Costas ring using pilot frequency self-adaptive sensitivity

Abstract

A Costas ring utilizing pilot frequency self-adaptive sensitivity comprises a frequency mixer which carries out frequency mixing processing on an input base band to obtain an in-phase component and an orthogonal component; the phase discriminator is used for calculating the in-phase component and the quadrature component, generating a phase error and providing the phase error to the loop filter; the loop filter is used for carrying out filtering and frequency phase calculation on the phase error so as to control the phase increment of the numerical control oscillator; the numerically-controlled oscillator generates local carriers according to the phase increment and sends the local carriers to the mixer; the combiner is used for combining the in-phase component and the orthogonal component into an output base band, and the output base band is the final output of the Costas ring; the correlator is used for calculating the cross-correlation between the output baseband and the local pilot frequency, and sending the calculated cross-correlation to the power estimator; the delayer is used for providing delay of a pilot symbol length and carrying out symbol alignment before an output baseband and a local pilot are input into the power estimator; a power estimator that estimates and outputs a signal-to-noise ratio based on the cross-correlation; according to the method, the dynamic characteristic and the anti-noise capability of the Costas ring on carrier tracking are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of communication, and particularly relates to a Costas loop using pilot adaptive sensitivity. BACKGROUND

[0002] Carrier synchronization technology is a key link in modern digital communication systems, and plays a crucial role in coherent demodulation. As a mature scheme for realizing carrier synchronization, the Costas loop is widely used in the synchronization processing of double-sideband suppressed carrier signals and phase shift keying (PSK) signals. In a broad sense, the Costas loop is actually a phase-locked loop (PLL) under specific phase tracking, and its basic principles and structures have been widely verified in communication systems.

[0003] As a closed-loop control system, the design of the phase detector and the loop filter in the Costas loop has a decisive influence on the dynamic performance of the system. Using a small-bandwidth loop filter and a low-gain phase detector can significantly improve the anti-noise performance of the system, but this will reduce the tolerance to frequency deviation; conversely, using a large-bandwidth loop filter and a high-gain phase detector can improve the adaptability of the system to frequency deviation changes, but its anti-noise performance will be significantly affected.

[0004] In terms of frequency deviation tracking, in a high-dynamic Doppler communication environment, when the carrier frequency deviation exceeds the capture range of the Costas loop, the system cannot achieve initial locking. Even if the initial locking is successful, severe frequency changes will also cause the tracking performance of the Costas loop to decline, and even cause the phenomenon of losing lock.

[0005] In terms of anti-noise performance, a low signal-to-noise ratio (SNR) environment not only increases the risk of false locking in the carrier synchronization process, but also easily causes communication link interruption, and each link interruption requires additional time and resources for re-synchronization. The existing Costas loop structure still has shortcomings when facing the above complex channel environment, and it is difficult to meet the strict requirements for synchronization performance under high-dynamic frequency deviation and low signal-to-noise ratio conditions.

[0006] The patent application file with publication number CN 105704080 A discloses an improved variable code rate carrier synchronization Costas loop design, aiming at variable code rate signals, dynamically calculating parameters according to the change of code rate through an external controller. The design features are that the parameters of the two-way low-pass filter inside and the loop filter are dynamically adjusted according to the code rate of the signal. However, due to the design scheme of controlling the parameters of the loop filter and the low-pass filter according to the code rate of the signal, and due to the existence and structure design of the low-pass filter which cannot perceive the signal characteristics, there are problems of delayed output signal and inability to dynamically adjust parameters according to signal characteristics. Therefore, it cannot meet the signal synchronization under the conditions of high dynamic frequency offset and low signal-to-noise ratio.

[0007] The patent application file with publication number CN 110324272 A discloses a fast carrier recovery system and method based on a combination ring, which uses the method of combining Costas ring and general carrier recovery ring to switch the use of Costas ring and general carrier recovery ring to improve the carrier tracking performance. The design features are that the traditional Costas ring is used to improve the rate of carrier tracking according to the fast tracking characteristics of the synchronization head single carrier signal. However, due to the design characteristics which cannot perceive the signal characteristics and cannot adjust the parameters of the loop internal structure, the loop structure is complex, and due to the existence of the low-pass filter, there are problems of delayed output signal and inability to dynamically adjust parameters according to signal characteristics. Therefore, it cannot meet the signal synchronization under the conditions of high dynamic frequency offset and low signal-to-noise ratio. SUMMARY

[0008] In order to overcome the shortcomings of the prior art, the purpose of the present application is to provide a pilot adaptive sensitivity Costas loop, which avoids the use of two-way low-pass filter by introducing a complex mixer, designs a phase discriminator and loop filter according to the adaptive sensitivity of signal-to-noise ratio, increases the perception ability of signal-to-noise ratio, and can adaptively adjust the loop bandwidth and output error jitter according to the signal-to-noise ratio of the signal link, has the advantages of simplifying the original design and greatly improving the dynamic characteristics and noise immunity of the Costas loop for carrier tracking.

