CA2290342A1 - Bit detection method in a radio communications system - Google Patents
Bit detection method in a radio communications system Download PDFInfo
- Publication number
- CA2290342A1 CA2290342A1 CA002290342A CA2290342A CA2290342A1 CA 2290342 A1 CA2290342 A1 CA 2290342A1 CA 002290342 A CA002290342 A CA 002290342A CA 2290342 A CA2290342 A CA 2290342A CA 2290342 A1 CA2290342 A1 CA 2290342A1
- Authority
- CA
- Canada
- Prior art keywords
- value
- bit
- bits
- equal
- values
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/08—Arrangements for detecting or preventing errors in the information received by repeating transmission, e.g. Verdan system
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
- H03M13/37—Decoding methods or techniques, not specific to the particular type of coding provided for in groups H03M13/03 - H03M13/35
- H03M13/43—Majority logic or threshold decoding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0045—Arrangements at the receiver end
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/06—Dc level restoring means; Bias distortion correction ; Decision circuits providing symbol by symbol detection
- H04L25/067—Dc level restoring means; Bias distortion correction ; Decision circuits providing symbol by symbol detection providing soft decisions, i.e. decisions together with an estimate of reliability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0056—Systems characterized by the type of code used
- H04L1/007—Unequal error protection
Abstract
The present invention relates generally to the detection of bits which are protected by repetition, and which, along with their repetitions, have associated soft values available which give a measurement of the reliability of their received values. In particular, the present invention may be used in speech encoding in the GSM mobile communications system, and more particularly to the detection of those class 2 bits called pulse5 bits which are not protected with channel coding. In enhanced full rate (EFS) transmission in GSM
there are 4 bits called pulse5 bits. These pulse5 bits are duplicated twice, giving three bits for each original bit, for a total of twelve pulse5 bits.
These bits have associated soft values, a probability measure of their reliability, that are produced by the equaliser. These associated soft values are then used in the present invention to improve detection of these pulse5 bits over the state of the art which detects the pulse5 bits only using a majority decision among the 3 bits generated, without using their associated soft values.
there are 4 bits called pulse5 bits. These pulse5 bits are duplicated twice, giving three bits for each original bit, for a total of twelve pulse5 bits.
These bits have associated soft values, a probability measure of their reliability, that are produced by the equaliser. These associated soft values are then used in the present invention to improve detection of these pulse5 bits over the state of the art which detects the pulse5 bits only using a majority decision among the 3 bits generated, without using their associated soft values.
Description
Bit Detection Method in a Radio Communications System FIELD OF THE INVENTION
The present invention relates generally to the detection of bits which are protected by repetition, and which, along with their repetitions, have soft values available which give a measurement of the reliability of their received values. In particular, the present invention may be used in the decoding process of speech parameters in the GSM mobile communications system, and more particularly to the detection of those class 2 bits called l0 pulses bits in enhanced full-rate speech (EFS) which are not protected with channel coding.
RELATED ART
In GSM, the speech is digitised and sliced into 20 ms pieces. A
20 ms speech frame is designed to contain 260 bits which is divided into three blocks of bits, class la, class 1b, and class 2, according to their level of protection. A speech frame sent every 20 ms gives a net bit rate of l3kbit/s. The normal mode of transmission in GSM is called full rate speech (FS).
Another option in GSM is enhanced full rate speech (EFS), which 2o is more efficient than ordinary full rate speech (FS). Although FS uses 260 bits to encode various speech parameters, EFS only needs 244 bits to encode the same information. Thus, in comparison to FS, EFS has 16 extra bits that can be used to protect some of the other bits. An 8 bit cyclic redundancy check (CRC) is used for error detection among some of the bits (i.e.
the class lb bits), while the other 8 of the 16 bits are used to protect some of the other bits known as pulses bits.
In EFS there are certain bits known as pulses bits generated by the speech encoder. These bits are not protected with channel coding, although they are protected by repeating them. In each SUBSTITUTE SHEET (RULE 26) speech frame there are four pulses bits that are protected by repeating each pulses bit two times, which gives a total of twelve (4x(1 pulses bit + 2 repetitions)) pulses bits in each speech frame.
The system must detect the values of each of the bits transmitted. Today, a pulses bit is detected by means of a majority decision. Since each pulses bit is repeated twice, a majority decision is made among the three resulting bits; if two of the three bits are Os then the decision will be taken that !U the value of the pulses bit as transmitted was equal to a 0. The GSM specification itself does not mention a particular detection method, but merely requires that each pulses bit must be repeated.
The prior art method still presents a problem, however, since it doesn't make any use of the "soft" values provided in the equaliser. These values provide a measure of the reliability of the bits received. A better method is proposed by the present invention that makes use of certain soft values provided by the equaliser.
In current GSM systems, a Viterbi equaliser is used which provides certain soft values about the received bits in addition to the bits themselves. The output of the equaliser gives both the value of the bit as received, and a measure of its reliability. The measure of the reliability is in the form of a probability that the bit is indeed equal to 0 or 1. By ignoring the soft values, present methods are not as reliable as that proposed by the present invention.
