JPH01200852A - Maximum likelihood receiver - Google Patents

Abstract

PURPOSE:To execute a satisfactory code series reproduction even when a linear distortion due to delay components is generated by generating fundamental waves corresponding to the delay components contained in a receiving digital signal, respectively, and obtaining a correlation between the fundamental wave and the receiving digital signal. CONSTITUTION:When a modulating wave is supplied to an input terminal 1, a correlation device 2 for extracting components obtains the correlation between a receiving signal and a code series, extracts a specific code series, and outputs a complex delay profile. A fundamental waveform generator 3 generates the fundamental waveform to the delay time of the delay profile, and it is supplied to a fundamental waveform correlation device 4. Further, the modulating wave is supplied to the correlation device 4, the correlation between each fundamental wave and a code series part is obtained, and it is stored into a correlation memory 5. A distance arithmetic device 7 obtains a signal space distance based on the contents of the memory 5, a degenerating pass memory 6 and a condition transition memory 9. A condition deciding device 8 determines a condition, the condition transition memory 9 considers that the modulating wave with the smallest signal space distance is received, and the corresponding code series is outputted to a decoder 10.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention is used for receiving and demodulating encoded signals. In particular, the present invention relates to a maximum likelihood receiver that receives signals via a wireless digital transmission path with delayed waveform distortion.

[Conventional technology]

Generally, when performing signal transmission, the signal sent from the transmitter is subjected to linear distortion due to the band characteristics of the transmission device and transmission medium disposed on the transmission path. Therefore, in digital signal transmission, code amount interference occurs and receiver performance deteriorates. Various equalizers have been proposed in the past to reduce the deterioration of transmission characteristics caused by such linear distortion, and a maximum likelihood receiver is known as one with particularly high performance.

FIG. 6 is a block diagram of a first conventional maximum likelihood receiver.

The received signal is supplied to input terminal 1, and further supplied to maximum likelihood code sequence estimator 63 via matched filter 61 and sampling circuit 62. The output of the maximum likelihood code sequence estimator 63 is connected to the output terminal 11.

The operation of this maximum likelihood receiver will be described using a complex number representation of a baseband detection waveform when carrier modulation is used as the modulation method and orthogonal detection is performed before equalizer processing.

The complex envelope S (t) of the linear modulation signal for transmitting the code string (a,) at a speed of 1/T is the basic waveform g (t)
Using , it is expressed as . An impulse response h (t) corresponding to the linear distortion of the transmission path acts on this. Therefore, the received wave R(t) at the input terminal 1 is as follows. Here, "*" represents a compound operation, and N (t) represents noise.

The matched filter 61 is matched to the impulse signal of g (t) * h (t), and has the characteristic that the signal-to-noise ratio of the output waveform is the highest. The sampling circuit 62 samples the output waveform of the matched filter 2 at every time T in synchronization with the complex envelope S (t). The maximum likelihood code sequence estimator 63 internally generates matched filter ideal outputs corresponding to possible combinations of code sequences (a,), and samples the waveform in synchronization with the complex envelope S (t). The signal spatial distance between the input signal and the actually input signal is calculated, and a combination of code strings (al) corresponding to the waveform with the closest distance is output to the output terminal 11.

FIG. 7 is a block diagram of a second conventional maximum likelihood receiver.

When using a nonlinear modulated wave, it is not possible to form a linear matched filter, so the modulated wave is approximately represented by a combination of several carrier waves. For example, TFM
(Tamed FM) and Gaussian baseband filter normalized 3dB bandwidth B, T (one side) is 0.21
GMSK (Gaussian-filtered
In order to approximately generate a modulated wave of (Minimum Shift Keying), the signal waveform of one time slot is divided into five waves for each time slot: a carrier wave with frequency f0 and a carrier wave with frequency offsets of ±Δf and ±2Δf. Connect them so that they are phase continuous.