[0009] In order to achieve the above purpose, the technical scheme adopted by the present application is:

[0010] A pilot adaptive sensitivity Costas loop, comprising:

[0011] The mixer 1 is used for mixing processing of the input baseband s IN , and performs frequency spectrum shift according to the local carrier s NCO output by the numerical control oscillator (NCO) 4; the input baseband s IN and the local carrier s NCOBoth are complex signals. Mixer 1 outputs two real signals, namely the processed in-phase components s. i and orthogonal components s q ;

[0012] Phase detector (PD) 2: Used to detect phase error and detect the in-phase component s of the mixer 1 output. i and orthogonal components s q Calculate and generate phase error Δ θ Provided to loop filter 3 to reflect the degree and direction of phase asynchrony;

[0013] Loop filter 3: Used to correct the phase error Δ generated by phase detector (PD) 2 θ Filtering and frequency-phase calculations are performed to control the phase increment of the numerically controlled oscillator (NCO)4.

[0014] Numerically controlled oscillator (NCO) 4: Receives phase increments from loop filter 3 According to phase increment Generate local carrier s NCO The signal is fed into mixer 1;

[0015] Combiner 5: Used to combine the in-phase component s of the mixer 1 output. i and orthogonal components s q Combined into output baseband s OUT Output baseband s OUT This is the final output of the Costas ring;

[0016] Correlator 6: Used to calculate the output baseband s of merger 5 OUT With local pilot s p cross-correlation R p The cross-correlation R is calculated. p Input to the power estimator (PE) 8;

[0017] Delay unit 7: Used to provide a delay of pilot symbol length D, reducing the output baseband s of combiner 5. OUT With local pilot s p Then perform sign alignment before inputting into the power estimator (PE) 8;

[0018] Power estimator (PE) 8: Used to calculate the input signal power Pr of delay unit 7 within the pilot symbol length D, based on the cross-correlation R of correlator 6 output. p To estimate and output the signal-to-noise ratio (SNR).

[0019] Furthermore, the phase detector (PD) 2 includes:

[0020] Multiplier 9: Used to multiply the in-phase component s of the mixer 1 output. iMultiply by the signal-to-noise ratio (SNR) to perform signal weighting, and obtain the weighted in-phase component s. i •SNR;

[0021] S-converter 10: Used to process the in-phase component s-s after weighting multiplier 9. i • SNR implements S-shaped transformation;

[0022] Multiplier 11: Used to combine the output of S-type converter 10 with the quadrature component s of the output of mixer 1. q Multiplication generates phase error Δ θ .

[0023] Furthermore, the loop filter 3 includes:

[0024] Proportional path gainer 12: Used to receive the phase error Δ generated by phase detector (PD) 2 θ This is then multiplied by the scaling gain α for rapid phase correction. The scaling gain α is a parameter associated with the loop bandwidth, which is derived from the signal-to-noise ratio (SNR).

[0025] Integrating path gainer 13: Used to receive the phase error Δ generated by phase detector (PD) 2 θ This is then multiplied by the integral gain β, which is used to adjust the loop's ability to eliminate steady-state frequency errors. The integral gain β is a parameter associated with the loop bandwidth, which is derived from the signal-to-noise ratio (SNR).

[0026] Adder 14: Used to add the output of the integrator path gainer 13 to the output of the delay unit 15, realizing digital integration and obtaining the frequency error Δ. f ;

[0027] Delay unit 15: Implements a unit sampling delay, used to adjust the frequency error Δ output by adder 14. f Delay by one sampling period for the integration calculation at the next sampling time of adder 14;

[0028] Limiter 16: Used to limit the frequency error Δ of the adder 14 output. f The purpose is to limit the phase increment of the numerically controlled oscillator (NCO)4. The rate of change;

[0029] Adder 17: Used to add the output of the proportional path gainer 12 to the output of the limiter 16 to generate a phase increment.

