JP3526271B2 - Decoding device and decoding method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a decoding apparatus and a decoding method for performing error correction decoding after normalizing data.

[0002]

2. Description of the Related Art In the third generation mobile communication system,
The error correction decoding process is performed after performing the respective processes in the order of (1) RAKE combining process, (2) physical channel decoding process, and (3) transport channel decoding process. In addition,
The RAKE combining process is to separate the preceding wave and the delayed wave from the received signal on which the preceding wave and the delayed wave, which have different delay times and have independent fading fluctuations and are superimposed, are separated in the multipath transmission line, and the delay times are aligned and weighted. Is performed to perform maximum ratio combination, and a diversity effect is obtained. Further, the physical channel decoding process is a process of performing intraframe deinterleaving or the like on a signal for each physical channel data sequence subjected to the RAKE combining process to divide it into transport channels. Further, the transport channel decoding process is a process of performing deinterleaving, rate matching, and error correction decoding over a plurality of frames on the signals of the transport channel data series that are individually divided.

When performing Viterbi decoding or turbo decoding in error correction decoding, if the indices of the received data series are not the same, the estimated probability density function and the actual signal distribution differ, and the likelihood between bits can be obtained correctly. Since it does not exist, the error correction capability is reduced.

Since the processing units of the above (1) to (3) are different from each other, it is necessary to use the same processing unit (index) in order not to reduce the error correction decoding capability. The process of aligning a plurality of data series represented by different indexes to a common index is called normalization.

That is, since the processing unit of RAKE combining processing is a slot and the processing unit of physical channel decoding processing is a frame, after performing RAKE combining processing,
Before performing the physical channel decoding process, the index of the received data sequence in each slot must be normalized within the frame. Similarly, the processing unit of the physical channel decoding process is a frame, and the processing unit of the transport channel decoding process is TTI (Transmit Time Interval).
Therefore, the index of the received data sequence in each frame is set to TT before the transport channel decoding process.
It has to be normalized between the frames in I. However, since the TTI has a value of 1, 2, 4 or 8 frames, it is not necessary to perform interframe normalization when the TTI is 1 frame.

Intraframe normalization should be performed before intraframe deinterleaving. This is because after the order of the slot data sequence is changed, it is necessary to perform inefficient normalization processing such as storing the shift number for each data. Similarly, interframe normalization should be done before interframe deinterleaving.

In the third generation mobile communication system, generally, the error correction decoding processing is performed by hardware or software having a finite bit length, and the normalization processing is performed when the bit length is limited. This is a mandatory process. The bit length is limited by extracting a predetermined bit position part of the received data sequence.

The normalization process and the bit length limitation after the conventional RAKE combining process and before the physical channel decoding process will be described below.

Slot number is m, data number is n, RA
If the data sequence after KE synthesis processing is X (m, n) and the index of each slot is 2 ^ (-(s (m) -S)) (^: operator representing a power), the normalized frame Data series Xf
(m, n) is represented by the following equation (1).

Xf (m, n) = X (m, n) × 2 ^ (-(s (m) -S)) (1) Here, the index S is S = min [s (m)]. S (m) is an integer of 0 or more, and min [s (m)] is a function for detecting the minimum value in the set [s (m)] of s (m). The second term on the right side of the equation (1) is 2
The exponent of 2 is used because the exponent of 2 can be replaced by a shift operation that can be realized with a lower processing amount than a normal multiplication process when processed by hardware or software (hereinafter, The description will be made by replacing "index" with "shift number"). Further, the reason why the minimum value is detected as the shift number S is that the exponent of 2 is expressed as a negative value in the equation (1). That is, the minimum shift number S corresponds to the maximum amplitude.

FIG. 9 is a diagram for explaining the limitation on the bit length. Note that FIG. 9 shows the case where the data represented by 16 bits is limited in bit length to 8 bits.