For example, in the present state of the art systems a majority decision is taken when two bits have the same value (e.g. two ls) in spite of the fact that their reliability might be low (i.e. low soft values). In addition, the reliability of the third bit, which can be of another value (e.g. 0), might be high. In this example, the state of the art would detect a value SUBSTITUTE SHEET (RULE 26) of 1, while the present invention would detect a value of 0, because a 0 was the most probable bit that was repeated ' according to the soft values.
These soft values are currently available in the decoder in present systems where they are normally used in the channel decoding process. The present invention proposes modifying the structure of current systems to then make use of these soft values to make a soft decision on the value of the pulses bits.
Simulations of this use of soft values instead of majority decision has shown an improved C/I and SNR performance of approximately 4.5 dB in the detection of these pulses bits.
SUMMARY OF THE INVENTION
As has been seen, there currently exists a problem with the present methods of using a simple majority decision to determine IS the value of the pulses bits in a GSM speech frame. The present invention aims to solve this problem by using those soft values provided by the equaliser. In current systems these soft values are used only for channel decoding and are not then available for use in detection of the pulses bits. Because they can 2o provide a measure of the reliability of the bits from the equaliser, they can increase the reliability of the decision taken that determines the value of the pulses bits.
Each speech frame in enhanced full rate (EFS) speech uses 244 of 260 available bits for encoding speech, leaving 16 bits 25 available for protection of other bits. Among the 260 bits are 4 bits called pulses bits. These 4 bits are each repeated twice, resulting in 12 total bits . For each of four total pulses bits > there are three soft values, the soft value of the original bit plus the soft values of two repetitions.
3o The equaliser produces a soft value for each of these three bits. In the first step of the method the bits whose values are equal to 0 are chosen. Then the soft values of these bits are added together to yield a value called SumSoftO, which will SUBSTITUTE SHEET (RULE 26) thereby be the sum of the soft values for all the bits equal to 0 for a given pulses bit.
The next step in the present invention is to similarly choose the bits whose values are equal to 1. Then the soft values of these bits are added together to yield a value called SumSoftl, which will thereby be the sum of the soft values for all the bits equal to 1 for a given pulses bit.
After the values for SumSoftO and SumSoftl are determined, a decision is taken according to the algorithm below:
to if SumSoftO>SumSoftl then pulses bit . 0 else if SumSoftO<SumSoftl then pulses bit . I
else t5 an ordinary majority decision takes place.
The present invention improves over the performance of present systems in the detection of pulses bits. A normal majority decision as performed by current systems can only correct one bit error. If two bit errors occur out of the three bits 2o allocated for each pulses bit, then an incorrect decision will be taken. By using the soft values for each of these bits, as a measure of their probability and reliability, performance is improved. Test simulations have indicated an improved C/I and SNR performance of approximately 4.5 dB in the detection of 25 these bits.
Although the preferred embodiment of the present invention focuses primarily on the detection of pulses bits in GSM
systems, the algorithm may be used in broader applications. The use of repetition to protect bits, as is done with pulses bits, 30 is not often used, due to its inefficiency. However, the SUBSTITUTE SHEET (RULE 26) additional availability of "soft" values for the bit and its repetitions makes possible the use of the present invention for ' detection of any bit which is protected by repetition and which has "soft" values available.
The present invention relates generally to the detection of bits which are protected by repetition, and which, along with their repetitions, have soft values available which give a measurement of the reliability of their received values. In particular, the present invention may be used in the decoding process of speech parameters in the GSM mobile communications system, and more particularly to the detection of those class 2 bits called l0 pulses bits in enhanced full-rate speech (EFS) which are not protected with channel coding.
RELATED ART
In GSM, the speech is digitised and sliced into 20 ms pieces. A
20 ms speech frame is designed to contain 260 bits which is divided into three blocks of bits, class la, class 1b, and class 2, according to their level of protection. A speech frame sent every 20 ms gives a net bit rate of l3kbit/s. The normal mode of transmission in GSM is called full rate speech (FS).
Another option in GSM is enhanced full rate speech (EFS), which 2o is more efficient than ordinary full rate speech (FS). Although FS uses 260 bits to encode various speech parameters, EFS only needs 244 bits to encode the same information. Thus, in comparison to FS, EFS has 16 extra bits that can be used to protect some of the other bits. An 8 bit cyclic redundancy check (CRC) is used for error detection among some of the bits (i.e.
the class lb bits), while the other 8 of the 16 bits are used to protect some of the other bits known as pulses bits.
In EFS there are certain bits known as pulses bits generated by the speech encoder. These bits are not protected with channel coding, although they are protected by repeating them. In each SUBSTITUTE SHEET (RULE 26) speech frame there are four pulses bits that are protected by repeating each pulses bit two times, which gives a total of twelve (4x(1 pulses bit + 2 repetitions)) pulses bits in each speech frame.
The system must detect the values of each of the bits transmitted. Today, a pulses bit is detected by means of a majority decision. Since each pulses bit is repeated twice, a majority decision is made among the three resulting bits; if two of the three bits are Os then the decision will be taken that !U the value of the pulses bit as transmitted was equal to a 0. The GSM specification itself does not mention a particular detection method, but merely requires that each pulses bit must be repeated.