[0030] The phase detector (PD)2 uses an S-type function for phase detection.

[0031] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0032] 1. Compared with the prior art, the application can simplify the hardware design structure, reduce signal processing delay and improve system real-time performance by adopting the complex mixer 1 and removing the IQ branch low-pass filter.

[0033] 2. The application can realize parameter dynamic adjustment by the improved S-type function phase discriminator (PD) 2 design and pilot-assisted power estimator (PE) 8, can adaptively adjust the sensitivity of the phase discriminator (PD) 2 and the bandwidth of the loop filter 3: reduce the loop bandwidth to suppress noise disturbance when the signal signal-to-noise ratio SNR is low, and increase the bandwidth to improve the tracking sensitivity when the signal signal-to-noise ratio SNR is high, effectively solving the performance contradiction of the traditional fixed parameter design when the signal-to-noise ratio fluctuates.

[0034] 3. The application realizes signal signal-to-noise ratio SNR sensing by using the output baseband s OUT of the combiner 5 to calculate the cross-correlation R p of the local pilot s p , can output the current Doppler shift and calculate the relative motion speed of the transmitting and receiving ends, and breaks through the limitation that the traditional structure cannot sense the motion parameters of the communication object.

[0035] 4. Gain brought by the improved phase discriminator (PD) 2:

[0036] The conclusion obtained when the signal signal-to-noise ratio SNR>>1 is consistent with the disturbance of the traditional Costas loop, because the noise itself is extremely small, and the influence on the signal disturbance is also very small.

[0037] When the signal signal-to-noise ratio SNR<<1, it can be proved that when the signal-to-noise ratio is low: (1). The phase error accumulation amount of the phase discriminator (PD) 2 caused by noise is 0 (E(Δ θ )=SNR·P s ·θ), which can prove that the noise will not affect the tracking performance within the accumulation time. (2). The disturbance amount of the phase discriminator (PD) 2 output of the application under low signal-to-noise ratio is SNR 2 times that of the traditional phase discrimination method. Therefore, the new phase discrimination method of the application can effectively resist noise and not affect the performance under high signal-to-noise ratio.

[0038] 5. Gain brought by the improved loop filter 3:

[0039] The loop bandwidth is controlled by the signal signal-to-noise ratio SNR, and then the α and β parameters are affected, and the two parameters directly affect the sensitivity of the loop filter 3 to the input. Since the bandwidth of the loop filter 3 is increased, the sensitivity of the Costas loop under high signal-to-noise ratio is increased; the bandwidth of the loop filter 3 is reduced, and the noise resistance of the Costas loop under low signal-to-noise ratio is increased;

[0040] In conclusion, the application improves the tracking accuracy under low signal-to-noise ratio and reduces the disturbance caused by noise by improving the phase detector (PD) 2 and the loop filter 3 together; the adaptive loop parameter adjustment realized by the complex mixer 1 and the pilot-assisted power estimator (PE) 8 has the advantages of reducing the output delay of the signal, improving the real-time performance of the system, improving the tracking accuracy under low signal-to-noise ratio and the sensitivity under high signal-to-noise ratio. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a schematic diagram of the improved Costas loop structure of the application.

[0042] Figure 2 is a schematic diagram of the improved phase detector (PD) 2 structure of the application.

[0043] Figure 3 is a schematic diagram of the improved loop filter 3 structure of the application. DETAILED DESCRIPTION

[0044] The application will be further described below in combination with the drawings and specific embodiments.

[0045] As shown in Figure 1 , the embodiment proposes an improved Costas loop for tracking and synchronizing BPSK signals. The embodiment takes a binary phase shift keying signal (BPSK, Binary Phase Shift Keying) as an example, and assumes that the received input baseband s IN is:

[0046]

[0047] Where A is the signal amplitude, m(t) is the binary symbol waveform, and takes values {-1, +1}, f c is the received residual carrier, and n(t) is the Gaussian white noise.

[0048] The mixer 1 is used for mixing the input baseband s IN , and the purpose is to perform spectrum shift according to the local carrier s NCO output by the numerically controlled oscillator (NCO) 4; the input baseband s IN and the local carrier s NCO are both complex signals.

