A Method and System for Single-Carrier Symbol Correction in Narrowband IoT Uplink Based on Costas Ring

Abstract

This invention discloses a method and system for uplink single-carrier symbol correction in narrowband IoT based on a Costas ring, supporting the narrowband IoT protocol specifications defined in 3GPP Release 13. This invention combines the Costas ring with the characteristics of narrowband IoT uplink, utilizing software-defined radio equipment to acquire uplink single-carrier signals. By reconstructing the Costas ring structure, a nonlinear phase detector and a low-pass filter are jointly designed to form a loop filter with joint gain. A digital frequency synthesis module replaces the traditional voltage-controlled oscillator, dynamically compensating for symbol frequency offset and phase distortion based on the error signal output from the Costas ring, thereby completing the demodulation of the real narrowband RF signal. This invention solves the symbol distortion problem caused by frequency offset, multipath effects, and hardware nonlinearity in narrowband IoT scenarios, and can significantly reduce the communication bit error rate compared to traditional channel estimation schemes.

Description

Technical Field

[0001] This invention relates to the field of low-power wide-area Internet of Things (IoT) technology, and more specifically, to a method and system for single-carrier symbol correction of narrowband IoT uplink based on Costas ring. Background Technology

[0002] Narrowband Internet of Things (NB-IoT), as a core technology of low-power wide-area IoT, has been widely used in smart cities, industrial monitoring, and other fields. Its uplink typically employs a single-carrier transmission mode to meet the low-power and low-cost requirements of terminal devices. However, in complex wireless environments, uplink signals are susceptible to frequency offset, phase noise, and multipath interference, leading to symbol timing deviations and phase distortion, severely reducing demodulation performance. Especially in edge coverage scenarios, the accuracy of the crystal oscillator in terminal devices is limited, with frequency offsets reaching hundreds of Hz to several kHz. Traditional synchronization algorithms struggle to achieve high-precision compensation, becoming a key bottleneck restricting system reliability.

[0003] In existing technologies, symbol synchronization and correction in the uplink of narrowband IoT mainly rely on schemes such as cyclic prefix matching, phase-locked loop tracking, or blind equalization. For example, time-domain synchronization methods based on cyclic prefixes determine symbol boundaries by matching the correlation between the received signal and the cyclic prefix. However, the cyclic prefix length in narrowband IoT is relatively short, and synchronization accuracy drops significantly under multipath interference, and it cannot distinguish the coupling effect of frequency offset and phase noise. Although the combination of phase-locked loop and Kalman filtering can track phase changes, it is prone to loss of lock in low signal-to-noise ratio environments, and the computational complexity is difficult to meet real-time requirements. Blind equalization and iterative compensation techniques require multiple convergence iterations, leading to increased processing latency and failing to meet the real-time communication needs of narrowband IoT. These schemes generally suffer from insufficient dynamic adaptability, resource overhead and performance contradictions, and poor multi-protocol compatibility. For example, traditional algorithms are difficult to cope with signal distortion caused by rapid terminal movement or sudden interference, the complex calculations required for high-precision synchronization conflict with the low power consumption requirements of terminals, and single-protocol optimization design leads to increased hardware redundancy in multi-mode base stations.

[0004] Costas ring technology, as a classic carrier synchronization scheme, was initially used for coherent demodulation of frequency-modulated and phase-modulated signals. It extracts phase errors through orthogonal mixing and low-pass filtering to achieve frequency offset tracking without pilot assistance. In recent years, improved digital Costas rings have been attempted for use in orthogonal frequency division multiplexing (OFDM) systems, but their application in narrowband single-carrier scenarios has not yet been fully explored. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing a method and system for single-carrier symbol correction in the uplink of narrowband Internet of Things (NB-IoT) based on a Costas Loop. By leveraging the closed-loop feedback mechanism of the Costas Loop and the efficient adaptation to the Software Radio (SDR) hardware platform, the symbol distortion problem caused by frequency offset, multipath effects, and hardware nonlinearity in NB-IoT scenarios is solved.

[0006] According to a first aspect of this specification, a method for single-carrier symbol correction in the uplink of a narrowband Internet of Things (IoT) based on a Costas ring is provided, the method comprising:

[0007] S1. Receive the NB-IoT uplink single-carrier signal through a software-defined radio device, extract the time-domain symbol sequence and divide it into several symbol blocks, perform initial frequency offset estimation and compensation on the symbol blocks, and then perform downsampling processing to obtain the downsampled sequence.

