and, in conjunction with the CORDIC algorithm, calculates instantaneous phase, phase difference demodulation, and CDR optimal sampling decision to form a fully digital cascaded demodulation link; the two stages of PLLs work together and coordinate, avoiding the low sensitivity of traditional noncoherent demodulation and eliminating the complex analog mixing and high-order algorithm overhead of traditional coherent demodulation, achieving the optimal collaborative effect of low power consumption, low hardware resources, and high sensitivity demodulation under multi-mode narrowband communication.

[0061] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A quasi-coherent demodulation method for multimode narrowband wireless communication, characterized in that, include: S1: Receive the I / Q quadrature components of the wireless communication signal, and calculate the instantaneous total phase φ_total[n] of the wireless communication signal through CORDIC coordinate transformation. The instantaneous total phase φ_total[n] includes the slowly varying carrier component and the GFSK modulation phase. S2: Input the instantaneous total phase φ_total[n] into the narrowband digital PLL carrier tracking module, and use a second-order narrowband digital phase-locked loop to track only the slowly varying carrier component in the instantaneous total phase φ_total[n], and output the carrier phase estimate φ_pll[n]. S3: Subtract the instantaneous total phase φ_total[n] from the carrier phase estimate φ_pll[n] to obtain the pure GFSK modulation phase φ_demod[n] that eliminates slow-varying carrier disturbances; S4: Perform phase differential processing on the pure GFSK modulation phase φ_demod[n] to obtain the frequency modulation baseband signal f_demod[n]; S5: Input the frequency modulation baseband signal f_demod[n] to the residual frequency difference and frequency drift removal module, and use the pure digital PLL again to remove the residual frequency difference and frequency drift signal to obtain the frequency signal after frequency offset correction; S6: Perform low-pass filtering on the frequency signal after frequency offset correction to remove high-frequency noise remaining in the signal, and obtain the filtered frequency signal. S7: Use the CDR module to recover clock data from the filtered frequency signal and obtain a sample of the moment when the eye diagram is at its maximum opening; S8: Perform a sign determination on the sample and output demodulation bits according to the positive or negative sign of the sample.

2. The quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S1, the instantaneous total phase φ_total[n] output by the CORDIC coordinate transformation has a range of [-π,π], and when the value overflows, it directly loops back.

3. The quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S2, the loop bandwidth of the narrowband digital PLL carrier tracking module is set to 3kHz-10kHz, with an optimal value of 5kHz, and the damping ratio ζ=0.707 to ensure that the loop only tracks the slowly varying carrier components and does not respond to the transitions of the high-speed modulation phase.

4. The quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S2, the slow-varying carrier component includes the phase component corresponding to the local oscillator low-frequency phase noise, crystal oscillator drift, or slow-varying frequency offset.

5. A quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S3, the formula for calculating the pure GFSK modulation phase φ_demod[n] is: φ_demod[n]=φ_total[n]-φ_pll[n], and the pure GFSK modulation phase φ_demod[n] has eliminated all slow-varying carrier disturbances and only retains GFSK modulation information.

6. The quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S4, the specific operation of the phase differential processing is: f_demod[n]=φ_demod[n]-φ_demod[n-1], thereby obtaining the frequency modulation baseband signal.

7. A quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S5, the residual frequency difference and frequency drift removal module uses a digital PLL to dynamically suppress the residual frequency difference and low-frequency drift signal in order to avoid interfering with the subsequent synchronization and decision module.

8. A quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S6, the low-pass filtering process includes FIR filtering and a bit-interval integrator.

9. A quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S7, the CDR module is used to ensure that the sampled sample at the optimal moment is obtained, that is, the sample at the moment when the eye diagram is at its maximum opening.

10. A quasi-coherent demodulation method for multimode narrowband wireless communication according to claim 1, characterized in that, In S8, the specific rule for the symbol decision is as follows: when the frequency value corresponding to the sample is greater than or equal to 0, it is decided as bit 1; when the frequency value corresponding to the sample is less than 0, it is decided as bit 0.

Images