When the maximum amplitude is large, the extraction position is shifted to the upper side (the side where the value is large) with reference to the decimal point position.
For example, in the case of FIG. 9, the shift number S is 3, the sign bit (B in FIG. 9), the integer part 5 digits, and the decimal part 2 digits (+31.75).
.About.-32.0) are extracted. On the contrary, when the maximum amplitude is small, the extraction position is shifted downward based on the decimal point position. For example, in the case of FIG. 9, the shift number S is 5, and the sign bit, the integer part 3 digits, and the decimal part 4 digits (+ 7.9325-−
8.0) is being extracted.

As described above, the conventional decoding method limits the bit length by normalizing based on the maximum amplitude. In addition, the intra-frame normalization and the inter-frame normalization are processed separately to reduce the capacity of the memory that stores the data series.

FIG. 10 shows an amplitude distribution when the output data after the RAKE combining processing is normalized and the bit length is limited to 8 bits and the error correction capability is high. Figure 1
At 0, the horizontal axis is the amplitude value and the vertical axis is the distribution amount. As shown in FIG. 10, in the case where the bit length is normalized and the bit length is limited to 8 bits, and the error correction capability is high, the amplitude distribution has the maximum distribution amount at the amplitude values “−127” and “+127”. , And the distribution amount is almost zero at the amplitude value “0”.

FIG. 11 is an example of the time amplitude response in the output data after the RAKE combining process, and shows the case where the amplitude value at a certain timing t ₁ is extremely larger than the amplitude values at other timings. Then, FIG. 12 shows R of FIG.
7 shows an amplitude distribution before normalization in output data after AKE combining processing.

[0016]

In the conventional decoding method,
Since the normalization is performed based on the maximum amplitude and the bit length is limited, in the case of FIG. 11, the slot data related to the timing other than t ₁ is almost underflowed (it becomes 0 as a soft decision value). . Therefore, if this is normalized, the distribution amount at the amplitude value “0” becomes maximum as in the amplitude distribution shown in FIG. 13, and the error correction capability deteriorates.

Such a state occurs under a fading environment, and frequently occurs as the moving speed becomes faster and the Doppler frequency becomes higher. Therefore, the BER (Bit Error Rate) characteristic in the fading environment is significantly deteriorated,
The call quality deteriorates, and in the worst case, a situation occurs where the call is suddenly disconnected.

The present invention has been made in view of the above circumstances, and provides a decoding apparatus and a decoding method which are not underflowed even if a normalization process is performed in a high-speed fading environment and have a high error correction capability. With the goal.

[0019]

Means for Solving the Problems] decoding apparatus of the present invention, 2
Normalize the number of shifts of the data series expressed in powers of
As an indicator, the number of shifts in the data series is averaged
Reference shift number calculating means for calculating the shift number, and the reference
Align the index by shifting the data series based on the number of shifts
Therefore, a configuration is provided that includes a normalization unit that normalizes the data sequence and an error correction decoding unit that performs the error correction decoding process on the normalized data series.

[0020]

With this configuration, the amplitude width of the data series to be normalized can be adaptively accommodated within the effective bit width, so that an extreme underflow does not occur in a high-speed fading environment and the likelihood of the data series is reduced. The degree of error correction can be effectively maintained and the error correction capability can be improved.

The decoding apparatus of the present invention, normalization means, by a value obtained by subtracting the reference number of shifts from the shift number of the data sequence data
The data sequence is right-shifted and normalized.

In the decoding apparatus of the present invention, the normalizing means subtracts 2 from the first value obtained by subtracting the reference shift number from the shift number of the data sequence.
The data series is right-shifted by the second value divided by and normalized.

In the decoding device of the present invention, the normalizing means has a first value of 0 obtained by subtracting the reference shift number from the shift number of the data sequence as 0.
De only the second value obtained by dividing the first value by 2 is equal to or greater than
The data series is right-shifted and normalized, and when the first value is smaller than 0, the data series is right-shifted by the first value and normalized.