The prior art method still presents a problem, however, since it doesn't make any use of the "soft" values provided in the equaliser. These values provide a measure of the reliability of the bits received. A better method is proposed by the present invention that makes use of certain soft values provided by the equaliser.
In current GSM systems, a Viterbi equaliser is used which provides certain soft values about the received bits in addition to the bits themselves. The output of the equaliser gives both the value of the bit as received, and a measure of its reliability. The measure of the reliability is in the form of a probability that the bit is indeed equal to 0 or 1. By ignoring the soft values, present methods are not as reliable as that proposed by the present invention.
For example, in the present state of the art systems a majority decision is taken when two bits have the same value (e.g. two ls) in spite of the fact that their reliability might be low (i.e. low soft values). In addition, the reliability of the third bit, which can be of another value (e.g. 0), might be high. In this example, the state of the art would detect a value SUBSTITUTE SHEET (RULE 26) of 1, while the present invention would detect a value of 0, because a 0 was the most probable bit that was repeated ' according to the soft values.
These soft values are currently available in the decoder in present systems where they are normally used in the channel decoding process. The present invention proposes modifying the structure of current systems to then make use of these soft values to make a soft decision on the value of the pulses bits.
Simulations of this use of soft values instead of majority decision has shown an improved C/I and SNR performance of approximately 4.5 dB in the detection of these pulses bits.
SUMMARY OF THE INVENTION
As has been seen, there currently exists a problem with the present methods of using a simple majority decision to determine IS the value of the pulses bits in a GSM speech frame. The present invention aims to solve this problem by using those soft values provided by the equaliser. In current systems these soft values are used only for channel decoding and are not then available for use in detection of the pulses bits. Because they can 2o provide a measure of the reliability of the bits from the equaliser, they can increase the reliability of the decision taken that determines the value of the pulses bits.
Each speech frame in enhanced full rate (EFS) speech uses 244 of 260 available bits for encoding speech, leaving 16 bits 25 available for protection of other bits. Among the 260 bits are 4 bits called pulses bits. These 4 bits are each repeated twice, resulting in 12 total bits . For each of four total pulses bits > there are three soft values, the soft value of the original bit plus the soft values of two repetitions.
3o The equaliser produces a soft value for each of these three bits. In the first step of the method the bits whose values are equal to 0 are chosen. Then the soft values of these bits are added together to yield a value called SumSoftO, which will SUBSTITUTE SHEET (RULE 26) thereby be the sum of the soft values for all the bits equal to 0 for a given pulses bit.
The next step in the present invention is to similarly choose the bits whose values are equal to 1. Then the soft values of these bits are added together to yield a value called SumSoftl, which will thereby be the sum of the soft values for all the bits equal to 1 for a given pulses bit.
After the values for SumSoftO and SumSoftl are determined, a decision is taken according to the algorithm below:
to if SumSoftO>SumSoftl then pulses bit . 0 else if SumSoftO<SumSoftl then pulses bit . I
else t5 an ordinary majority decision takes place.
The present invention improves over the performance of present systems in the detection of pulses bits. A normal majority decision as performed by current systems can only correct one bit error. If two bit errors occur out of the three bits 2o allocated for each pulses bit, then an incorrect decision will be taken. By using the soft values for each of these bits, as a measure of their probability and reliability, performance is improved. Test simulations have indicated an improved C/I and SNR performance of approximately 4.5 dB in the detection of 25 these bits.
Although the preferred embodiment of the present invention focuses primarily on the detection of pulses bits in GSM
systems, the algorithm may be used in broader applications. The use of repetition to protect bits, as is done with pulses bits, 30 is not often used, due to its inefficiency. However, the SUBSTITUTE SHEET (RULE 26) additional availability of "soft" values for the bit and its repetitions makes possible the use of the present invention for ' detection of any bit which is protected by repetition and which has "soft" values available.
The present invention will now be described in more detail with reference to preferred embodiments of the present invention, given only by way of example, and illustrated in the accompanying drawings, in which:
to FIG. 1 is a flowchart illustrating the method performed according to the present invention where comparisons are performed between variables SumSoftl and SumSoftO.
FIG. 2 is a flowchart illustrating the method performed according to the present invention where the comparison of t5 Figure 1 is performed according to a specific algorithm.
FIG. 3 is a flowchart illustrating the method performed according to the present invention where the total of a bit and its repetitions is odd.
FIG. 4 is a flowchart illustrating the method performed 20 according to the present invention where the total of a bit and its repetitions is even.
FIG. S is a diagram of part of a GSM mobile communications system used in reception of radio signals.
FIG. 6 is a diagram of a GSM enhanced full-rate (EFS) speech 25 frame, and how various bits are allocated among class 1a, class lb, and class 2.
FIG. 7 is a flowchart illustrating the preferred embodiment performed according to the present invention.
FIG. 8 is a graph illustrating the results of a simulation 30 illustrating the improved performance of the present invention SUBSTITUTE SHEET (RULE 26) WO 98/53574 PCTlSE98100915 s over the prior art in the detection of pulses bits in a GSM
system.
