Broadband image rejection method based on combination of LS algorithm and MP model

Abstract

The invention discloses a broadband image rejection method based on combination of an LS algorithm and an MP model, and belongs to the field of communication. According to the method, a method of combining a least square algorithm (LS algorithm) and a memory polynomial model (MP) describing nonlinear response is adopted, and local oscillator leakage, mirror image frequency distortion and in-band uneven distortion in a zero intermediate frequency transceiver passband are effectively compensated; in-band uneven distortion is effectively suppressed, the suppression effect on mirror image frequency distortion is better, and the applicable frequency band range is wider; the method not only can be used for independently carrying out image rejection compensation on a zero-intermediate-frequency transmitter or receiver, but also can be used for carrying out closed-loop adaptive image rejection compensation on a zero-intermediate-frequency transceiver, so that the problem that two IQ imbalance distortions of the transmitter and the receiver are aliasing together and cannot be distinguished and corrected when no reference instrument exists is solved; and a technical support is provided for a more ideal test channel environment of the zero-intermediate-frequency signal transceiver.

Description

Technical Field

[0001] The invention belongs to the field of communications, and in particular relates to a broadband image suppression method based on the combination of an LS algorithm and an MP model. Background Art

[0002] With the development of modern communication systems, the image suppression technology of traditional narrowband receivers can no longer meet the requirements of various communication standards. Wideband image suppression technology is mainly used to solve the image interference problem in broadband receivers and adapt to the requirements of various communication frequency band standards. However, driven by the rapid development of wireless communication technology, this technology faces new opportunities and challenges in suppression effect, compatibility and reusability design, low power design, high integration design, design difficulty and cost control.

[0003] The existing broadband image suppression technology mainly includes high image suppression mixer design, precise hardware circuit design and debugging, digital domain signal compensation algorithm, etc. The high image suppression mixer design mainly adopts double balanced mixer, which is small in size, highly integrated, and covers the entire S band. In the S band, the image suppression is greater than 20dB, and the single sideband suppression is greater than 20dB, which effectively suppresses the sideband distortion of the image frequency. The precise hardware circuit design package uses multi-channel down-conversion and intermediate frequency filtering to change the frequency of the local oscillator and intermediate frequency filter; there are also methods such as using a 0.45V low-power noise amplifier, and developing a variable resonant LC resonant circuit, a common-gate amplifier broadband gain enhancement and image suppression circuit. In addition, there is also a broadband image suppression mixer designed using TSMC's 0.18um CMOS process that works between 3 and 3.4GHz, and the applicable frequency band range is not wide. Compared with the first two broadband image suppression technologies, the digital domain signal compensation algorithm is implemented by software, which has the advantages of high reusability, low power consumption, low implementation complexity and low cost. At present, broadband image suppression technology in the digital domain mainly includes parameter estimation compensation method based on statistical characteristics of IQ imbalance model and blind estimation calibration compensation method based on LS algorithm. It can effectively suppress image frequency distortion, and the method is simple and has low implementation complexity.

[0004] The existing broadband image suppression technology mainly includes mixer design with high image suppression, precise hardware circuit design and debugging, and signal compensation algorithm in the digital domain.

[0005] Disadvantages of existing technology:

[0006] The design method of mixers with high image suppression mainly uses double-balanced mixers, which are relatively small in size and highly integrated, but have certain requirements on hardware, high implementation complexity, low compatibility and reusability, and limited frequency band range and suppression effect.

[0007] Precise hardware circuit design utilizes multi-channel down-conversion and intermediate frequency filtering, changes the frequency of the local oscillator and intermediate frequency filter, and adopts 0.45V low-power noise amplifier, and develops variable resonant LC resonant circuit, broadband gain enhancement of common-gate amplifier and image suppression circuit. Such methods have high hardware requirements, including precise wiring, matching component selection and high-quality manufacturing process, so the cost and implementation complexity are also high, compatibility and reusability are low, integration is low, and the applicable frequency band range is also limited.