[0008] S2, Construct the Costas ring correction structure, input the downsampled sequence of the symbol block into the Costas ring correction structure to complete the frequency offset estimation of the downsampled sequence;

[0009] The Costas ring correction structure is a closed-loop feedback structure consisting of a nonlinear phase detector, a low-pass filter, and a digital frequency synthesis module. The phase detector extracts the phase difference sequence through a sliding window and uses an I / Q-path separated symbol phase detection method to send the output signal to the low-pass filter. The phase detector and the low-pass filter are jointly designed, and the parameters in the system function of the low-pass filter are fused to form a loop filter with joint gain. The output signal of the loop filter is sent to the digital frequency synthesis module to generate an oscillation signal with a phase of its input value, and multiplied with the downsampled sequence of the input symbol block to obtain the loop output signal.

[0010] S3 uses the demodulation result of the NB-IoT uplink demodulation reference signal to perform phase rotation on the symbol constellation diagram phase, completes phase correction, and outputs the symbol sequence after frequency and phase correction.

[0011] Furthermore, the extraction of the time-domain symbol sequence specifically involves: receiving the NB-IoT uplink single-carrier signal, performing down-conversion and analog-to-digital conversion, and then suppressing out-of-band interference through a bandpass filter to extract the time-domain symbol sequence; the initial frequency offset estimation adopts an autocorrelation algorithm based on cyclic prefixes, combined with a cross-correlation algorithm for pilot symbols.

[0012] Furthermore, the extraction of the phase difference sequence in the phase detector is specifically as follows: a sliding window differential algorithm is used to perform complex conjugate multiplication on adjacent sampling points within the window to extract the phase difference sequence. The window length is dynamically adjusted to the range of 8-32 sampling points and is adaptively selected based on the SNR detection result: when SNR>10dB, a short window is enabled; when SNR<5dB, a long window is switched to.

[0013] Furthermore, the expression for the I / Q-channel separated symbolic phase detection method used by the phase detector is as follows:

[0014]

[0015] Among them, V D (t) represents the output of the phase detector and the input of the low-pass filter, s I (t) and s Q (t) represent the real and imaginary parts of the downsampled sequence s(t) of the symbol block at time t, respectively, and Sign(·) represents the sign function;

[0016] The system function F(z) of the low-pass filter is expressed as follows:

[0017]

[0018] Where z is a complex variable in the complex plane, and C1 and C2 are constants;

[0019] The phase detector and low-pass filter are jointly designed and combined into a loop filter, using a joint gain G. lf Instead of choosing constants C1 and C2, the expression is as follows:

[0020] O lf (t+1)=O lf (t)+G lf V D (t)

[0021] Among them, O lf (t),O lf (t+1) represents the outputs of the loop filter at time t and time t+1, respectively.

[0022] Furthermore, after each symbol block is processed, the Costas ring demodulates the loop output signal and performs correlation analysis with the locally reconstructed constellation diagram, calculates the error as a feedback quantity, and dynamically adjusts the joint gain value.

[0023] Furthermore, the symbol-by-symbol block iterative verification is performed within the Costas ring, specifically: phase alignment is performed on the time-domain sampling points of each symbol to eliminate residual errors of frequency offset; cyclic redundancy check is continuously performed in each retransmission of the NB-IoT terminal to determine the reliability of the current data packet; when the cyclic redundancy check passes, the final symbol sequence after frequency and phase correction is output.

[0024] Furthermore, the corrected symbols are input into the NB-IoT narrowband channel estimation module to extract the channel impulse response, and the compensation parameters of the Costas ring, namely the joint gain and the sliding window length, are dynamically adjusted according to the delay spread and multipath intensity.

[0025] Furthermore, hardware-level optimizations are performed to address resource constraints of the software-defined radio (SDR), including: for the USRP, converting the floating-point calculations of the Costas ring to 8-bit fixed-point operations and accelerating the parallel pipelined processing of phase detection and compensation using CUDA; for the USRP, deploying the hardware acceleration module of the Costas ring through an FPGA programmable interface to match the peak rate requirements of NB-IoT; and implementing a symbol block-level sleep mechanism on the SDR to disable the calculation function of the Costas ring when there is no data input.

[0026] Furthermore, the symbol correction method is compatible with at least one of the following narrowband IoT protocols: the 3.75kHz single-carrier uplink mode in the NB-IoT protocol, and the case of using only a single carrier in the 15kHz uplink mode of the NB-IoT protocol; the symbol correction method supports joint correction with the downlink, including: pre-calibrating the initial frequency offset of the Costas ring using the downlink reference signal, optimizing the joint gain and sliding window length through uplink and downlink channel reciprocity, and multiplexing the Costas ring hardware resources in time division duplex (TDD) mode.

[0027] According to a second aspect of this specification, a narrowband Internet of Things uplink single-carrier symbol correction system based on a Costas ring is provided, the system comprising:

[0028] The signal receiving and preprocessing module is used to receive the NB-IoT uplink single carrier signal through the software radio device, extract the time domain symbol sequence and divide it into several symbol blocks, perform initial frequency offset estimation and compensation on the symbol blocks, and then perform downsampling processing to obtain the downsampled sequence.