DETAILED DESCRIPTION
In Figure 1 is shown a flowchart of the present invention as used in detection of the value of a bit received in a communications system. In certain systems bits are protected by repetition of their values.
In addition, certain systems, such as the GSM system described in more detail below, also have "soft" values for all the t0 received bits. These soft values are a measure of the reliability of the bit received. If the bit is received as a 1, for example, the soft value gives a measure of the probability that the bit is actually a 1. These soft values can be used to increase the efficiency of the use of repeated bits. Normally systems that repeat bits, repeat them an even number of times.
This will then give a total odd number of bits from the original bit plus its repetitions . The system will then make a decision as to the value of the original bit transmitted by making a majority decision based on the value of the majority of the original bit and its repetitions.
In Figure 1 is shown a flowchart of the present invention that takes advantage of the availability of soft values to improve upon this majority decision method. The use of soft values will be described in more detail below in the description of the preferred embodiment detection of pulses bits in a GSM system.
In the first step 110 of the method as shown in Figure 1, the bits among those bits to be protected whose values are equal to 0 are chosen. These will be the bits from the original bit and its repetitions. Then the soft values of these bits are added together, step 120, to yield a value called SumSoftO, which will thereby be the sum of the soft values for all the bits equal to 0 among a given bit and its repetitions.
SUBSTITUTE SHEET (RULE 26) The next step in the present invention is to similarly choose, step 130, the bits from the original and its repetitions whose values are equal to 1. Then the soft values of these bits are added together, step 140, to yield a value called SumSoftl, which will thereby be the sum of the soft values for all the bits equal to 1 among a given bit and its repetitions.
After the values for SumSoftO and SumSoftl are determined, a comparison, step 150, between the two is taken and a decision, step 160, is then made as to the value of the original bit transmitted based on the result of this comparison.
The comparison can be made using a variety of algorithms, however the preferred method is illustrated in Figure 2. The values for SumSoftO and SumSofti are determined, steps 210-240, corresponding to steps 110-140 in Figure 1. Then, after the values for SumSoftO and SumSoftl are determined, a decision is taken according to the algorithm below:
if SuznSoftO>SumSoftl ( step 250 ) then bit . 0 (step 260) else if SumSoftO<SumSoftl(step 270) then bit . 1 (step 280).
A particular refinement of the invention can be seen in Figure 3. It is still possible after step 270 in Figure 2, corresponding to step 370 in Figure 3, that it is neither true that SumSoftO>SumSoftl or SumSoftO<SumSoftl. If an even number of repetitions of the original bit were made then an odd number of total bits is available from the original bit and its repetitions. In this situation a simple majority decision will be made, step 390, among the original bit and its repetitions.
For example, suppose the original bit is 0 and there are four repetitions, one equal to 0 and three equal to 1. However, it is neither true that SumSoftO>SumSoftl or SumSoftO<SumSoftl. Then a SUBSTITUTE SHEET (RULE 26) simple majority decision will be made among the five bits that the value of the original bit is a 1.
In Figure 4 is illustrated the situation where an odd number of repetitions of the original bit is performed. In this situation a simple majority decision cannot be performed because the total number of bits, original plus repetitions, is even. Instead, the value of the highest soft value is chosen, step 490. The system then checks, step 491, to see if a majority of 1s or Os has this highest soft value. The value of the original bit transmitted is lU then chosen, step 492, to be the value of the bit having this highest soft value. For example, if there are two Os and one 1 all having the same highest soft value, then the system chooses 0 as the value of the original bit.
If an equal number of is and Os have this same highest soft value then the step is repeated, step 493, for the next highest soft value. At some point a soft value will be reached which will have a majority of is or Os among the original bit and its repetitions whose soft values equal this value. The value of the original bit transmitted will then be set, step 492, to be this value .
In Figure 5 is a block diagram illustrating part of a radio receiver system as used in a GSM system. This will illustrate the preferred embodiment of the present invention. An antenna 510 of the receiver 520 of a base station, for instance, receives radio signals over a certain radio channel. The signals (e. g. data, speech) transmitted over this channel may become strongly distorted due to fading, for example, so that the transmitted bursts give rise to a distorted speech frame.
The receiver 520 converts a radio signal to a baseband signal.
The receiver 520 then sends this baseband signal to the equaliser 530 where the signal is then demodulated. The equaliser 530 also compensates for the distortion (e. g.
multipath propagation, time dispersion, etc.) caused by the SUBSTITUTE SHEET (RULE 26) The next step in the present inv radio channel. For example, in GSM a Viterbi equaliser 530 is typically used for this purpose.
"Soft" information is obtained from the equaliser 530 and is a measure of the reliability of each bit. There are several different ways of obtaining "soft" information that gives various measurements of this reliability. For example, the soft values can be obtained in the following way in a system using Trellis decoding. To each state in the Trellis (the Viterbi algorithm is used) two metrics are calculated, Mo and M1. To o each state there is also a soft value calculated defined as follows: SoftValue=~Mo-M1~. The used soft values correspond to the decided path in the Trellis.