[0049] When the improved Costas loop is locked, the input baseband s IN and the local carrier s NCO output by the numerically controlled oscillator (NCO) 4 are mixed via the mixer 1, and the in-phase component s i and the quadrature component s q after processing have a signal center frequency reduced to zero, at which time the in-phase component s iand quadrature component s q The expression:

[0050] s i = AM(t)cos(θ) + n i (t)

[0051] s q = -AM(t)sin(θ) + n q (t)

[0052] where θ is the phase difference between the input baseband s IN and the local carrier s NCO after mixing by mixer 1, the in-phase component s i and the quadrature component s q satisfy the in-phase component signal energy P(s i )<0, and the quadrature component signal energy P(s q )<0.

[0053] Phase detector (PD) 2: for detecting phase error, calculating the in-phase component s i and the quadrature component s q output by mixer 1, and generating phase error Δ θ provided to loop filter 3, reflecting the degree and direction of phase misalignment; the S-shaped function used in phase detector (PD) 2 is exemplified by tanh in this embodiment, as shown in Figure 2 Further, phase detector (PD) 2 includes:

[0054] Multiplier 9: multiplying the in-phase component s i output by mixer 1 with signal-to-noise ratio SNR, performing signal weighting, and obtaining weighted in-phase component s i · SNR;

[0055] S-shaped transformer 10: implementing S-shaped transformation on the weighted in-phase component s i · SNR output by multiplier 9; the S-shaped function of this embodiment is exemplified by tanh;

[0056] Multiplier 11: multiplying the output of S-shaped transformer 10 with the quadrature component s q output by mixer 1, and generating phase error Δ θ .

[0057] For phase detector (PD) 2, more specifically:

[0058] The mathematical expression for the operation of phase error Δ θ output by phase detector (PD) 2:

[0059] Δ θ = tanh(s i·SNR)·s q

[0060] like Figure 3 As shown, loop filter 3 is used to filter the phase error Δ generated by phase detector (PD) 2. θ Filtering and frequency-phase calculations are performed to control the phase increment of the numerically controlled oscillator (NCO)4. Furthermore, the loop filter 3 includes:

[0061] Proportional path gainer 12: Receives the phase error Δ generated by phase detector (PD) 2. θ This is then multiplied by the scaling gain α for rapid phase correction. The scaling gain α is a parameter associated with the loop bandwidth, which is derived from the signal-to-noise ratio (SNR).

[0062] Integrating path gainer 13: Phase error Δ generated by receiver phase detector (PD) 2 θ This is then multiplied by the integral gain β, which is used to adjust the loop's ability to eliminate steady-state frequency errors. The integral gain β is a parameter associated with the loop bandwidth, which is derived from the signal-to-noise ratio (SNR).

[0063] The proportional gain α is given by the following formula:

[0064]

[0065] The integral gain β is given by the following equation:

[0066]

[0067] in, B is the damping factor. L This is the loop bandwidth of loop filter 3; where the loop bandwidth B L Controlled by signal-to-noise ratio:

[0068] B L =g(SNR)

[0069] In the above formula, g(·) represents the signal-to-noise ratio (SNR) to the loop bandwidth B. L For transitive relationships, a linear function relationship is generally sufficient to meet most application requirements.

[0070] Adder 14: Adds the output of the integrator path gainer 13 to the output of the delay unit 15 to achieve digital integration and obtain the frequency error Δ. f ;

[0071] Delay unit 15: Implements a unit sampling delay, and converts the frequency error Δ output by adder 14 to the unit sampling delay. f The delay is one sampling period, which is used for the integration calculation at the next sampling time of adder 14.

[0072] For adder 14 and delay 15, more specifically:

[0073] Δ f [n] = Δ f [n-1] + β·Δ θ

[0074] where Δ f [n] represents the frequency error obtained by adder 14 at the current time, Δ f [n-1] represents the frequency error from delay 15 at the previous time.

[0075] For frequency error Δ f , additionally, to calculate the relative velocity v d between the transmitter and receiver, the following formula is used:

[0076]

[0077] where λ is the carrier wavelength;

[0078] Clipper 16: used to limit the frequency error Δ f output by adder 14, for the purpose of limiting the rate of change of the phase increment of numerically controlled oscillator (NCO) 4.