[0029] The frequency correction module is used to construct the Costas ring correction structure, input the downsampled sequence of the symbol block into the Costas ring correction structure, and complete the frequency offset estimation of the downsampled sequence.

[0030] The phase correction module is used to perform phase rotation on the symbol constellation diagram phase using the demodulation result of the NB-IoT uplink demodulation reference signal, thereby completing phase correction.

[0031] The signal output module is used to output the symbol sequence after frequency and phase correction to the demodulation module.

[0032] According to a third aspect of this specification, an electronic device is provided, including a memory and a processor, the memory being coupled to the processor; wherein the memory is used to store program data, and the processor is used to execute the program data to implement the narrowband Internet of Things uplink single-carrier symbol correction method based on Costas ring as described in the first aspect.

[0033] According to a fourth aspect of this specification, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the narrowband Internet of Things uplink single-carrier symbol correction method based on a Costas ring as described in the first aspect.

[0034] According to a fifth aspect of this specification, a computer program product is provided, comprising a computer program / instructions that, when executed by a processor, implement the narrowband Internet of Things uplink single-carrier symbol correction method based on a Costas ring as described in the first aspect.

[0035] The beneficial effects of this invention are as follows: The symbol correction method and system of this invention support the narrowband IoT protocol specifications specified in 3GPP Release 13. This invention innovatively combines the Costas ring with the uplink characteristics of narrowband IoT, utilizes software-defined radio equipment to collect NB-IoT uplink single-carrier signals, reconstructs the Costas ring correction structure, and jointly designs a nonlinear phase detector and a low-pass filter. The parameters in the system function of the low-pass filter are fused to form a loop filter with joint gain. A DDS module replaces the VCO in the traditional Costas ring design. Based on the error signal output by the Costas ring, symbol frequency offset and phase distortion are dynamically compensated, thereby completing the demodulation of real narrowband RF signals. In single-carrier mode, field signal acquisition and testing were conducted in common narrowband IoT usage environments such as indoors, corridors, playgrounds, roads, and factories, verifying the feasibility of the proposed solution. Compared to existing hardware implementation strategies, this invention features a special design for single-carrier mode, which significantly reduces the communication bit error rate compared to traditional channel estimation schemes. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 A flowchart illustrating a narrowband IoT uplink single-carrier symbol correction method based on a Costas ring, as an exemplary embodiment;

[0038] Figure 2 A schematic diagram of the structure of a Costas ring is shown as an exemplary embodiment;

[0039] Figure 3 A symbolic constellation diagram before the Costas ring, as shown in an exemplary embodiment;

[0040] Figure 4 A symbolic constellation diagram after passing through the Costas ring, as shown in an exemplary embodiment;

[0041] Figure 5 This is a schematic diagram of the structure of an electronic device as an exemplary embodiment. Detailed Implementation

[0042] To better understand the technical solution of this application, the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0043] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0044] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0045] This invention deeply integrates Costas ring closed-loop control theory with innovative software-defined radio hardware architecture. Addressing the symbol distortion problem in NB-IoT uplinks caused by Doppler frequency shift, crystal oscillator drift, and multipath interference, it proposes an adaptive correction method and system based on dynamic compensation. Within the 3GPP Release 13 standard framework, this solution focuses on overcoming the frequency offset tracking capability bottleneck of traditional phase-locked loops (PLLs) in low signal-to-noise ratio narrowband scenarios. By introducing a Costas ring nonlinear phase compensation mechanism, it achieves symbol-level distortion correction and energy efficiency optimization for resource-constrained devices.

[0046] like Figure 1 As shown, the present invention provides a narrowband IoT uplink single-carrier symbol correction method based on a Costas ring, and the specific implementation steps are as follows:

[0047] S1: Uplink single-carrier signal reception and preprocessing, specifically including the following sub-steps:

[0048] Step S1.1: Receive the NB-IoT uplink single-carrier signal sent by the NB-IoT terminal through the radio frequency front-end of a software-defined radio device such as USRP N210 or HackRF One. After down-conversion and analog-to-digital conversion, suppress out-of-band interference through a bandpass filter and extract the time-domain symbol sequence.

[0049] Step S1.2: Preliminary packet localization is performed based on the envelope detection algorithm. Utilizing the high-precision clock synchronization function of the software-defined radio, the time-domain symbol sequence is divided into fixed-length symbol blocks (e.g., 128 sampling points per block), with each block corresponding to one symbol. Initial frequency offset estimation and compensation are performed for each symbol block. The initial frequency offset estimation employs an autocorrelation algorithm based on the cyclic prefix, combined with a cross-correlation algorithm for pilot symbols, to control the frequency offset error within ±50Hz.