The equaliser 530 sends the signal to the deinterleaver 540, which reorders the bitstream so that the bitstream is in the same order going into the channel decoder as it was when it originally left the channel encoder.
This signal is then sent to a channel decoder 550 to perform the opposite operation of that performed by the channel encoder on the transmitter side, i.e. to recover transmitted information 2o from the known redundant bits and the known channel coding (e. g.
a convolution code).
The decoded speech frames are delivered from the channel decoder 550 to the speech decoder 560, speech-frame by speech-frame, where a complete synthesis of received speech frames is effected so as to deliver speech signals to a converter 570 which converts speech signals from 13-bit uniform PCM to an 8-bit/A-law PCM used in the public-switched telephone network (PSTN).
In Figure 6 can be seen a speech frame 600 which contains 260 bits in accordance with the GSM recommendation. The speech frame is divided into three blocks of which each defines one of three different classes. One block of 50 bits is assigned to class 1a 610, one block of 132 bits is assigned to class Lb 620, and the remaining block of 78 bits is assigned to class 2 630. The 260 SUBSTITUTE SHEET (RULE 26~
bits are delivered from the speech encoder to form the digitised speech after speech encoding. Since a speech frame composed of 260 bits is sent every 20 ms, this gives an effective transmission rate of 13 kbit/s in full rate (FS) transmission.
5 In enhanced full rate (EFS) transmission the relevant speech parameters are coded using only 244 bits, giving a rate of slightly more than 12 kbit/s.
Class la 610: This block 610 of 50 bits are most sensitive to transmission error and cause the most problematic consequences to with regard to the intelligibility of the transmitted and decoded speech. When errors are found in these bits, large parts of the immediately preceding, correct speech frame are repeated.
This error detection is performed with the aid of three parity bits 640 which are added to the 50 data bits as control bits.
Class lb 620: This block of 132 bits is not protected by parity bits. Four bits are added as so-called tail bits 650.
These 132 bits are not equally as sensitive with regard to the intelligibility to transmission bit errors occurring as compared to the class la bits.
2o Both the class 1a and 1b bits, including the three parity bits and the four tail bits, are grouped together and convolution coded.
Class 2 630: These 78 bits are the bits least susceptible to error and are not protected at all by additional bits, as in the case of class la and lb.
These three blocks 610, 620, 630, in a speech frame contain 50+132+78=260 bits, not including the 3 parity and 4 tail bits, which, when added, give a complete total of 267 bits. Of these 267 bits, 53+136=189 bits (see second group 660) are convolution 3o encoded 265 with the rate=1/2, so a further 189 .bits are added.
This results in a total of 378+78=456 bits in the final group 670 coming from the channel encoder, which can then be interleaved for inclusion in the TDMA frames.
SUBSTITUTE SHEET (RULE 26) As mentioned above, in the enhanced full rate (EFS) mode of transmission in GSM there are used only 244 of the 260 bits available due to the greater efficiency of the speech encoding method. This means that there are an additional 16 bits that can be used to protect the other 244 bits. An 8-bit cyclic redundancy check (CRC) 680 is used for error detection among the class lb bits, while the other 8 bits 695 are used to protect a group of class 2 bits known as pulses bits 690 . These pulses bits 690 are not protected with channel coding.
!0 The remaining 8 of 16 bits available in EFS are used to protect the pulses bits 690 by repetition 695. There are four pulses bits 690 in each speech frame. These four bits 690 are each then duplicated 695, using the 8 free bits and forming a total of twelve pulses bits 690, 695.
t5 The pulses bits 690 must then be detected. The GSM specification doesn't require a particular detection method for the pulses bits, but merely requires that they be repeated. As has been mentioned above, current techniques detect the pulses bits by taking the value of each pulses bit and the values of its two 2o repetitions. A decision is then taken whereby the value of the pulses bit is decided to be the majority of the three values.
This is efficient, but not as reliable as possible using the soft values from the equaliser 530, Figure 5.
As discussed above, the equaliser 530 provides additional 25 information in the form of soft values for each bit, including the pulses bits. The present invention makes use of these soft values which provide a measure of the reliability of each bit.
In the preferred embodiment the method of the present invention will be implemented in the channel decoder 550, where the soft 30 values for all bits are available. These values are normally used in the channel decoding process.
For each of four total pulses bits there are three soft values, the soft value of the original bit plus the soft values of two SUBSTITUTE SHEET (RULE 26) WO 98!53574 PCTlSE98/00915 repetitions. The equaliser produces a soft value for each of these three bits. In the first step 710 of the method as shown in Figure 7, the bits whose values are equal to 0 are chosen.
Then the soft values of these bits are added together, step 720, to yield a value called SumSoftO, which will thereby be the sum of the soft values for all the bits equal to 0 for a given pulses bit.
The next step in the present invention is to similarly choose, step 730, the bits whose values are equal to 1. Then the soft values of these bits are added together, step 740, to yield a value called SumSofti, which will thereby be the sum of the soft values for all the bits equal to 1 for a given pulses bit.
After the values for SumSoftO and SumSoftl are determined, a decision is taken according to the algorithm below:
if SumSoftO>SumSoftl (step 750) then pulses bit . 0 (step 760) else if SumSoftO<Su~aSoftl (step 770) then pulses bit . 1 (step 780) else an ordinary majority decision takes place (step 790).