In the decoding device of the present invention, the normalizing means sets the first value obtained by subtracting the reference shift number from the shift number of the data sequence to 0.
If it is above, normalize the data series without shifting,
When the first value is smaller than 0, the data series is right-shifted by a second value obtained by dividing the first value by 2 and normalized.

In the decoding apparatus of the present invention, the normalizing means sets the first value to 2 when the first value obtained by subtracting the reference shift number from the shift number of the data sequence is 1/2 or more of the finite bit width. The data series is right-shifted and normalized by the divided second value,
When the first value is smaller than ½ of the finite bit width, the data series is right-shifted by the first value and normalized.

In the decoding device of the present invention, the normalizing means shifts the data sequence when the first value obtained by subtracting the reference shift number from the shift number of the data sequence is 1/2 or more of the finite bit width. > Normalize without adding, and the first value is 1/2 of finite bit width
De only the second value obtained by dividing the first value by 2 if less than
The data sequence is right-shifted and normalized.

With these configurations, it is possible to avoid underflow and overflow by devising normalization processing.

The decoding device of the present invention uses the shift number of the data sequence represented by a power of 2 as an index for normalization, and R
First reference shift number calculating means for averaging the shift numbers of the data sequence in slot units after AKE combining to calculate a first reference shift number, and shifting the data sequence in slot units based on the first reference shift number. By aligning the indices
A first normalization unit for normalizing Ri, and the second reference shift count calculation means for calculating a second number of reference shift by averaging the number of shift data series of the first normalization unit frame output from, By shifting the data sequence in frame units based on the second reference shift number to align the indices
A second normalization unit for normalizing a configuration comprising an error correction decoding means for performing error correction decoding processing on the data sequence output from the second normalization unit taken.

With this configuration, the amplitude width of the data series to be normalized can be adaptively accommodated within the effective bit width, so that an extreme underflow does not occur in a fast fading environment, and the likelihood of the data series is reduced. The degree of error correction can be effectively maintained and the error correction capability can be improved.

In the decoding device of the present invention, the first normalizing means is
A configuration is adopted in which the data sequence is right-shifted by a value obtained by subtracting the first reference shift number from the shift number of the data sequence in slot units, and is normalized.

In the decoding device of the present invention, the second normalizing means is
A configuration is adopted in which the data series is right-shifted by a value obtained by subtracting the second reference shift number from the number of shifts of the data series in units of frames and normalized.

With these configurations, it is possible to avoid underflow and overflow by devising normalization processing.

The communication terminal device of the present invention has a configuration including any one of the above decoding devices. Further, the base station apparatus of the present invention has a configuration including any one of the above decoding apparatuses.

With these configurations, the error correction capability can be enhanced, so that high quality wireless communication can be performed.

[0036]

The program of the present invention uses the shift number of the data series represented by a power of 2 as an index for normalization,
A standard shift number calculation function that averages the shift numbers of the data series to calculate the standard shift number, and a positive value by aligning the indexes by shifting the data series based on the standard shift number.
A normalization function for normalizing and an error correction decoding function for performing error correction decoding processing on a normalized data series are realized.

The program of the present invention uses the shift number of the data series represented by a power of 2 as an index related to normalization,
A first reference shift count calculation function for calculating a first reference number of shifting the number of shift data series slot unit after RAKE combining by averaging shifts the data series of slot units based on the number of the first reference shift To align the indices
A first normalization function for more normalizing, a second reference shift number calculating function for averaging the shift numbers of the normalized frame-based data series to calculate a second reference shift number, and the second reference shift number a data series of frames on the basis of
A second normalization function for normalizing by shifting and aligning the exponents; an error correction decoding function for performing error correction decoding processing on the normalized data series over a plurality of frames;
To realize.

With these programs, the amplitude width of the data series to be normalized can be adaptively accommodated within the effective bit width, so that extreme underflow does not occur in a high-speed fading environment and the data series of the data series is not generated. The likelihood can be effectively maintained and the error correction capability can be improved.