[0050] Step S1.3: Perform downsampling processing on the compensated symbol blocks, for example, reduce the sampling rate to 1 / 2 of the original sampling rate to adapt it to the symbol rate detection range of the Costas ring, and obtain the downsampling sequence of each symbol block.

[0051] S2: Construction and distortion analysis of the Costas ring, specifically including the following sub-steps:

[0052] Step S2.1: Construct a Costas ring correction structure based on the downsampling sequence of the symbol block. This Costas ring correction structure is a closed-loop feedback structure consisting of a nonlinear phase detector, a low-pass filter, and a digital frequency synthesis (DDS) module.

[0053] In a nonlinear phase detector, the phase difference sequence between adjacent sampling points is extracted using a sliding window (e.g., the window length is set to 1 / 8 of the symbol block length) to identify local distortion features of the symbol waveform. In this design, an I / Q-channel separated symbol phase detection method is used, and its specific expression is as follows:

[0054] V D (t)=s Q (t)Sing(s I (t))-s I (t)Sign(s Q (t))

[0055] Among them, V D (t) represents the output of the phase detector and the input of the low-pass filter, s I (t) and s Q (t) represents the real and imaginary parts of the downsampled sequence s(t) of the symbol block at time t, respectively, and Sign(·) represents the sign function.

[0056] The output signal of the nonlinear phase detector is fed into a low-pass filter for processing. The system function F(z) of the low-pass filter design is expressed as follows:

[0057]

[0058] Where z is a complex variable in the complex plane, and C1 and C2 are constants that can be flexibly set according to the actual application scenario. The most important influencing factor is the modulation mode of the signal. From the perspective of digital signal processing algorithm design, it is advisable to design the phase detector and low-pass filter together, fusing some parameters to reduce the experimental cost of parameter determination. Accordingly, this invention combines the expressions of the low-pass filter and phase detector into a loop filter, using a joint gain G. lf Instead of choosing constants C1 and C2, its expression is as follows:

[0059] O lf (t+1)=O lf (t)+G lf V D (t)

[0060] Among them, O lf This corresponds to the output of the loop filter, which is also the output of the original low-pass filter. This signal is actually the frequency error signal of the downsampled sequence of the symbol block, and it is sent as the excitation signal to the DDS module for subsequent iterative signal processing. Considering the simplicity of digital signal processing algorithms, this invention uses a DDS module instead of the voltage-controlled oscillator (VCO) in the traditional Costas ring design, which can directly generate the phase as its input value 0. lf The oscillation signal.

[0061] Step S2.2: Input the downsampled sequence s(t) of the symbol block into the Costas ring to start the closed-loop feedback frequency offset correction process and complete the frequency offset estimation of the downsampled sequence. The specific structure of the Costas ring is as follows: Figure 2 As shown. The final loop output signal s o (t)=s I (t)+js Q (t), where j is the imaginary unit.

[0062] After distortion analysis iteration, the NB-IoT uplink single-carrier signal has completed frequency offset correction, and its symbol constellation diagram has obvious four-block cluster characteristics of QPSK modulation symbols. However, since the initial phase offset cannot be corrected by the Costas ring, it is necessary to use the demodulation result of the NB-IoT uplink demodulation reference signal (DMRS) to perform complex multiplication phase rotation on the symbol constellation diagram phase.

[0063] Iterative verification is performed symbol-by-symbol block within the Costas ring. Specifically, phase alignment is performed on the time-domain sampling points of each symbol to eliminate residual frequency offset errors, thereby further eliminating constellation diagram ambiguity caused by frequency offset and noise in communication. Considering the low-power characteristics of NB-IoT terminals, the communication signal-to-noise ratio is often not high. Therefore, cyclic redundancy check (CRC) needs to be continuously performed in each retransmission by the NB-IoT terminal to determine the reliability of the current data packet. When the CRC passes, the final symbol sequence after frequency and phase correction is output.

[0064] Experiments show that when the HackRF One platform is used as an RF receiver, the correction delay of a single symbol block is less than 200μs, and after three iterations, the error is reduced from 22% to below 3.5%.

[0065] S3: Narrowband channel adaptation and resource optimization, specifically including the following sub-steps:

[0066] Step S3.1: Input the corrected symbol into the NB-IoT narrowband channel estimation module (MMSE), extract the channel impulse response (CIR), and dynamically adjust the compensation parameters of the Costas ring, namely the joint gain and the sliding window length, according to the delay spread and multipath intensity.

[0067] Step S3.2: Perform hardware-level optimization to address the resource limitations of the SDR device, specifically:

[0068] Fixed-point arithmetic and parallelization concept: For the USRP N210 software radio device, the floating-point calculation of the Costas ring is converted into 8-bit fixed-point arithmetic, and the parallel pipelined processing of phase detection and compensation is accelerated by CUDA, so that the FPGA resource utilization of USRP N210 is reduced to 35%.