To explain FIG. 7 by way of example, a received wave input from input terminal 1 is supplied to multipliers 71-1 to 71-5. Multiplier 7
1-1 to 71-5 further include e,=cos
2π(fc+2Δf)t ez=cos2π(fc+Δf)te,=cos2πfct e,=CO52π(fc−Δf)tes=C092π(fc−2Δf)t are input. The outputs of the multipliers 71-1 to 71-5 are respectively output to the integrator 7.
2-1 to 72-5. These multipliers 71-
1 to 71-5 and integrators 72-1 to 72-5,
The correlation between the received wave and each of the five carrier waves is determined.

These multipliers 71-1 to 71-5 and integrator 72-
1 to 72-5 to form a matched filter corresponding to the waveform of one time slot, and its output is sent to a sampling circuit 62-5.
1 to 62-5. Maximum likelihood code sequence estimator 6
3 estimates a code sequence based on the values sampled by the sampling circuits 62-1 to 62-5. When the constraint length of the code sequence is long, the hardware scale of the maximum likelihood code sequence estimator 63 increases exponentially with respect to the constraint length. Therefore, the signal is usually estimated using the Viterbi algorithm.

The maximum likelihood code sequence estimator 63 outputs the estimated code sequence from the output terminal 11.

[Problem that the invention seeks to solve]

However, if linear distortion is added to the nonlinear modulated wave due to the band characteristics of the transmission path, the received wave cannot be expressed by a linear combination as shown in equation (2), making it difficult to obtain a matched filter. Met. Therefore, even if a conventional maximum likelihood receiver for nonlinear modulated signals is used as is, the error rate characteristics cannot be improved.

SUMMARY OF THE INVENTION An object of the present invention is to solve the above problems and provide a maximum likelihood receiver that can receive and demodulate nonlinearly modulated signals with linear distortion.

[Means for solving problems]

The maximum likelihood receiver of the present invention includes a means for detecting a component of a specific code sequence included in a received digital signal and determining a delay component included in the signal, and generating a fundamental wave corresponding to each of the delay components. It is characterized by including means for.

Preferably, the means for determining the delay component includes a component extraction correlator that measures the delay time and amplitude of the delay component by correlating the received digital signal with a pre-stored code sequence. Further, it is preferable that the means for generating the fundamental wave is configured to generate the fundamental wave with the time slot length of the delay component.

This maximum likelihood receiver further includes a correlation memory that stores the output of the correlation means for a predetermined time slot as a maximum likelihood code sequence estimation means, and a plurality of states up to the current state estimated on the receiving side. A state transition memory that stores transition history data, multiple degenerate paths that extract the relative relationship of state transition history, the relationship between the inner product value of the signal waveforms of each time slot in each path, and one A degenerate path memory that stores the relationship between the next degenerate path following the degenerate path, the storage contents of the correlation memory, the output of the means for determining the delay component, the storage contents of the state transition memory, and the degenerate path memory. A distance calculator calculates the signal spatial distance within a time slot to the next state following multiple current states, and a state determiner determines the next state using the Viterbi algorithm. It is preferable to include means for outputting a code of a code sequence corresponding to the state history of the state in which the cumulative signal spatial distance of the plurality of current states is the smallest.

[For production]

The maximum likelihood receiver of the present invention generates a fundamental wave corresponding to each delay component included in a received digital signal, and finds a correlation between this fundamental wave and the received digital signal. For this reason,
Even if linear distortion occurs due to delay components, code sequences can be reproduced favorably.

Furthermore, the maximum likelihood receiver of the present invention is configured to recursively determine the signal spatial distance for each time slot, and can simplify state determination using a degenerate path table.

〔Example〕

FIG. 1 is a block diagram of a maximum likelihood receiver according to an embodiment of the present invention.

Here, the modulated wave is a TFM signal or ab'r=Q.

This will be explained using the No. 21 GMSK signal as an example. These two types of signals are almost equivalent signals, and j2πfct S (t) = E (t) e −
−(3)jφ(1) E (t) = e −−・・
...-(4) is expressed. Here, 2πfc=ω. is the angular frequency of the carrier and E (t) is the complex amplitude.