[0040]

In the decoding method of the present invention, the shift number of the data sequence expressed by a power of 2 is used as an index for normalization, and the shift numbers of the data sequence are averaged to calculate the reference shift number. Then, the data series is shifted based on the reference shift number to align the indices
And a normalization step for performing the error correction decoding processing for performing the error correction decoding processing on the normalized data series.

In the decoding method of the present invention, the shift number of the data sequence represented by a power of 2 is used as an index for normalization, and R
A first reference shift number calculating step of averaging the shift numbers of the data sequence in slot units after AKE combining to calculate a first reference shift number; and shifting the data sequence in slot units based on the first reference shift number. By aligning the indices
A first normalizing step for normalizing the data, a second standard shift number calculating step for averaging the normal frame-unit shift numbers of the data series to calculate a second standard shift number,
Based on the standard shift number, the data series in frame units is
And a second normalization step of normalizing by aligning the exponents, and an error correction decoding step of performing an error correction decoding process on the normalized data series over a plurality of frames.

By these methods, the amplitude width of the data series to be normalized can be adaptively accommodated within the effective bit width, so that an extreme underflow does not occur in the fast fading environment and the data series of the data series is not generated. The likelihood can be effectively maintained and the error correction capability can be improved.

[0044]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present inventor has completed the present invention by paying attention to the fact that even if an overflow occurs in a portion where the amplitude value is extremely larger than the others, the error correction capability is not significantly affected. That is, the essence of the present invention is to adaptively fit the amplitude width of the data sequence to be normalized within the effective bit width by averaging the shift numbers and calculating the reference shift number.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

(Embodiment 1) FIG.1 is a block diagram showing the configuration of a decoding apparatus according to Embodiment 1 of the present invention.

The RAKE combining unit 101 inputs the received data after despreading and performs RAKE combining, limits the data after RAKE combining to a predetermined bit width, and limits the data memory 10.
2, and the shift number when limiting the bit length is written in the shift number memory 103.

The data memory 102 stores the RAKE-combined data in slot units. Shift number memory 10
3 stores the shift number of each slot related to the data after RAKE combining.

The reference shift number calculator 104 averages the shift numbers stored in the shift number memory 103 to calculate a first reference shift number. The reference shift number calculation unit 104
Details of the operation will be described later.

The normalizing section 105 includes a shift number memory 103.
The RAKE-combined data stored in the data memory 102 is normalized within the frame on the basis of the shift number and the first reference shift number stored in. The normalization unit 10
Details of operation 5 will be described later.

The deinterleaver 106 includes a normalization unit 10
Intra-frame deinterleaving, etc. is performed on the data normalized in 5 to divide the data into transport channels,
Write to the data memory 107.

The data memory 107 stores the data divided for each transport channel in frame units. The shift number memory 108 stores the first reference shift number as the shift number of the frame.

The reference shift number calculation unit 109 averages the shift numbers stored in the shift number memory 108 to calculate a second reference shift number. The normalization unit 110 normalizes the data for each transport channel stored in the data memory 107 within the TTI based on the shift number stored in the shift number memory 108 and the second reference shift number.

The error correction decoding unit 111 includes a normalization unit 110.
Inter-frame de-interleaving, rate matching, error correction decoding, etc. are performed on the normalized data.

Next, the operation of the reference shift number calculating section 104 will be described in detail.

Assuming that the number of slot shifts is s (m) [m = 0 to the number of physical channels × the number of slots-1], the reference shift number calculation unit 104 calculates the number of shifts in the following steps (1) to (6). The first standard shift number S ₁ is calculated by averaging.