An example of this decision process can be illustrated. Assume that a pulses bit is equal to 0. It is repeated to yield three Os, but when detected two 1s and one 0 result. The soft values from the equaliser give low soft values for the is (i.e. low reliability) while it gives a high soft value for the 0 (i.e. a high reliability). The result is that the sum of the soft values for both is is still less than the soft value for the 0. Thus, the variable SumSoftO is greater than the variable SumSoftl (i.e. SumSoftO>SumSoftl), so that the decision will be taken -that the pulses bit is equal to 0 (i.e. pulses bit . 0).
SUBSTITUTE SHEET (RULE 26) ,, In this example the existing detection would detect a 1 but the present invention would detect a 0 because the reliability as derived from the soft values was greater than the reliability of a 1.
Figure 8 shows a result of a simulation comparing the present invention with the prior art method of using a majority decision. It can be seen that the present invention improves the residual bit error for pulses bits (rber~ulse5) performance between 3.4 and 4.6 dB, depending on the propagation conditions, IO and if it is interference or sensitivity that is tested.
The embodiments described above serve merely as illustration and not as limitation. It will be apparent to one of ordinary skill in the art that departures may be made from the embodiments described above without departing from the spirit and scope of t5 the invention. The invention should not be regarded as being limited to the examples described, but should be regarded instead as being equal in scope to the following claims.
SUBSTITUTE SHEET (RULE 26)
to FIG. 1 is a flowchart illustrating the method performed according to the present invention where comparisons are performed between variables SumSoftl and SumSoftO.
FIG. 2 is a flowchart illustrating the method performed according to the present invention where the comparison of t5 Figure 1 is performed according to a specific algorithm.
FIG. 3 is a flowchart illustrating the method performed according to the present invention where the total of a bit and its repetitions is odd.
FIG. 4 is a flowchart illustrating the method performed 20 according to the present invention where the total of a bit and its repetitions is even.
FIG. S is a diagram of part of a GSM mobile communications system used in reception of radio signals.
FIG. 6 is a diagram of a GSM enhanced full-rate (EFS) speech 25 frame, and how various bits are allocated among class 1a, class lb, and class 2.
FIG. 7 is a flowchart illustrating the preferred embodiment performed according to the present invention.
FIG. 8 is a graph illustrating the results of a simulation 30 illustrating the improved performance of the present invention SUBSTITUTE SHEET (RULE 26) WO 98/53574 PCTlSE98100915 s over the prior art in the detection of pulses bits in a GSM
system.
DETAILED DESCRIPTION
In Figure 1 is shown a flowchart of the present invention as used in detection of the value of a bit received in a communications system. In certain systems bits are protected by repetition of their values.
In addition, certain systems, such as the GSM system described in more detail below, also have "soft" values for all the t0 received bits. These soft values are a measure of the reliability of the bit received. If the bit is received as a 1, for example, the soft value gives a measure of the probability that the bit is actually a 1. These soft values can be used to increase the efficiency of the use of repeated bits. Normally systems that repeat bits, repeat them an even number of times.
This will then give a total odd number of bits from the original bit plus its repetitions . The system will then make a decision as to the value of the original bit transmitted by making a majority decision based on the value of the majority of the original bit and its repetitions.
In Figure 1 is shown a flowchart of the present invention that takes advantage of the availability of soft values to improve upon this majority decision method. The use of soft values will be described in more detail below in the description of the preferred embodiment detection of pulses bits in a GSM system.
In the first step 110 of the method as shown in Figure 1, the bits among those bits to be protected whose values are equal to 0 are chosen. These will be the bits from the original bit and its repetitions. Then the soft values of these bits are added together, step 120, to yield a value called SumSoftO, which will thereby be the sum of the soft values for all the bits equal to 0 among a given bit and its repetitions.
SUBSTITUTE SHEET (RULE 26) The next step in the present invention is to similarly choose, step 130, the bits from the original and its repetitions whose values are equal to 1. Then the soft values of these bits are added together, step 140, to yield a value called SumSoftl, which will thereby be the sum of the soft values for all the bits equal to 1 among a given bit and its repetitions.
After the values for SumSoftO and SumSoftl are determined, a comparison, step 150, between the two is taken and a decision, step 160, is then made as to the value of the original bit transmitted based on the result of this comparison.
The comparison can be made using a variety of algorithms, however the preferred method is illustrated in Figure 2. The values for SumSoftO and SumSofti are determined, steps 210-240, corresponding to steps 110-140 in Figure 1. Then, after the values for SumSoftO and SumSoftl are determined, a decision is taken according to the algorithm below:
if SuznSoftO>SumSoftl ( step 250 ) then bit . 0 (step 260) else if SumSoftO<SumSoftl(step 270) then bit . 1 (step 280).
A particular refinement of the invention can be seen in Figure 3. It is still possible after step 270 in Figure 2, corresponding to step 370 in Figure 3, that it is neither true that SumSoftO>SumSoftl or SumSoftO<SumSoftl. If an even number of repetitions of the original bit were made then an odd number of total bits is available from the original bit and its repetitions. In this situation a simple majority decision will be made, step 390, among the original bit and its repetitions.