FIG. 2 shows a trellis representing changes in φ(1). Here, a trellis for four time slots is shown. Each grid point represents a state, and the number in the figure indicates the state. The same number represents the same state, and there are 16 states. The time lock ■ and the time slot ■ form the same trellis as the time slot I and the time slot ■. Also, I(t)
The envelope represented by = cosφ (1) has the waveform shown in FIG.

Returning to the explanation of the embodiment shown in FIG. 1, a modulated wave expressed by equation (3) is supplied to the input terminal 1'. This modulated wave includes a signal modulated with a specific code sequence at regular intervals as a frame signal. The component extraction correlator 2 correlates the received signal with a pre-stored code sequence,
A specific code sequence is extracted from the received signal. At this time, if a plurality of delayed wave components are superimposed on the received signal, the output of the component extraction correlator 2 has a complex delay profile (α., α1) according to the delay time difference of each component and the amplitude of the delayed wave. , α2
, -) are output. Here, α1 represents a complex coefficient of the i-th delay component. This delay profile is supplied to the basic waveform generator 3. The basic waveform generator 3 generates a basic waveform for one time slot as shown in FIG. 3 for the delay time of the delay profile. This basic waveform is supplied to each corresponding correlator in the basic waveform correlator 4.

The modulated wave at input terminal 1 further includes a signal of a code sequence portion that transmits information, and this is supplied to basic waveform correlator 4 . The fundamental waveform correlator 4 correlates each fundamental wave from the fundamental waveform generator 3 with the code sequence part. This correlation output is stored in a corresponding area in the correlation memory 5. Each area is provided with a storage capacity corresponding to the signal constraint. When the modulated wave has the waveform shown in FIG. 2, the constraint length is four time slots, so four stages of memory are required.

The distance calculator 7 calculates the signal spatial distance based on the stored contents of the correlation memory 5, the degenerate path memory 6, and the state transition memory 9. The signal spatial distance is the complex envelope R of the received signal.
(t) and some expected signal waveforms Uh(t), and is expressed as mT. Here, the complex envelope R(t) is the direct wave E
(t), and E (t-T
), then R(t) = α. E(t)+α, E(t-T)+
N(t)・・・−・−・・・・−・・・(6) It becomes. However, N (t) represents noise. Also, Uk
(t) is expressed as U k (t) = αQEk (t) + αr Ek” (t
) ”-・ (7) Here, Ek(t), E
k'(t) is the fundamental waveform output by the fundamental waveform generator 3.

Therefore, the integrand F (t) is: F(t) = l R(t) l ” + lα.l'
l Eh(t)12+1αll'' l Ek'(t
)12+2Re[α. R(t) Eh”(t))+2
Re Cat R(t) Eh-”(t))+2Re
Ca. at Ek(t) Ek'(t), 1-...-
・---1---(8) becomes. The first term, second term, and third term are signal levels, and their integral values can be easily obtained. Sections 4 and 5
Since the term is the correlation between the received signal and the fundamental wave, it can be obtained from the stored contents of the correlation memory 5. However, since which of the memory contents to use depends on the state history, the state transition memory 9 that stores the state transition history
See. Since the signal spatial distance can be determined recursively using equation (5), the state transition memory 9 also stores the signal spatial distance up to (m-1)T. The sixth term represents the inner product value between the basic signal waveforms. For this value,
The pre-calculated values are stored in the degenerate path memory 6 as a degenerate path table (.

The degenerate path table is a grouping of relatively similar paths determined by state transitions. For example, the path that transitions to 13.16.1.2.9 shown in bold line in FIG. 2 is considered to be an example of a degenerate path as shown in FIG. In addition to the above-mentioned inner product values, the degenerate path memory 6 stores paths that can be taken in the next step as a table. This table allows the next state to be determined. In this way, for state k, when transitioning to the two next states (
m-1) Signal spatial distance from T to mT d k l S
d k 2 is found.