(1) S ₁ = 0 (2) m = 0 (3) S ₁ = S ₁ + s (m) (4) m = m + 1 (5) m = physical channel number × slot number;
Repeat (3) and (4). (6) S ₁ = S ₁ / (number of physical channels x number of slots):
(Rounded up after the decimal point)

The reference shift number calculation unit 109 also averages the shift numbers by the same operation as the reference shift number calculation unit 104 to calculate the second reference shift number S ₂ . In this case, the number of physical channels × the number of slots is read as TTI / 10 of the transport channel.

Further, although the decimal point is rounded up in (6) above, it may be rounded down.

Next, the operation of the normalizing section 105 will be described in detail.

The normalizing unit 105 uses the input data series In Dat
a [m] [n], the shift number s [m] of each slot stored in the shift number memory 103, and the first reference shift number S ₁ are used to output data sequence Out Data [m] by the flow shown in FIG. [n]
To generate. In FIG. 2, N is the number of data in one slot, and M is the number of physical channels × the number of slots.

The normalizing section 105 operates according to the flow of FIG. 2 to shift the number of shifts s [m] of each slot to the first reference number of shifts S _{1 for} the input data series In Data [m] [n]. Is shifted right by the value obtained by subtracting
ut Data [m] [n] can be generated.

The normalizing unit 110 is also the normalizing unit 1 described above.
Output data series Out Data [m] [n] by the same operation as 05
To generate. In this case, the number of physical channels × the number of slots is read as TTI / 10 of the transport channel.

FIG. 3 shows the amplitude distribution when the shift numbers are averaged to calculate the reference shift number, and the distribution amount is maximum at the average amplitude value and almost zero at the amplitude value "0". Distribution.

By thus averaging the shift numbers and calculating the reference shift number, the amplitude width of the data series to be normalized can be adaptively accommodated within the effective bit width, so that in a high-speed fading environment. It is possible to effectively maintain the likelihood of the data series and to enhance the error correction capability without causing an extreme underflow. By averaging the shift numbers and calculating the reference shift number, an overflow occurs in a portion whose amplitude value is extremely larger than the others, but this does not significantly affect the error correction capability.

(Second Embodiment) In the second embodiment, the operation of the normalizing section 105 different from the first embodiment will be described in detail.

The normalization unit 105 uses the input data series In Dat
a [m] [n], the shift number s [m] of each slot stored in the shift number memory 103, and the first reference shift number S ₁ are used to output data sequence Out Data [m] according to the flow shown in FIG. [n]
To generate. In FIG. 4, N is the number of data in one slot, and M is the number of physical channels × the number of slots.

The normalizing section 105 operates according to the flow shown in FIG. 4, whereby the shift number s [m] of each slot to the first reference shift number S _{1 for} the input data series In Data [m] [n]. The output data series Out Data [m] [n] can be generated by performing a right shift by ½ of the value obtained by subtracting.

As a result, the probability of avoiding overflow can be increased as compared with the first embodiment.

The normalizing unit 110 is also the normalizing unit 1 described above.
Output data series Out Data [m] [n] by the same operation as 05
To generate. In this case, the number of physical channels × the number of slots is read as TTI / 10 of the transport channel.

(Third Embodiment) In the third embodiment, the operation of normalizing section 105, which is different from the first and second embodiments, will be described in detail.

The normalization unit 105 uses the input data series In Dat
a [m] [n], the shift number s [m] of each slot stored in the shift number memory 103, and the first reference shift number S ₁ are used to output data sequence Out Data [m] by the flow shown in FIG. [n]
To generate. In FIG. 5, N is the number of data in one slot, and M is the number of physical channels × the number of slots.

The normalizing unit 105 operates according to the flow of FIG. 5 to input when the value S obtained by subtracting the first reference shift number S ₁ from the shift number s [m] of each slot is 0 or more. The output data series Out Data [m] [n] can be generated by right shifting the data series In Data [m] [n] by S / 2. In addition, the normalization unit 105 subtracts the first reference shift number S ₁ from the shift number s [m] of each slot to obtain a value S.
When is negative, the input data series In Data [m] [n] is right-shifted by S and the output data series Out Data [m]
[n] can be generated.