For example, suppose the original bit is 0 and there are four repetitions, one equal to 0 and three equal to 1. However, it is neither true that SumSoftO>SumSoftl or SumSoftO<SumSoftl. Then a SUBSTITUTE SHEET (RULE 26) simple majority decision will be made among the five bits that the value of the original bit is a 1.
In Figure 4 is illustrated the situation where an odd number of repetitions of the original bit is performed. In this situation a simple majority decision cannot be performed because the total number of bits, original plus repetitions, is even. Instead, the value of the highest soft value is chosen, step 490. The system then checks, step 491, to see if a majority of 1s or Os has this highest soft value. The value of the original bit transmitted is lU then chosen, step 492, to be the value of the bit having this highest soft value. For example, if there are two Os and one 1 all having the same highest soft value, then the system chooses 0 as the value of the original bit.
If an equal number of is and Os have this same highest soft value then the step is repeated, step 493, for the next highest soft value. At some point a soft value will be reached which will have a majority of is or Os among the original bit and its repetitions whose soft values equal this value. The value of the original bit transmitted will then be set, step 492, to be this value .
In Figure 5 is a block diagram illustrating part of a radio receiver system as used in a GSM system. This will illustrate the preferred embodiment of the present invention. An antenna 510 of the receiver 520 of a base station, for instance, receives radio signals over a certain radio channel. The signals (e. g. data, speech) transmitted over this channel may become strongly distorted due to fading, for example, so that the transmitted bursts give rise to a distorted speech frame.
The receiver 520 converts a radio signal to a baseband signal.
The receiver 520 then sends this baseband signal to the equaliser 530 where the signal is then demodulated. The equaliser 530 also compensates for the distortion (e. g.
multipath propagation, time dispersion, etc.) caused by the SUBSTITUTE SHEET (RULE 26) The next step in the present inv radio channel. For example, in GSM a Viterbi equaliser 530 is typically used for this purpose.
"Soft" information is obtained from the equaliser 530 and is a measure of the reliability of each bit. There are several different ways of obtaining "soft" information that gives various measurements of this reliability. For example, the soft values can be obtained in the following way in a system using Trellis decoding. To each state in the Trellis (the Viterbi algorithm is used) two metrics are calculated, Mo and M1. To o each state there is also a soft value calculated defined as follows: SoftValue=~Mo-M1~. The used soft values correspond to the decided path in the Trellis.
The equaliser 530 sends the signal to the deinterleaver 540, which reorders the bitstream so that the bitstream is in the same order going into the channel decoder as it was when it originally left the channel encoder.
This signal is then sent to a channel decoder 550 to perform the opposite operation of that performed by the channel encoder on the transmitter side, i.e. to recover transmitted information 2o from the known redundant bits and the known channel coding (e. g.
a convolution code).
The decoded speech frames are delivered from the channel decoder 550 to the speech decoder 560, speech-frame by speech-frame, where a complete synthesis of received speech frames is effected so as to deliver speech signals to a converter 570 which converts speech signals from 13-bit uniform PCM to an 8-bit/A-law PCM used in the public-switched telephone network (PSTN).
In Figure 6 can be seen a speech frame 600 which contains 260 bits in accordance with the GSM recommendation. The speech frame is divided into three blocks of which each defines one of three different classes. One block of 50 bits is assigned to class 1a 610, one block of 132 bits is assigned to class Lb 620, and the remaining block of 78 bits is assigned to class 2 630. The 260 SUBSTITUTE SHEET (RULE 26~
bits are delivered from the speech encoder to form the digitised speech after speech encoding. Since a speech frame composed of 260 bits is sent every 20 ms, this gives an effective transmission rate of 13 kbit/s in full rate (FS) transmission.
5 In enhanced full rate (EFS) transmission the relevant speech parameters are coded using only 244 bits, giving a rate of slightly more than 12 kbit/s.
Class la 610: This block 610 of 50 bits are most sensitive to transmission error and cause the most problematic consequences to with regard to the intelligibility of the transmitted and decoded speech. When errors are found in these bits, large parts of the immediately preceding, correct speech frame are repeated.
This error detection is performed with the aid of three parity bits 640 which are added to the 50 data bits as control bits.
Class lb 620: This block of 132 bits is not protected by parity bits. Four bits are added as so-called tail bits 650.
These 132 bits are not equally as sensitive with regard to the intelligibility to transmission bit errors occurring as compared to the class la bits.
2o Both the class 1a and 1b bits, including the three parity bits and the four tail bits, are grouped together and convolution coded.
Class 2 630: These 78 bits are the bits least susceptible to error and are not protected at all by additional bits, as in the case of class la and lb.
These three blocks 610, 620, 630, in a speech frame contain 50+132+78=260 bits, not including the 3 parity and 4 tail bits, which, when added, give a complete total of 267 bits. Of these 267 bits, 53+136=189 bits (see second group 660) are convolution 3o encoded 265 with the rate=1/2, so a further 189 .bits are added.