[0008] Currently, broadband image suppression technology in the digital domain mainly includes parameter estimation compensation methods based on the IQ imbalance model with statistical characteristics and blind estimation calibration compensation methods based on the LS algorithm. The methods are simple and low in complexity, but the compensation effect is limited and does not take into account local oscillator leakage distortion and in-band uneven distortion. Summary of the invention

[0009] In view of the above technical problems existing in the prior art, the present invention proposes a broadband image suppression method based on the combination of LS (Least Square) algorithm and MP (Memory polynomial) model, which has a reasonable design, overcomes the shortcomings of the prior art and has good effects.

[0010] In order to achieve the above object, the present invention adopts the following technical solution:

[0011] A broadband image suppression method based on a combination of an LS algorithm and an MP model comprises the following steps:

[0012] Step 1: Perform parameter training; specifically includes the following steps:

[0013] Step 1.1: Eliminate local oscillator leakage;

[0014] Eliminate the error caused by local oscillator leakage and perform local oscillator leakage compensation on IQ data respectively;

[0015] Step 1.2: Data alignment;

[0016] Step 1.3: Configure the training model and order;

[0017] Step 1.4: Estimate compensation parameters of in-band uneven distortion;

[0018] Step 1.5: Compensate for the uneven distortion parameters within the band;

[0019] After obtaining the compensation parameters of the in-band uneven distortion, the in-band unevenness compensation is performed on the processed received signal;

[0020] Step 1.6: Estimate compensation parameters of image frequency distortion;

[0021] Step 1.7: After obtaining the compensation parameters of the image frequency distortion, the parameter training process is completed;

[0022] Step 2: Perform parameter compensation; specifically including the following steps:

[0023] Step 2.1: Compensate for mirror distortion;

[0024] Step 2.2: Evaluate the inhibitory effect;

[0025] If: the result of the judgment is that the inhibitory effect is not achieved, then execute step 1.3;

[0026] Or if the result is that the inhibitory effect is achieved, execute step 2.3:

[0027] Step 2.3: Image frequency distortion compensation;

[0028] Step 2.4: Compensate for uneven distortion within the band;

[0029] Step 2.5: Eliminate local oscillator leakage;

[0030] Step 2.6: End.

[0031] Preferably, in step 1.2, data alignment is performed, that is, data alignment is performed on the original input signal and the received signal, including integer alignment and fractional alignment processes, so as to eliminate errors caused by sampling at the receiving end.

[0032] Preferably, in step 1.4, the compensation parameters of the in-band uneven distortion include time domain filtering, power and phase compensation and LS algorithm solution, and the filter cutoff frequency and order are correctly configured to ensure the accuracy of the training results.

[0033] Preferably, in step 1.6, the compensation parameters of the mirror frequency distortion include time domain filtering, power and phase compensation and MP model algorithm solution, and the filter cutoff frequency, order and order of the MP model are correctly configured to ensure the accuracy of the training results.

[0034] Beneficial technical effects brought by the present invention:

[0035] The present invention adopts a method combining a least squares algorithm (LS algorithm) and a memory polynomial model (MP) describing nonlinear response to effectively compensate for local oscillator leakage, image frequency distortion and in-band uneven distortion in the passband of a zero intermediate frequency transceiver; compared with other compensation methods using statistical characteristics and blind estimation calibration algorithms such as the LS algorithm, the present invention effectively suppresses in-band uneven distortion, has a better suppression effect on image frequency distortion, and is applicable to a wider frequency band.

[0036] The present invention can not only be used to perform image suppression compensation on a zero intermediate frequency transmitter or receiver separately, but also can perform closed-loop adaptive image suppression compensation on a zero intermediate frequency transceiver, thereby solving the problem that the two types of IQ imbalance distortion of the transmitter and receiver are mixed together and cannot be distinguished and corrected when there is no reference instrument, and provides technical support for a more ideal test channel environment for a zero intermediate frequency signal transceiver. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0038] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0039] The present invention utilizes a method combining the LS algorithm and the MP model to estimate the in-band unevenness parameters and the image distortion parameters respectively, and reduces or eliminates the image frequency distortion and in-band unevenness distortion caused by image interference on the basis of eliminating the local oscillator leakage distortion by compensating the signals at the transmitting end / receiving end in the digital domain.

