CN107634926B - High-order modulation soft decision method combining channel information - Google Patents

Abstract

The invention relates to a high-order modulation soft decision method of combined channel information. Specifically, in the implementation process of the OFDM system, equalization is not performed separately, but channel equalization is performed in combination with channel information. The method is simpler to realize, and the resources of the divider are saved; secondly, the method is not influenced by the equalization process, and particularly under the condition of low signal-to-noise ratio or deep fading channel, the influence of noise cannot be amplified, so that the sensitivity of the receiver can be improved.

Description

High-order modulation soft decision method combining channel information

Technical Field

The invention belongs to the technical field of mobile communication systems, and particularly relates to a soft decision device and method for combining channel information.

Background

In the transmission scheme of the broadband wireless communication system, an ofdm (orthogonal frequency division multiplexing) technology is a typical transmission system, and has the characteristics of relatively simple implementation, high spectrum utilization rate and the like, and is widely applied to various communication standards. Generally, in the OFDM transmission technology, when the length of cp (cyclic prefix) is not less than the maximum delay of the channel, and the starting point of the fft (fast Fourier transform) window of the OFDM symbol is determined within its guard interval, the receiver samples the received signal sample

FFT conversion is carried out to obtain

Wherein, X_l[k]、Y_l[k]、H_l[k]And Z_l[k]Respectively representing the transmitted symbol, the received symbol, the frequency response of the channel and the frequency domain noise on the kth subcarrier of the ith symbol. Thus, an OFDM system can be considered in the frequency domain as a frequency domain system as shown in fig. 1, when the received signal is equal to the product of the input symbols and the frequency response of the channel. In summary, in the absence of noise, the received signal only needs to be divided by the channel, i.e. the transmitted signal can be detected by a single-tap equalizer.

In a mobile communication system using an OFDM transmission scheme, in order to obtain higher spectrum utilization and coding gain, a data traffic channel generally uses a coded modulation system combining advanced channel coding techniques (e.g., Turbo code and LDPC code) and high-order modulation techniques (16QAM, 64QAM, etc.). In such systems, a combination of decoding and demodulation is often used, and the demodulator does not determine the codeword at the input of the decoder, but only performs a temporary estimation of the various symbols, which is called a "soft" decision. It may not lose some information useful to the decoder. Soft-decision decoding generally has some degree of improvement in performance over hard-decision decoding. Generally, with soft decisions, the SNR of the signal will have a 2dB advantage over hard decisions.

In a typical OFDM system implementation process, a pilot sequence is generally used to obtain an estimated value of a channel, on the basis, data information is equalized according to the obtained channel information, and finally, soft decision is performed according to symbol information obtained by equalization processing. In fact, the implementation of equalization before soft decision leads to performance loss of the OFDM system and increases the computational complexity of implementation, and to solve this problem, we propose a method that does not perform equalization alone, but directly performs soft decision on channel information obtained by joint estimation.

Disclosure of Invention

From the frequency domain, any subcarrier of the OFDM system can be equivalent to a single-input single-output system, and the transceiving relation can be expressed as

Y＝HX+Z (2)

Where Y is the frequency domain received signal, X is the frequency domain transmitted signal, H is the frequency domain response of the channel, typically the complex flat fading channel gain, and Z is the frequency domain noise. In general, the noise considered is white Gaussian noise with a mean of zero and a variance of

The conventional method generally performs equalization processing on the received data signal to obtain symbol information of the transmitted signal, and is described by taking ls (least square) equalization as an example

At this time, the process of the present invention,

mean value is zero and variance is

Gaussian random variable of (2). Thus, with the known transmitted symbol X, the symbol is received

The condition PDF (probability function) of (1) is

Order to

And

representing symbol sets with 1 and 0 in the ith bit, respectively, the log-likelihood ratio (LLR) or soft output value of the ith bit is defined as

If all symbols etc. are possible (i.e. p (X) is a constant), then equation (4) can be approximated as

Wherein the content of the first and second substances,

and

is defined as

If in a coding block

Is constant, and equation (5) can be further simplified to

Different from the traditional soft decision method, the soft decision realization method combining the channel information is provided, and a separate equalization processing process is not needed. From equation (2), it can be found that the conditional PDF of the received symbol Y is given by the transmitted symbol X

According to equation (7), the equalization process is not considered here, but the LLR or soft output value of the l-th bit is calculated directly from the received data signal and the channel information estimated from the pilot sequence, and is defined as

Similarly, if all symbols etc. are likely to occur (i.e. p (X) is a constant), then equation (8) can be approximated as

Further assume that in one coding block

Is constant, equation (9) can be further simplified to

The invention provides a soft decision method of combined channel information, which is used for a digital communication system and designs a corresponding realization method in an OFDM broadband communication system. Compared with the traditional soft decision method, the method has the following two main advantages: firstly, the equalization processing is not needed to be carried out independently in the implementation process, and a divider is saved in implementation, so that the calculation complexity required by implementation is reduced; secondly, the method is not influenced by the equalization process, and particularly under the condition of low signal-to-noise ratio or deep fading channel, the influence of noise cannot be amplified, so that the sensitivity of the receiver can be improved.