As a result, the probability that underflow can be avoided can be greatly increased as compared with the first embodiment.

The normalizing unit 110 is also the normalizing unit 1 described above.
Output data series Out Data [m] [n] by the same operation as 05
To generate. In this case, the number of physical channels × the number of slots is read as TTI / 10 of the transport channel.

(Embodiment 4) In Embodiment 4, the operation of normalization section 105 different from Embodiments 1 to 3 will be explained in detail.

The normalization unit 105 uses the input data series In Dat
a [m] [n], the shift number s [m] of each slot stored in the shift number memory 103, and the first reference shift number S _{1 according} to the flow shown in FIG. [n]
To generate. In FIG. 6, N is the number of data in one slot, and M is the number of physical channels × the number of slots.

The normalizing section 105 operates according to the flow of FIG. 6 to input the input data when the value S obtained by subtracting the first reference shift number S ₁ from the shift number s [m] of each slot is 0 or more. The output data sequence Out Data [m] [n] can be generated by setting the shift number of the sequence In Data [m] [n] to 0. In addition, the normalization unit 105 determines the shift number s of each slot.
When the value S obtained by subtracting the first reference shift number S ₁ from [m] is negative, the input data series In Data [m] [n] is right-shifted by S / 2 and the output data series Out Data [m] [n] can be generated.

As a result, the probability of avoiding underflow and overflow can be greatly increased as compared with the first embodiment.

The normalizing unit 110 is also the normalizing unit 1 described above.
Output data series Out Data [m] [n] by the same operation as 05
To generate. In this case, the number of physical channels × the number of slots is read as TTI / 10 of the transport channel.

(Fifth Embodiment) In the fifth embodiment, the operation of the normalizing section 105 different from the first to fourth embodiments will be described in detail.

The normalizing unit 105 uses the input data series In Dat
a [m] [n], the shift number s [m] of each slot stored in the shift number memory 103, the first reference shift number S ₁ and the effective bit width X are used to output the data sequence according to the flow shown in FIG. Generate Out Data [m] [n]. In FIG. 7, N
Is the number of data in one slot, and M is the number of physical channels × the number of slots.

The normalizing unit 105 operates according to the flow of FIG. 7 so that the value S obtained by subtracting the first reference shift number S ₁ from the shift number s [m] of each slot is ½ of the effective bit width X. Input data series In Data [m] [n]
Right shift by S / 2 for output data series Ou
t Data [m] [n] can be generated.

As a result, in addition to the effects of the first embodiment,
It also has the effect of avoiding underflow.

The normalizing unit 110 is also the normalizing unit 1 described above.
Output data series Out Data [m] [n] by the same operation as 05
To generate. In this case, the number of physical channels × the number of slots is read as TTI / 10 of the transport channel.

(Sixth Embodiment) In the sixth embodiment, the operation of normalization section 105 different from the first to fifth embodiments will be described in detail.

The normalization unit 105 uses the input data series In Dat
a [m] [n], the shift number s [m] of each slot stored in the shift number memory 103, the first reference shift number S ₁ and the effective bit width X are used to output data series by the flow shown in FIG. Generate Out Data [m] [n]. In FIG. 8, N
Is the number of data in one slot, and M is the number of physical channels × the number of slots.

The normalizing section 105 operates according to the flow of FIG. 8 so that the value S obtained by subtracting the first reference shift number S ₁ from the shift number s [m] of each slot is ½ of the effective bit width X. Input data series In Data [m] [n]
The output data sequence Out Data [m] [n] can be generated with the shift number of 0 as 0. Further, the normalization unit 105, when the value S obtained by subtracting the first reference shift number S ₁ from the shift number s [m] of each slot is smaller than 1/2 of the effective bit width X, the input data series In Data [m The output data sequence Out Data [m] [n] can be generated by right-shifting S [/ 2] to [] [n].