[0069] Power consumption control: Implement a symbol block-level sleep mechanism on software-defined radio devices such as HackRF One or USRP N210 to disable the calculation function of the Costas ring when there is no data input, thereby reducing power consumption.

[0070] Step S3.3: For the USRP N210 software-defined radio device, deploy the hardware acceleration module of the Costas ring through its FPGA programmable interface to achieve a processing throughput of 150,000 symbols per second, which fully matches the peak rate requirements of NB-IoT.

[0071] S4: SDR experimental verification and performance evaluation, which includes the following sub-steps:

[0072] Step S4.1: Build a dual-node closed-loop test platform, use USRP N210 to simulate a base station receiver and HackRFOne to simulate a terminal transmitter, and verify the performance of the solution in a real channel scenario.

[0073] Step S4.2: Under extreme conditions of frequency offset ±7kHz and signal-to-noise ratio (SNR) of 3dB, the method of the present invention reduces the symbol error rate (SER) from 14% of the conventional LMS equalizer to 0.7%, and the end-to-end delay is less than 1.5ms.

[0074] Step S4.3: Monitor the Costas ring iteration process in real time through HackRF's visual debugging interface, enabling engineers to quickly locate distortion sources and optimize parameters, improving debugging efficiency by 60%.

[0075] The adaptive parameter configuration of the Costas loop in this invention includes: ① dynamically adjusting the joint gain of the loop filter based on the signal-to-noise ratio; ② dynamically adjusting the sliding window length of the phase detector based on the current signal quality.

[0076] The symbol correction method of this invention is compatible with at least one of the following narrowband Internet of Things (IoT) protocols: ① 3.75kHz single-carrier uplink mode in the NB-IoT protocol; ② the case of using only a single carrier in the 15kHz uplink mode of the NB-IoT protocol.

[0077] The symbol correction method of this invention supports joint correction with the downlink, including: ① pre-calibrating the initial frequency offset of the Costas ring using the downlink reference signal (NRS); ② optimizing the joint gain and sliding window length through uplink and downlink channel reciprocity; ③ multiplexing the Costas ring hardware resources in time division duplex (TDD) mode.

[0078] Using the above-described narrowband IoT uplink single-carrier symbol correction method based on Costas rings, the following technical effects can be achieved:

[0079] ① High precision and low complexity: The local feedback mechanism of the Costas ring compresses the computational complexity to 25% of the traditional equalization algorithm, achieving 99.5% symbol recovery accuracy on the USRP N210.

[0080] ② Interference resistance: In a multipath environment with a frequency offset of ±7kHz and a delay spread of 15μs, the bit error rate (BER) remains consistently below 10. -5 This represents an improvement of two orders of magnitude compared to existing solutions.

[0081] ③SDR platform compatibility: Supports hybrid deployment of HackRF and USRP, requiring no additional development compared to Raspberry Pi or FPGA solutions.

[0082] ④ Narrowband optimization: The compensation frequency domain range is limited to accommodate the 180kHz narrowband characteristics, avoiding out-of-band noise interference, and the downsampling technique reduces the computational load by 50%.

[0083] In the following example, the hardware layer employs a dual-platform collaborative architecture using the USRP N210 general-purpose software-defined radio and the HackRF One, while the software layer deploys a real-time signal processing pipeline and an FPGA hardware acceleration module, forming an end-to-end correction link from the RF front-end to baseband processing. Full-link verification was completed on both the HackRF One and USRP N210, significantly improving symbol demodulation accuracy and end-to-end communication system energy efficiency. The specific implementation process of this example is described below.

[0084] (1) Hardware platform construction, signal reception and preprocessing

[0085] At the RF receiver, the USRP N210 is used as the main processing node, configured for single-antenna reception. The center frequency is set in the 800MHz band, conforming to the 3GPP Band 20 specification, and the default sampling rate is configured to 1.92MHz, matching the minimum resolution requirements of different bandwidth modes in NB-IoT. In the hardware preprocessing stage, the received uplink single-carrier signal is first subjected to a two-stage downconversion operation: the first stage uses the ADL5375 RF chip (UBX-40 RF daughterboard) to perform analog domain mixing, shifting the signal to a 70MHz intermediate frequency; the second stage performs quadrature demodulation in the digital domain, generating I / Q baseband signals. To address the phase continuity requirements unique to narrowband signals, the analog-to-digital conversion stage uses a 12-bit precision ADC, coupled with an anti-aliasing filter bank (cutoff frequency ±100kHz, roll-off factor 0.25), ensuring out-of-band interference suppression greater than 40dB, and extracting the timing symbol sequence.