Drawings

FIG. 1 is a frequency domain equivalent model of the OFDM system;

FIG. 2 is a block diagram of a joint channel information soft decision implementation in an OFDM system;

FIG. 3 is a block diagram of a conventional soft decision implementation in an OFDM system;

FIG. 4 shows the physical significance of the thresholds in the LLR quantization process;

fig. 5 is a plot of bit error performance for the conventional soft decision method and the inventive joint channel information soft decision method.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the scope of the present invention.

The embodiment of the invention discloses an implementation case of a soft decision method for combining channel information in an OFDM system, and a specific implementation block diagram is shown in FIG. 2. In an OFDM system, a signal synchronized at a receiving end is subjected to serial/parallel conversion and then is subjected to FFT module to obtain a frequency domain signal, on the basis, a Preamble signal is used for obtaining a channel information estimation value, then, parallel/serial conversion is carried out on a data signal and channel information, soft decision is carried out according to channel information (CSI) obtained by estimation and received data information to obtain bit level soft information, and finally, the soft information is input into a channel decoding module to obtain final decoding output.

The block diagram of the conventional soft decision implementation in the OFDM system is shown in fig. 3, and it can be seen that compared with the present invention, its data is recovered to the transmission signal through channel equalization before being subjected to soft decision, and then the soft decision is performed. Taking the OFDM system using the block pilot as an example to illustrate the channel information processing part in the two soft decision manners shown in fig. 2 and fig. 3, in the combined channel information soft decision method shown in fig. 2, when the Least Square (LS) channel estimation and the simple linear interpolation method are used, only the channel estimation needs to be performed according to the equation (11)

Wherein the content of the first and second substances,

representing the frequency domain response estimate of the channel, C₁And C₂Respectively representing coefficients of linear interpolation, X_PA frequency domain transmission signal representing a pilot frequency,

and

frequency domain received signals respectively representing two adjacent pilots, ()^*Representing a conjugate operation. Note that in an actual OFDM system, a Chu sequence or the like having a constant modulus property is often used for a frequency domain transmission signal of a pilot, and thus | X in equation (11)_P|²And (3) constant representation, and division is not needed in the specific implementation process.

In the conventional soft-decision method shown in fig. 3, similarly, when the LS channel estimation and the simple linear interpolation method are adopted, the channel estimation and equalization are required according to equation (12)

Where Y represents the frequency domain received signal of the data, X_eqRepresenting a frequency domain estimate of the data signal. It can be found by the analysis of formula (12),

the divider in the equalizer shown in equation (12) is indispensable because it varies with time and transmission environment. In summary, the joint channel information soft decision method shown in fig. 2 saves a divider compared with the conventional soft decision method, and thus requires less computational resources for implementation.

The core of the soft information quantization unit is to select a specific threshold (embodied as truncation processing in the implementation process) for the LLR value, and particularly, for multilevel QAM modulation, different thresholds need to be selected for different bits. Selecting a particular threshold for an LLR value can reduce to some extent the bit width required to quantize the LLR value, and the physical significance of the threshold values is shown in fig. 4. The smaller the threshold value is, the soft decision is performed only for the case outside the circle in the quantization process in the graph, and the case inside the circle is correspondingly hard decision. Thus, the solid line circle indicates that the selected threshold is small, while the dashed line circle indicates that the selected threshold is large. In fact, when the threshold is selected to be small, the effect of soft decision is close to that of hard decision, therefore, the threshold value should take a large value, and when the threshold exceeds a certain value, it indicates that the current signal value is located near a certain constellation point, it belongs to the constellation point with probability 1. Therefore, it is reasonable to choose an appropriate threshold.

The soft decision method of the joint channel information is applied to an OFDM system and is tested in a practical environment, specific parameters of the OFDM system for testing are shown in a table I, a bit error rate curve of the joint channel information soft decision method adopting the traditional soft decision method and the invention is shown in a figure 5, and it can be obviously seen that compared with the traditional soft decision method, the bit error performance of the system can be obviously improved.

Table I simulation of OFDM system parameters

Claims (1)

1. A high-order modulation soft decision method of combined channel information, in the process of realizing OFDM system, not only do equalization alone, but also combine channel information to carry out channel equalization, its characteristic is, specifically: in an OFDM system, a signal synchronized at a receiving end is subjected to serial/parallel conversion and then is subjected to FFT module to obtain a frequency domain signal, on the basis, a Preamble signal is used for obtaining a channel information estimation value, then, parallel/serial conversion is carried out on a data signal and channel information, soft decision is carried out according to a modulation mode of the channel information and the data signal to obtain soft information, and finally, the soft information is input into a channel decoding module to obtain final decoding output;

when the least square channel estimation and linear interpolation method is adopted, only the channel estimation is carried out according to the formula (11)

In the formula (I), the compound is shown in the specification,

representing the frequency domain response estimate of the channel, C₁And C₂Respectively representing coefficients of linear interpolation, X_PA frequency domain transmission signal representing a pilot frequency,

and

frequency domain received signals respectively representing two adjacent pilots, ()^*Represents a conjugate operation, | X_P|²Representing a constant.

Images