As a result, in addition to the effects of the first embodiment,
It also has the effect of avoiding underflow and overflow.

The normalizing unit 110 is also the normalizing unit 1 described above.
Output data series Out Data [m] [n] by the same operation as 05
To generate. In this case, the number of physical channels × the number of slots is read as TTI / 10 of the transport channel.

[0091]

As described above, according to the present invention,
By averaging the number of shifts and calculating the reference number of shifts, the amplitude width of the data sequence to be normalized can be adaptively accommodated within the effective bit width, which causes an extreme underflow in a high-speed fading environment. It is possible to effectively maintain the likelihood of the data series and improve the error correction capability. Furthermore, by devising the normalization process, underflow and overflow can be avoided.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a decoding device according to a first embodiment of the present invention.

FIG. 2 is a flowchart of normalization processing of the decoding device according to the above embodiment.

FIG. 3 is a diagram showing an amplitude distribution after normalization by the decoding device according to the above embodiment.

FIG. 4 is a flowchart of normalization processing of the decoding device according to the second embodiment of the present invention.

FIG. 5 is a flowchart of normalization processing of the decoding device according to the third embodiment of the present invention.

FIG. 6 is a flowchart of normalization processing of the decoding device according to the fourth embodiment of the present invention.

FIG. 7 is a flowchart of normalization processing of the decoding device according to the fifth embodiment of the present invention.

FIG. 8 is a flowchart of normalization processing of the decoding device according to the sixth embodiment of the present invention.

FIG. 9 is a diagram for explaining the limitation of bit length.

FIG. 10 is a diagram showing an amplitude distribution when the error correction capability is high.

FIG. 11 is a diagram showing an example of a time amplitude response in output data after RAKE combining processing.

12 is a diagram showing an amplitude distribution before normalization in the output data after the RAKE combining process of FIG.

FIG. 13 is a diagram showing an amplitude distribution after normalization by a conventional decoding device.

[Explanation of symbols]

101 RAKE synthesizer 102 data memory 103 shift number memory 104 Reference shift number calculation unit 105 Normalization part 106 Deinterleaver 107 data memory 108 shift memory 109 Reference shift number calculation unit 110 Normalization part 111 Error correction decoding unit

Front page continued (58) Fields surveyed (Int.Cl. ⁷ , DB name) H03M 3/00-11/00 G06F 11/10 330 H03M 13/45 H04B 1/707 H04L 1/00

Claims (16)