[0086] After preprocessing, the time-domain symbol sequence enters the symbol segmentation module. Based on the NB-IoT single-carrier frequency division multiple access (SC-FDMA) frame structure, the time-domain symbol sequence is divided into blocks of a fixed length of 128 sampling points. A 10 / 9-point cyclic prefix is ​​extracted from the header of each symbol block (based on the current symbol block's position in the time slot). In the initial frequency offset estimation stage, a composite estimation algorithm combining the cyclic prefix and pilot symbols is designed: first, a coarse frequency offset estimate is calculated using the correlation of the cyclic prefix, and then the frequency offset error is converged to within ±200Hz through complex rotation compensation; subsequently, the pre-set Zadoff-Chu sequence pilot symbols within the frame are extracted, and a least-squares fine frequency offset estimation is performed, ultimately controlling the frequency offset error within ±50Hz.

[0087] The downsampling stage uses a multiphase filter bank to achieve a 2x decimation, which compresses the data processing volume to 50% of the original signal while maintaining a symbol rate of 7.5 ksps, providing computing power redundancy for subsequent real-time processing.

[0088] (2) Costas Ring Correction Core Architecture Design

[0089] A Costas ring correction structure is constructed, which is a closed-loop feedback structure consisting of a nonlinear phase detector, a low-pass filter, and a DDS module.

[0090] Nonlinear phase detector: Employs a sliding window differential algorithm, with the window length dynamically adjusted between 8 and 32 sampling points. The selection is adaptively based on the signal-to-noise ratio (SNR) detection results: a short window (8 points) is used to improve response speed when SNR > 10dB; a long window (32 points) is switched to enhance noise immunity when SNR < 5dB. Specifically, complex conjugate multiplication is performed on adjacent sampling points within the window to extract the phase difference sequence.

[0091] Low-pass filter: The nonlinear phase detector and low-pass filter are jointly designed, and the parameters in the system function of the low-pass filter are fused to form a loop filter with joint gain and dynamic adjustment of loop bandwidth. After processing each symbol block, the demodulated loop output signal is correlated with the locally reconstructed QPSK constellation diagram, and the error is calculated as feedback to dynamically adjust the joint gain value.

[0092] The DDS module generates an oscillating signal with a specific phase based on the output signal of the loop filter, and multiplies it by the downsampled sequence of the input symbol block to obtain the loop output signal s. o (t).

[0093] (3) Resource constraint optimization and energy efficiency improvement

[0094] To address the stringent power consumption limitations of narrowband IoT terminals, a low-power adaptive architecture was implemented on the HackRF One platform. At the hardware level, a symbol block-level sleep mechanism was deployed: when the FPGA detects that the error of three consecutive symbol blocks is less than 5%, the computation function of the Costas ring is automatically shut down, maintaining only the basic frequency offset tracking loop, reducing the overall power consumption from 2.8W to 0.7W; when the error fluctuation exceeds a threshold, full-function operation is resumed within 200μs via an interrupt wake-up mechanism. At the algorithm level, fixed-point transformation was implemented, converting the floating-point calculations of the Costas ring into 8-bit fixed-point operations, thereby reducing computational overhead.

[0095] (4) Multi-scenario verification and performance analysis

[0096] A closed-loop test platform was built, including the USRP N210, HackRF One, and commercial NB-IoT terminals. Complete testing was conducted in real-world NB-IoT application scenarios such as school buildings, playgrounds, roads, and factories. Under extreme test conditions (factory environment, calculated frequency offset exceeding 8kHz), the system demonstrated excellent robustness: after 5 iterations of the Costas ring, the constellation diagram showed clear convergence (the constellation diagrams before and after the Costas ring are shown below). Figure 3 , Figure 4As shown in the figure, the symbol error rate (SER) stabilized within the range of 0.7%-1.2%, a 98.5% reduction compared to when correction was not enabled; end-to-end processing latency was controlled within 1.2ms, meeting the 3GPP requirement of 1ms latency for URLLC services. Actual deployment testing showed that during uplink transmission in the 800MHz band, the average number of bit error events decreased sharply from 950 before correction to 12, and the additional power consumption of the terminal due to retransmission decreased by 76%. Through a visual debugging interface, the symbol constellation convergence process and EVM heatmap can be observed in real time, quickly locating crystal oscillator anomalies caused by equipment aging. It is worth noting that this strategy has strong generalization capabilities for the uplink side of various low-power IoT terminals using frequency shift keying as the modulation scheme, and can be effectively migrated and applied.