[0079] so that frequency error Δ f satisfies:

[0080] f min < Δ f < f max

[0081] where f min is the minimum frequency offset, f max is the maximum frequency offset;

[0082] Adder 17: adds the output of proportional path gain 12 and the output of clipper 16 to generate phase increment

[0083]

[0084] Numerically controlled oscillator (NCO) 4: receives phase increment from loop filter 3 to generate local carrier s NCO , which is sent to mixer 1; more specifically,

[0085] s NCO [n] = exp{-j(φ[n])}

[0086]

[0087] where φ[n] represents the output phase of the current time of the numerically controlled oscillator (NCO) 4, and φ[n-1] represents the output phase of the last time of the numerically controlled oscillator (NCO) 4;

[0088] Combiner 5: used to combine the in-phase component s i and the quadrature component s q outputted by the mixer 1 into an output baseband s OUT , the output baseband s OUT is the final output of the Costas loop;

[0089] s OUT = s i + js q

[0090] Correlator 6: used to calculate the cross-correlation R OUT of the output baseband s p of the combiner 5 and the local pilot s p , the calculated cross-correlation R p is sent to the power estimator (PE) 8; the following formula is the cross-correlation R p outputted by the correlator 6:

[0091]

[0092] or expressed as:

[0093]

[0094] where s p is the pilot with a power of 1 stored by the local receiver, and D is the pilot symbol length;

[0095] Delay 7: used to provide a delay of a pilot symbol length D, and to perform symbol alignment on the output baseband s OUT of the combiner 5 and the local pilot s p before inputting them again to the power estimator (PE) 8;

[0096] Power estimator (PE) 8: based on the cross-correlation R p outputted by the correlator 6, to estimate and output the signal-to-noise ratio SNR.

[0097] As to the power estimator (PE) 8, more specifically:

[0098] Supposing that the cross-correlation R p peak is obtained at t0 time or m0 sampling point, the pilot signal power P s and the input signal power P r are obtained by the following formula:

[0099]

[0100] Noise power is obtained from

[0101]

[0102] The signal-to-noise ratio SNR is obtained as

[0103]

[0104] Experimental analysis

[0105] The performance gain over the conventional Costas loop is demonstrated as follows:

[0106] For the phase detector:

[0107] The expression for the operation of the phase detector 2 (the S-shaped function in this embodiment is exemplified by tanh): Figure 3 ) is:

[0108] Δ θ = tanh(s i · SNR) · s q

[0109] When the signal-to-noise ratio SNR » 1, the noise n(t) satisfies:

[0110] n(t) « 0

[0111] and when the in-phase component s i > 0, tanh(·) tends to 1; and when s i < 0, tanh(·) tends to -1, in which case the phase error Δ θ :

[0112] Δ θ = sign(s i ) · s q

[0113] The symbol waveform m(t) = ± 1 of the BPSK signal and the phase error Δ θ can be approximated as:

[0114] Δ θ = m 2 (t) sin(θ) = sin(θ)

[0115] When Δ θ is small (because the loop is locked):

[0116] sin(θ) « θ

[0117] Δ θ = sin(θ) « θ

[0118] The phase error Δθ The output is the phase difference θ;

[0119] When the signal-to-noise ratio SNR << 1:

[0120]

[0121] where A is the signal amplitude, Pn is the noise power, and P s is the signal power.

[0122] Considering the instantaneous energy of the analysis signal, let the in-phase component s i be weighted by the signal-to-noise ratio x1 = s i · SNR:

[0123]

[0124] The symbol waveform m(t) and the cos(θ) amplitude are limited, and the noise power >> the signal power P s , both terms of the above equation are << 1.

[0125] Considering the Taylor expansion of the tanh function

[0126]

[0127] There is an approximation tanh(x1) ≈ x1 = s i · SNR, and the phase error Δ θ can be approximated as:

[0128] Δ θ ≈ s i · s q · SNR

[0129] Δ θ

[0130] ≈ SNR· [A 2 m 2 (t) · cos θ · sin θ + Am(t) · cos θ · n q (t) + Am(t) · sin θ · n i (t) + n i (t) · n q (t)]

[0131] The mathematical expectation E[Δ θ ] of the phase error can be used to characterize the steady-state characteristics:

[0132] Considering the equal probability of the symbol waveform m(t) = ±1, the four terms in the mathematical expectation E[Δ θ ] of the phase error are respectively:

[0133]

[0134] E[Am(t)]·n Q (t)] = E[m(t)]·E[n Q (t)] = 0

[0135] E[Am(t)·n I (t)] = E[m(t)]·E[n I (t)] = 0

[0136] E[n I (t)·n Q (t)] = 0

[0137] Thus the average error function:

[0138]

[0139] When the phase difference θ is very small (again because of the loop locking), we have:

[0140] E(Δ θ ) = SNR·P s ·θ

[0141] It can be seen that the accumulation of the output phase error is mainly affected by the signal-to-noise ratio at low signal-to-noise ratio, and the output phase error is weighted by the signal energy at the same time.