The state determiner 8 determines the state using the Viterbi algorithm since the transition destination state is transitioned from two states. When a new state is determined by the state determiner 8, the state transition memory 9 considers that a modulated wave corresponding to the state with the smallest signal spatial distance Dk(n) has been received among the states, and stores the code corresponding to the modulated wave. The sequence is output to the decoder 10.

FIG. 5 shows the error rate characteristics of the receiver of this embodiment and the conventional receiver. This characteristic diagram shows the error rate characteristics when the direct wave component is "1" and the delayed wave component with a delay time of 3T is eJ. As shown in this figure, the characteristics of the receiver of this embodiment are greatly improved, and even when a nonlinear modulated wave is received as a discrete delayed wave, it can be received and demodulated with good characteristics.

〔Effect of the invention〕

As explained above, the maximum likelihood receiver of the present invention can satisfactorily receive a signal even when linear distortion occurs in a nonlinear modulated wave due to the band characteristics of a transmission path.

Moreover, since the maximum likelihood receiver of the present invention can simplify calculations by using a degenerate path table, it has the advantage of being able to perform high-speed processing. Furthermore, calculation processing is simplified and power consumption can be reduced.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a maximum likelihood receiver according to an embodiment of the present invention. FIG. 2 is a diagram showing a trellis representing changes in φ(1). FIG. 3 is a diagram showing a trellis of an envelope represented by I (t)=cosφ(1). FIG. 4 is a diagram showing an example of a degenerate path. FIG. 5 is a diagram showing the error rate characteristics of the receiver of this embodiment and the conventional receiver. FIG. 6 is a block diagram of a first conventional maximum likelihood receiver. FIG. 7 is a block diagram of a second conventional maximum likelihood receiver. 1...Input terminal, 2...Correlator for component extraction, 3...
- Basic waveform generator, 4... Basic waveform correlator, 5...
Correlation memory, 6... Degenerate path memory, 7... Distance calculator, 8... State determiner, 9... State transition memory, 10... Decoder, 11... Output terminal, 61...
Matched filter, 62.62-1 to 62-5... Sampling circuit, 63... Maximum likelihood code sequence estimator, 71-1 to 71
-5... Multiplier, 72-1 to 72-5... Integrator. Patent applicant: Nippon Telegraph and Telephone Corporation, agent: Naotaka Ide, patent attorney. Time slow 2 to fan 2 RoCNR (d81 tail 5 mouth

Claims (1)

[Claims] 1. Fundamental wave generation means for generating a fundamental wave corresponding to a modulated wave of a received digital signal; Correlation means for determining the correlation between the fundamental wave generated by this means and the received digital signal; and maximum likelihood code sequence estimating means for calculating a maximum likelihood code sequence for the output of the correlation means, wherein the fundamental wave generating means detects a component of a specific code sequence included in the received digital signal. A means of determining the delay component included in the signal (
2); and means (3) for generating a fundamental wave corresponding to each of the delayed components. 2. The maximum likelihood code sequence estimating means includes a correlation memory (5) that stores the output of the correlation means over a predetermined time slot, and state transition history data up to a plurality of current states estimated on the receiving side. A state transition memory (9) that stores the relative relationships of state transition histories, multiple degenerate paths extracted from which the relative relationships of state transition histories are extracted, the relationship between the inner product values of the signal waveforms of each time slot in each path, and one a degenerate path memory (6) that stores the relationship of the next degenerate path following the degenerate path, the stored contents of the correlation memory, the output of the means for determining the delay component, and the stored contents of the state transition memory. , a distance calculator (7) that calculates the signal spatial distance within a time slot to the next state following multiple current states based on the stored contents of the degenerate path memory, and a distance calculator (7) that calculates the next state using the Viterbi algorithm. 2. A state determining device according to claim 1, comprising: a state determiner for making a determination; and means (9, 10) for outputting a code of a code sequence corresponding to a state history of a state in which the accumulated signal spatial distance of the plurality of determined current states is the smallest. Maximum likelihood receiver.