[0097] The technical value of this invention lies in overcoming the challenge of symbol correction in the uplink of narrowband IoT under high dynamic and low signal-to-noise ratio environments. For example, in extreme scenarios with frequency offsets as high as ±2.5kHz, its symbol error rate can be lower than 10⁻³, more than three times higher than traditional solutions. Through a lightweight Costas ring algorithm, computational complexity is reduced by 60%, adapting to the low-power constraints of terminal devices. The dynamic parameter loading mechanism enables a single hardware platform to be compatible with multiple protocols, reducing base station upgrade costs. As 5G-A and 6G evolve towards integrated air-space-ground networks, narrowband IoT will be widely used in non-terrestrial network scenarios. This invention provides key technical support for reliable transmission in high dynamic and large frequency offset environments, and has significant industrialization prospects.

[0098] On the other hand, embodiments of the present invention provide a narrowband IoT uplink single-carrier symbol correction system based on a Costas ring, used to implement the above-mentioned narrowband IoT uplink single-carrier symbol correction method based on a Costas ring. The system includes:

[0099] The signal receiving and preprocessing module is used to receive the NB-IoT uplink single carrier signal through the software radio device, extract the time domain symbol sequence and divide it into several symbol blocks, perform initial frequency offset estimation and compensation on the symbol blocks, and then perform downsampling processing to obtain the downsampled sequence.

[0100] The frequency correction module is used to construct the Costas ring correction structure, input the downsampled sequence of the symbol block into the Costas ring correction structure, and complete the frequency offset estimation of the downsampled sequence.

[0101] The phase correction module is used to perform phase rotation on the symbol constellation diagram phase using the demodulation result of the NB-IoT uplink demodulation reference signal, thereby completing phase correction.

[0102] The signal output module is used to output the symbol sequence after frequency and phase correction to the demodulation module.

[0103] Regarding the system in the above embodiments, the specific manner in which each module performs its operations has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0104] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0105] Accordingly, this application also provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the narrowband IoT uplink single-carrier symbol correction method based on Costas rings as described above. Figure 5 The diagram shown illustrates a hardware structure of any data processing-capable device for implementing the narrowband IoT uplink single-carrier symbol correction method based on a Costas ring, as provided in this embodiment of the invention. (Except for...) Figure 5 In addition to the processor, memory, and network interface shown, any data processing device in the embodiment may also include other hardware depending on the actual function of the data processing device, which will not be described in detail here.

[0106] Accordingly, this application also provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, implement the aforementioned narrowband IoT uplink single-carrier symbol correction method based on a Costas ring. The computer-readable storage medium can be an internal storage unit of any data-processing device as described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units of any data-processing device and external storage devices. The computer-readable storage medium is used to store the computer program and other programs and data required by the data-processing device, and can also be used to temporarily store data that has been output or will be output.

[0107] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0108] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

[0109] The above description is merely a preferred embodiment of the present invention. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make many possible variations and modifications to the technical solutions of the present invention using the methods and techniques disclosed above, or modify them into equivalent embodiments with equivalent changes, without departing from the scope of the technical solutions of the present invention. Therefore, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solutions of the present invention shall still fall within the protection scope of the technical solutions of the present invention.

Claims

1. A method for correction of single carrier symbols for uplink in narrowband internet of things based on a Costa's loop, characterized in that, The method comprises the following steps: S1, receiving an NB-IoT uplink single carrier signal through a software radio device, extracting a time domain symbol sequence and dividing it into several symbol blocks, performing initial frequency offset estimation and compensation on the symbol blocks, and then performing downsampling processing to obtain a downsampled sequence; S2, constructing a Costas loop correction structure, inputting the downsampled sequence of the symbol blocks into the Costas loop correction structure, and completing frequency offset estimation of the downsampled sequence; The Costas loop correction structure is a closed-loop feedback structure composed of a nonlinear phase discriminator, a low-pass filter and a digital frequency synthesis module; the phase discriminator extracts a phase difference sequence through a sliding window, adopts an I / Q path separation type symbol phase discrimination method, and sends an output signal into a low-pass filter; The phase discriminator and the low-pass filter are jointly designed, the parameters in the system function of the low-pass filter are fused, a loop filter with joint gain is formed, the output signal of the loop filter is sent into a digital frequency synthesis module, an oscillation signal with a phase being the input value is generated, and the oscillation signal is multiplied by the downsampled sequence of the input symbol block to obtain a loop output signal; the expression of the I / Q path separation type symbol phase discrimination method adopted by the phase discriminator is as follows: V D (t) = s Q (t) Sign(s I (t)) - s I (t) Sign(s Q (t)) Among them, V D (t) represents the output of the phase detector and the input of the low-pass filter, s I (t) and s Q (t) represent the real and imaginary parts of the downsampled sequence s(t) of the symbol block at time t, respectively, and Sign(·) represents the sign function; The system function F(z) of the low-pass filter is expressed as follows: Wherein, z is a complex variable on a complex plane, C1 and C2 are constants; The phase detector is jointly designed with a low-pass filter and combined into a loop filter with a joint gain G lf Instead of the selection of constants C1 and C2, the expression is as follows: O lf (t+1) = O lf (t) + G lf V D (t) wherein O lf (t),O lf (t+1) are the outputs of the loop filter at time t and t+1, respectively; S3, phase rotation is performed on the symbol constellation phase by using the demodulation result of the NB-IoT uplink demodulation reference signal, phase correction is completed, and a symbol sequence after frequency and phase correction is output.