[0142] Now let's analyze its instantaneous error. Considering that the phase error contribution in the accumulation time of the mathematical expectation is completely derived from the high-order small signal quantity, let's now analyze the impact of noise.

[0143] Let the phase error Δ θ be composed of a signal term Δ s and a noise term Δ n :

[0144] Δ θ = Δ s + Δ n

[0145]

[0146] Δ n = SNR·[Am(t)cosθ·n q (t) + Am(t)sinθ·n i (t) + n i (t)n q (t)]

[0147] The variance of the signal term Δ s :

[0148]

[0149] Re-decomposed noise term Δ n For the first-order noise term and the second-order noise term:

[0150]

[0151] Noise term Δ n The first-order noise term in the middle Variance:

[0152]

[0153] Noise term Δ n The second-order noise term in the middle Variance:

[0154]

[0155] Therefore, the variance of the noise term Δ n

[0156]

[0157] Contrast analysis of the phase detector of the classical Costas loop:

[0158] Phase error Δ' of the classical Costas loop θ :

[0159] Δ' θ = s' i · s' q

[0160] = A 2 m 2 (t)cosθsinθ+A 2 m(t)cosθ·n q (t)+A 2 m(t)sinθ·n i (t)+n i (t)n q (t) According to the above derivation analogy, the signal quantity Δ' of the classical Costas loop s Variance and noise term Δ' n Variance:

[0161]

[0162] The ratio of the output energy of the phase detector proposed in the present application to that of the traditional phase detector:

[0163]

[0164] ​It can be seen that when SNR is much less than 1, the output jitter of the phase discriminator (PD) 2 proposed by the application is much less than that of the classical phase discriminator, and this feature enables the phase discriminator to track the carrier offset well and be affected by noise as little as possible when the noise is large.

[0165] The key point and protection point of the application are:

[0166] 1. The output of the numerically controlled oscillator (NCO) 4 is a complex signal, which has the ability to represent negative frequencies, and can be directly mixed with the input signal to complete spectrum shifting (the traditional scheme is to mix two coherent carriers and then filter). The technical features are that the IQ path independent low-pass filter is cancelled, and the complex domain mixing method is used, which does not produce image frequencies.

[0167] 2. The method and process of using pilot inter-correlation operation. The technical features are that a signal-to-noise ratio estimator is constructed, and the signal-to-noise ratio SNR is calculated by the correlation peak R p (0) The method for calculating signal power, noise power and signal-to-noise ratio.

[0168] 3. The output Δ f of the loop filter integration branch can be regarded as the channel Doppler in an ideal case. The technical features are that the carrier wavelength λ and the output Δ f of the loop filter integration branch are used to calculate the relative motion speed v d between the transmitter and the receiver (the traditional structure does not use this information because it cannot determine the loop locking state).

[0169] 4. The adaptive variable feature of the bandwidth of the loop filter 3, characterized in that the bandwidth of the loop filter 3 is controlled by the signal-to-noise ratio SNR, and a response function g(·) is used to transfer the signal-to-noise ratio SNR to the bandwidth of the loop filter.

[0170] 5. The improved phase discriminator (PD) 2 of the application uses the signal-to-noise ratio SNR as the sensitivity, and the output phase error Δ θ is affected by noise. The technical features are that the signal-to-noise ratio SNR is introduced to control the sensitivity of the phase discriminator (PD) 2, and a sigmoid saturation function (tanh is used as an example in the embodiment) is used to realize high sensitivity at high signal-to-noise ratio and low sensitivity at low signal-to-noise ratio (the purpose is to reduce the influence of noise).