[0040] The specific technical solution is described as follows:

[0041] 1. First, eliminate the error caused by local oscillator leakage and perform local oscillator leakage compensation on the IQ data respectively;

[0042] 2. Next, data alignment is performed, that is, the original input signal (baseband signal) is aligned with the received signal, including integer alignment and fractional alignment processes, to eliminate the error caused by sampling at the receiving end;

[0043] 3. Then, the compensation parameters of the uneven distortion in the band are estimated, including time domain filtering, power and phase compensation, and LS algorithm solution. The filter cutoff frequency and order need to be correctly configured to ensure the accuracy of the training results.

[0044] 4. After obtaining the compensation parameters of the in-band uneven distortion, the in-band unevenness compensation is performed on the processed received signal;

[0045] 5. Then estimate the compensation parameters of the image frequency distortion, including time domain filtering, power and phase compensation, and MP model algorithm solution. In addition to the filter cutoff frequency and order, the order of the MP model must also be correctly configured to ensure the accuracy of the training results;

[0046] 6. After obtaining the compensation parameters of the image frequency distortion, the parameter training process is completed;

[0047] 7. Next is the compensation process. The compensation methods of the transmitter and receiver are slightly different. Taking the transmitter as an example, the baseband signal is compensated for the image frequency and the in-band unevenness in turn.

[0048] 8. Finally, the compensation effect is verified at the output end of the transmitter / receiver, and the evaluation parameters include NMSE, EVM, etc.

[0049] The specific process is as follows Figure 1 shown.

[0050] The present invention uses a method combining the least squares algorithm (LS algorithm) and the memory polynomial model (MP) to estimate the in-band unevenness parameters and the image distortion parameters respectively, and reduces or eliminates the image frequency distortion and in-band unevenness distortion caused by image interference on the basis of eliminating the local oscillator leakage distortion by compensating the signal of the transmitter / receiver in the digital domain, thereby improving the overall performance of the transmitter / receiver. The present invention is applicable to various communication standard requirements, the image suppression degree is more than 50dB, and the error vector magnitude (EVM) is also greatly improved.

[0051] Of course, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by technicians in this technical field within the essential scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A broadband image suppression method based on the combination of LS algorithm and MP model, characterized in that: The steps include: Step 1: Perform parameter training; specifically includes the following steps: Step 1.1: Eliminate local oscillator leakage; Eliminate the error caused by local oscillator leakage and perform local oscillator leakage compensation on IQ data respectively; Step 1.2: Data alignment; Step 1.3: Configure the training model and order; Step 1.4: Estimate compensation parameters of in-band uneven distortion; Step 1.5: Compensate for the uneven distortion parameters within the band; After obtaining the compensation parameters of the in-band uneven distortion, the in-band unevenness compensation is performed on the processed received signal; Step 1.6: Estimate compensation parameters of image frequency distortion; Step 1.7: After obtaining the compensation parameters of the image frequency distortion, the parameter training process is completed; Step 2: Perform parameter compensation; specifically including the following steps: Step 2.1: Compensate for mirror distortion; Step 2.2: Evaluate the inhibitory effect; If: the result of the judgment is that the inhibitory effect is not achieved, then execute step 1.3; Or if the result is that the inhibitory effect is achieved, execute step 2.3: Step 2.3: Image frequency distortion compensation; Step 2.4: Compensate for uneven distortion within the band; Step 2.5: Eliminate local oscillator leakage; Step 2.6: End.

2. The broadband image suppression method based on the combination of LS algorithm and MP model according to claim 1, characterized in that: In step 1.2, data alignment is performed, that is, data alignment is performed on the original input signal and the received signal, including integer alignment and fractional alignment processes, to eliminate errors caused by sampling at the receiving end.

3. The broadband image suppression method based on the combination of LS algorithm and MP model according to claim 1, characterized in that: In step 1.4, the compensation parameters of the in-band uneven distortion, including time domain filtering, power and phase compensation and LS algorithm solution, correctly configure the filter cutoff frequency and order to ensure the accuracy of the training results.

4. The broadband image suppression method based on the combination of LS algorithm and MP model according to claim 1, characterized in that: In step 1.6, the compensation parameters of the image frequency distortion, including time domain filtering, power and phase compensation and MP model algorithm solution, correctly configure the filter cutoff frequency, order and the order of the MP model to ensure the accuracy of the training results.

Images