2. The Costas ring based narrowband internet of things uplink single carrier symbol correction method according to claim 1, characterized in that, The extraction of the time domain symbol sequence is specifically as follows: after receiving the NB-IoT uplink single carrier signal, performing frequency down-conversion and analog-to-digital conversion, and then suppressing out-of-band interference through a band-pass filter, the time domain symbol sequence is extracted; the initial frequency offset estimation adopts a self-correlation algorithm based on a cyclic prefix and a cross-correlation algorithm of a pilot symbol.

3. The Costas ring based NB-IoT uplink single carrier symbol correction method according to claim 1, characterized in that, The extraction of the phase difference sequence in the phase discriminator is specifically as follows: a sliding window difference algorithm is adopted, complex conjugate multiplication operation is performed on adjacent sampling points in the window, the phase difference sequence is extracted, the window length is dynamically adjusted in the interval of 8-32 sampling points, and adaptive selection is performed according to the SNR detection result: when SNR>10dB, a short window is enabled; when SNR<5dB, a long window is switched to.

4. The Costas ring based NB-IoT uplink single carrier symbol correction method according to claim 1, characterized in that, After the Costas loop completes the processing of each symbol block, the demodulation loop output signal is output and correlation analysis is performed with a locally reconstructed constellation, an error is calculated as a feedback amount, and the joint gain value is dynamically adjusted.

5. The Costas ring based NB-IoT uplink single carrier symbol correction method according to claim 1, characterized in that, The Costas loop performs iterative verification on a symbol-by-symbol block basis, specifically as follows: the time domain sampling points of each symbol are phase-aligned to eliminate residual errors of the frequency offset; cyclic redundancy check is continuously performed in the NB-IoT terminal in the previous retransmission to determine the reliability of the current data packet; when the cyclic redundancy check passes, a final symbol sequence after frequency and phase correction is output.

6. The Costas ring based NB-IoT uplink single carrier symbol correction method according to claim 1, characterized in that, The corrected symbol is input into an NB-IoT narrowband channel estimation module, a channel impulse response is extracted, and the compensation parameters of the Costas loop, i.e. the joint gain and the sliding window length, are dynamically adjusted according to the time delay spread and the multipath intensity.

7. The Costas ring based NB-IoT uplink single carrier symbol correction method according to claim 1, c h a r a c t e r i z e d b y, The hardware-level optimization of the resource limitation of the software radio device includes: for the USRP, converting the floating-point calculation of the Costa loop into 8-bit fixed-point operation, and realizing the parallel pipeline processing of phase detection and compensation through CUDA acceleration; for the USRP, deploying the hardware acceleration module of the Costa loop through the FPGA programmable interface to match the peak rate requirement of the NB-IoT; and realizing the symbol block-level sleep mechanism on the software radio device to close the calculation function of the Costa loop when there is no data input.

8. The Costas ring based NB-IoT uplink single carrier symbol correction method of claim 1, wherein, The symbol correction method is compatible with at least one of the following narrowband Internet of Things protocols: the 3.75 kHz single-carrier uplink mode in the NB-IoT protocol, and the case of using only a single carrier in the 15 kHz uplink mode in the NB-IoT protocol. The symbol correction method supports joint correction with the downlink, including: pre-calibrating the initial frequency offset of the Costa loop by using the downlink reference signal, optimizing the joint gain and the sliding window length through the uplink-downlink channel reciprocity, and multiplexing the hardware resources of the Costa loop in the time division duplex (TDD) mode.

9. A Costas ring based narrowband internet of things uplink single carrier symbol correction system, characterized in that, The system for implementing the method according to any one of claims 1-8 comprises: a signal receiving and preprocessing module configured to receive an NB-IoT uplink single-carrier signal through a software radio device, extract a time-domain symbol sequence and divide it into a plurality of symbol blocks, perform initial frequency offset estimation and compensation on the symbol blocks, and then perform down-sampling processing to obtain a down-sampled sequence; a frequency correction module configured to construct a Costa loop correction structure, input the down-sampled sequence of the symbol blocks into the Costa loop correction structure, and complete frequency offset estimation of the down-sampled sequence; a phase correction module configured to perform phase rotation on the symbol constellation phase by using the demodulation result of the NB-IoT uplink demodulation reference signal, and complete phase correction; a signal output module configured to output the symbol sequence after frequency and phase correction to a demodulation module.

Images