[0171] Based on the performance bottleneck of the existing Costas loop in high dynamic and Doppler frequency offset correction, and the technical defects of insufficient synchronization reliability in a low signal-to-noise ratio environment, the application aims to solve the following key problems:

[0172] 1. Eliminate hardware dependency and improve real-time performance: In view of the hardware resource consumption and processing delay caused by IQ path independent low-pass filter in traditional complex mixing frequency architecture, the invention directly eliminates the image interference through baseband complex signal processing, simplifies the hardware configuration (reduces two low-pass filters) while reducing the processing delay.

[0173] 2. Overcome the influence of dynamic fluctuation of signal-to-noise ratio: In view of the contradiction between capture speed and tracking accuracy when the signal-to-noise ratio fluctuates in the design of fixed loop bandwidth, the invention adjusts the loop filter (B L ) parameter in the technical scheme, better adapts to the changing channel conditions, and is particularly suitable for processing high dynamic Doppler correction and carrier frequency offset correction in extremely low signal-to-noise ratio scenarios.

[0174] 3. Realize the motion parameter sensing function: The invention breaks through the limitation of traditional loop that can only recover the carrier parameters, estimates the Doppler of the communication signal, and further estimates the relative motion speed of the communication parties, while also more accurately estimating the signal-to-noise ratio of the communication parties.

[0175] 4. By using the phase discriminator with signal-to-noise ratio parameter to phase, enhance the phase sensitivity, use the characteristics of S function to suppress large signal disturbance, improve the sensitivity of weak signal in low signal-to-noise ratio SNR, and suppress strong noise jitter in high signal-to-noise ratio SNR.

[0176] Compared with the prior art, the innovation of the invention lies in:

[0177] 1. The classic Costas loop needs to configure I and Q low-pass filters after complex mixing frequency to suppress image frequency interference, which increases hardware resource consumption and introduces processing delay. The invention uses mixer 1 to directly process complex signals at baseband, eliminates image frequency interference without using physical low-pass filters, simplifies hardware implementation, and reduces system complexity and processing delay.

[0178] 2. The traditional Costas loop can only recover the carrier frequency and phase through phase tracking, and cannot directly perceive the motion parameters of the communication object. The invention uses combiner 5, correlator 6, delay device 7 and power estimator (PE) 8, uses the output baseband s OUT of combiner 5 to calculate the cross-correlation R p with the local pilot s p , realize signal-to-noise ratio SNR sensing, and use the internal structure of the Costas loop to estimate the Doppler of the communication signal when the local oscillators of the transmitter and receiver have no significant frequency difference, and further estimate the relative motion speed of the communication parties.

[0179] 3. The Costas loop tracks the carrier frequency offset by relying on the loop filter to process the phase error signal. The traditional solution has a contradiction between the acquisition speed and the tracking accuracy when the signal-to-noise ratio fluctuates. The present invention uses a correlator 6, a delay device 7 and a power estimator (PE) 8 to estimate the signal-to-noise ratio of the input signal by correlating the pilot sequence of the input signal with the local pilot, and to dynamically adjust the bandwidth of the loop filter 3; a larger loop bandwidth is used when the signal-to-noise ratio is high to improve the acquisition speed, and a smaller loop bandwidth is used when the signal-to-noise ratio is low to improve the tracking accuracy, thus balancing the contradiction between the acquisition speed and the tracking accuracy in high dynamic scenarios, and also accurately estimating the signal-to-noise ratio of the communication parties.

[0180] 4. The traditional Costas loop uses a multiplier or an arctangent function as a phase detector, and the phase detector (PD) 2 used in the present invention uses an S-shaped function (in the embodiment, the hyperbolic tangent function tanh) for phase detection, and the signal-to-noise ratio weighted by the power estimator (PE) 8 is used to complete the phase detection. The S-shaped function is approximately linear near zero, and tends to saturation away from zero. The phase detector used in the present invention has good sensitivity to small signal changes, strong suppression ability to large signal disturbances, and realizes natural amplitude limiting of noise, thus improving the stability of the system.

[0181] In addition, the technical solutions of the present invention are:

[0182] 1. The S-shaped function used in the phase detector (PD) 2 is analyzed by taking tanh as an example in the embodiment. Any S-shaped function can achieve the effect in actual application.

[0183] 2. The mixer 1 in the present invention can still be replaced by the traditional method (i.e. the scheme of real signal mixing + low pass filtering).

[0184] 3. The loop filter 3 is exemplified by a proportional-integral controller, and those that can achieve the same method can be extended to PID (proportional-integral-derivative) controllers and higher / lower order loop filters.