(57) [Claims] 1. A reference shift number calculating means for calculating a reference shift number by averaging the shift numbers of the data series, using the shift number of the data series represented by a power of 2 as an index for normalization, and the reference shift. Shift data series based on number
A decoding device comprising: a normalization means for performing normalization by aligning the exponents, and an error correction decoding means for performing an error correction decoding process on the normalized data series. 2. The decoding device according to claim 1 , wherein the normalizing means normalizes the data series by right-shifting by a value obtained by subtracting the reference shift number from the shift number of the data series . 3. The normalizing means uses only the second value obtained by dividing the first value obtained by subtracting the reference shift number from the shift number of the data series by 2.
Characterized by right-shifting and normalizing the data series
The decoding device according to claim 1 . 4. The normalizing means shifts the data series to the right by a second value obtained by dividing the first value by 2 when the first value obtained by subtracting the reference shift number from the shift number of the data series is 0 or more. The decoding device according to claim 1, wherein the data sequence is right-shifted by the first value to normalize when the first value is smaller than 0. 5. The normalizing means de- embeds when the first value obtained by subtracting the reference shift number from the shift number of the data series is 0 or more.
Normalized without shifting over data sequence, only said second value where the first value is divided by 2 the first value less than zero data
Claims, characterized in that to normalize Right shifts the sequence
1. The decoding device according to 1 . 6. The normalizing means has a first value obtained by subtracting the reference shift number from the shift number of the data series is 1/2 of a finite bit width.
De only the second value obtained by dividing the first value by 2 is equal to or greater than
Data sequence is right-shifted and normalized, and if the first value is smaller than 1/2 of the finite bit width, only the first value is data.
Claims, characterized in that to normalize Right shifts the sequence
1. The decoding device according to 1 . 7. The normalizing means has a first value obtained by subtracting the reference shift number from the shift number of the data series is 1/2 of a finite bit width.
If it is above, normalize the data series without shifting,
Decoding according to claim 1, characterized in that said first value is normalized by right shifting the data sequence a second value obtained by dividing by 2 the first value when less than 1/2 of a finite bit width apparatus. 8. A first reference shift number is calculated by averaging the shift numbers of the data sequence in slot units after RAKE combining, using the shift number of the data sequence represented by a power of 2 as an index relating to normalization. 1 reference shift number calculation means, and a slot-based data sequence based on the first reference shift number
And a first normalizing means for normalizing by aligning the exponents , and a second standard shift number for averaging the shift numbers of the data series in frame units output from the first normalizing means. and second reference shift count calculating means, the data sequence for each frame based on the second reference shift number
A second normalizing means for normalizing by shifting and aligning the exponents, and an error correction decoding means for performing error correction decoding processing on the data sequence output from the second normalization means are provided. Decoding device. 9. The first normalizing means uses only a value obtained by subtracting the first reference shift number from the shift number of the data sequence in slot units.
Characterized by right-shifting and normalizing the data series
The decoding device according to claim 8 . 10. The second normalization means according to claim 8 or characterized by normalizing by right shifting the data sequence by a value obtained by subtracting the second number reference shift from the shift number of the data sequence for each frame The decoding device according to claim 9 . 11. A communication terminal device comprising the decoding device according to any one of claims 1 to 10 . 12. A base station device comprising the decoding device according to any one of claims 1 to 10 . 13. A reference shift number calculation function for averaging the shift numbers of a data sequence to calculate a reference shift number, using a shift number of a data sequence represented by a power of 2 as an index for normalization, and the reference shift. the data series based on the number
A program for realizing a normalization function for performing normalization by shifting and aligning exponents, and an error correction decoding function for performing error correction decoding processing on a normalized data series. 14. A first reference shift number is calculated by averaging the shift numbers of the data sequence in slot units after RAKE combining, using the shift number of the data sequence represented by a power of 2 as an index relating to normalization. First reference shift number calculation function, and first normalization by shifting a data series in slot units based on the first reference shift number and aligning indexes
A normalization function, a second reference shift number calculation function for averaging the shift numbers of the normalized data sequence in frame units to calculate a second reference shift number, and a second reference shift number calculation function in frame units based on the second reference shift number. Shift data series and align indices
Program for realizing a second normalization function for normalizing, and the error correction decoding function for performing error correction decoding processing on the data sequence across multiple frames normalized, the by obtaining. 15. A reference shift number calculating step of calculating a reference shift number by averaging the shift numbers of the data series, using the shift number of the data series represented by a power of 2 as an index for normalization, and the reference shift. the data series based on the number
A decoding method including a normalization step of performing normalization by shifting and aligning exponents, and an error correction decoding step of performing error correction decoding processing on the normalized data series. 16. A first reference shift number is calculated by averaging the shift numbers of the data sequence in slot units after RAKE combining, using the shift number of the data sequence represented by a power of 2 as an index for normalization. 1 standard shift number calculation step, and normalizing by shifting the data series in slot units based on the first standard shift number to align the exponents
A normalization step, a second reference shift number calculation step of averaging the shift numbers of the normalized data series in frame units to calculate a second reference shift number, and a second reference shift number calculation step in frame units based on the second reference shift number. Shift data series and align indices
Decoding method comprising a second normalization step of normalizing, an error correction decoding process for performing error correction decoding processing on the data sequence across multiple frames normalized, the by obtaining.