[0185] 4. For the output signal s OUT , the method of the present invention combines two real signals, i.e. the in-phase component s i and the quadrature component s q into a complex baseband signal, so that the error quantity represented by the quadrature component s q can be transmitted backward. Therefore, the scheme of taking s OUT = s i as a real signal output should also be protected by the present invention. For this change, the combiner 5 will no longer be needed.

Claims

1. A Costas loop utilizing pilot-frequency adaptive sensitivity, characterized in that, include: Mixer (1): used for mixing the input baseband signal. IN The mixing process is based on the local carrier s output by the numerically controlled oscillator (NCO) (4). NCO Perform spectrum shifting; its input baseband s IN and local carriers NCO Both are complex signals. The mixer (1) outputs two real signals, namely the processed in-phase components s. i and orthogonal components s q ; Phase detector (PD)(2): Used to detect phase error and measure the in-phase component s of the mixer (1) output. i and orthogonal components s q Calculate and generate phase error Δ θ Provided to the loop filter (3) to reflect the degree and direction of phase asynchrony; Loop filter (3): used to filter the phase error Δ generated by the phase detector (PD) (2). θ Filtering and frequency-phase calculations are performed to control the phase increment of the numerically controlled oscillator (NCO) (4). Numerically controlled oscillator (NCO) (4): used to receive phase increments from the loop filter (3). According to phase increment Generate local carrier s NCO , and send it into mixer (1); Combiner (5): Used to combine the in-phase component s of the mixer (1) output. i and orthogonal components s q Combined into output baseband s OUT Output baseband s OUT This is the final output of the Costas ring; Correlator (6): Used to calculate the output baseband s of the merger (5) OUT With local pilot s p cross-correlation R p The cross-correlation R is calculated. p Input to the power estimator (PE)(8); Delay unit (7): Used to provide a delay of a pilot symbol length D, which will reduce the output baseband of the combiner (5) by s. OUT With local pilot s p Then, sign alignment is performed before inputting into the power estimator (PE)(8); Power estimator (PE)(8): Used to calculate the input signal power Pr of the delay unit (7) within the pilot symbol length D, and based on the cross-correlation R of the correlator (6) output. p To estimate and output the signal-to-noise ratio (SNR).

2. A Costas loop utilizing pilot-adaptive sensitivity according to claim 1, characterized in that, The phase detector (PD)(2) includes: Multiplier (9): Used to convert the in-phase component s of the mixer (1) output. i Multiply by the signal-to-noise ratio (SNR) to perform signal weighting, and obtain the weighted in-phase component s. i •SNR; S-type converter (10): used to weight the in-phase component s of the multiplier (9). i • SNR implements S-shaped transformation; Multiplier (11): Used to combine the output of the S-type converter (10) with the quadrature component s of the output of the mixer (1). q Multiplication generates phase error Δ θ .

3. A Costas loop utilizing pilot adaptive sensitivity according to claim 1, characterized in that, The loop filter (3) includes: Proportional path gainer (12): Used to receive the phase error Δ generated by the phase detector (PD) (2). θ This is then multiplied by the scaling gain α for rapid phase correction. The scaling gain α is a parameter associated with the loop bandwidth, which is derived from the signal-to-noise ratio (SNR). Integrating path gainer (13): Used to receive the phase error Δ generated by the phase detector (PD) (2). θ This is then multiplied by the integral gain β, which is used to adjust the loop's ability to eliminate steady-state frequency errors. The integral gain β is a parameter associated with the loop bandwidth, which is derived from the signal-to-noise ratio (SNR). Adder (14): Used to add the output of the integral path gainer (13) and the output of the delay unit (15) to realize the digital integration function and obtain the frequency error Δ. f ; Delay unit (15): Implements a unit sampling delay, used to adjust the frequency error Δ output by the adder (14). f Delay by one sampling period for the integral calculation of the next sampling time of adder (14); Limiter (16): Used to limit the frequency error Δ of the adder (14) output. f The purpose is to limit the phase increment of the numerically controlled oscillator (NCO)(4). The rate of change; Adder (17): Used to add the output of the proportional path gainer (12) to the output of the limiter (16) to generate a phase increment.

4. A Costas loop utilizing pilot-adaptive sensitivity according to claim 1, characterized in that, The phase detector (PD)(2) uses an S-type function for phase detection.

Images