This results in a total of 378+78=456 bits in the final group 670 coming from the channel encoder, which can then be interleaved for inclusion in the TDMA frames.
SUBSTITUTE SHEET (RULE 26) As mentioned above, in the enhanced full rate (EFS) mode of transmission in GSM there are used only 244 of the 260 bits available due to the greater efficiency of the speech encoding method. This means that there are an additional 16 bits that can be used to protect the other 244 bits. An 8-bit cyclic redundancy check (CRC) 680 is used for error detection among the class lb bits, while the other 8 bits 695 are used to protect a group of class 2 bits known as pulses bits 690 . These pulses bits 690 are not protected with channel coding.
!0 The remaining 8 of 16 bits available in EFS are used to protect the pulses bits 690 by repetition 695. There are four pulses bits 690 in each speech frame. These four bits 690 are each then duplicated 695, using the 8 free bits and forming a total of twelve pulses bits 690, 695.
t5 The pulses bits 690 must then be detected. The GSM specification doesn't require a particular detection method for the pulses bits, but merely requires that they be repeated. As has been mentioned above, current techniques detect the pulses bits by taking the value of each pulses bit and the values of its two 2o repetitions. A decision is then taken whereby the value of the pulses bit is decided to be the majority of the three values.
This is efficient, but not as reliable as possible using the soft values from the equaliser 530, Figure 5.
As discussed above, the equaliser 530 provides additional 25 information in the form of soft values for each bit, including the pulses bits. The present invention makes use of these soft values which provide a measure of the reliability of each bit.
In the preferred embodiment the method of the present invention will be implemented in the channel decoder 550, where the soft 30 values for all bits are available. These values are normally used in the channel decoding process.
For each of four total pulses bits there are three soft values, the soft value of the original bit plus the soft values of two SUBSTITUTE SHEET (RULE 26) WO 98!53574 PCTlSE98/00915 repetitions. The equaliser produces a soft value for each of these three bits. In the first step 710 of the method as shown in Figure 7, the bits whose values are equal to 0 are chosen.
Then the soft values of these bits are added together, step 720, to yield a value called SumSoftO, which will thereby be the sum of the soft values for all the bits equal to 0 for a given pulses bit.
The next step in the present invention is to similarly choose, step 730, the bits whose values are equal to 1. Then the soft values of these bits are added together, step 740, to yield a value called SumSofti, which will thereby be the sum of the soft values for all the bits equal to 1 for a given pulses bit.
After the values for SumSoftO and SumSoftl are determined, a decision is taken according to the algorithm below:
if SumSoftO>SumSoftl (step 750) then pulses bit . 0 (step 760) else if SumSoftO<Su~aSoftl (step 770) then pulses bit . 1 (step 780) else an ordinary majority decision takes place (step 790).
An example of this decision process can be illustrated. Assume that a pulses bit is equal to 0. It is repeated to yield three Os, but when detected two 1s and one 0 result. The soft values from the equaliser give low soft values for the is (i.e. low reliability) while it gives a high soft value for the 0 (i.e. a high reliability). The result is that the sum of the soft values for both is is still less than the soft value for the 0. Thus, the variable SumSoftO is greater than the variable SumSoftl (i.e. SumSoftO>SumSoftl), so that the decision will be taken -that the pulses bit is equal to 0 (i.e. pulses bit . 0).
SUBSTITUTE SHEET (RULE 26) ,, In this example the existing detection would detect a 1 but the present invention would detect a 0 because the reliability as derived from the soft values was greater than the reliability of a 1.
Figure 8 shows a result of a simulation comparing the present invention with the prior art method of using a majority decision. It can be seen that the present invention improves the residual bit error for pulses bits (rber~ulse5) performance between 3.4 and 4.6 dB, depending on the propagation conditions, IO and if it is interference or sensitivity that is tested.
The embodiments described above serve merely as illustration and not as limitation. It will be apparent to one of ordinary skill in the art that departures may be made from the embodiments described above without departing from the spirit and scope of t5 the invention. The invention should not be regarded as being limited to the examples described, but should be regarded instead as being equal in scope to the following claims.
SUBSTITUTE SHEET (RULE 26)
Claims (9)
1. A method for the determination of the value of a received bit in a radio communication system, said bit being repeated at least once, said bit and its said repetitions having associated soft values, said method characterized by:
adding together the associated soft values of those of said bit and its said repetitions which are equal to 0, thereby yielding a value SumSoft0;
adding together the associated soft values of those of said bit and its said repetitions which are equal to 1, thereby yielding a value SumSoft1;
performing a comparison of said values of SumSoft1 and SumSoft0;
determining the value of said bit based on the outcome of said comparison.
adding together the associated soft values of those of said bit and its said repetitions which are equal to 0, thereby yielding a value SumSoft0;
adding together the associated soft values of those of said bit and its said repetitions which are equal to 1, thereby yielding a value SumSoft1;
performing a comparison of said values of SumSoft1 and SumSoft0;
determining the value of said bit based on the outcome of said comparison.
2. The method of claim 1, wherein said step of determining the value of said bit further comprises the steps of:
deciding that said value of said bit is equal to 0 if said value of SumSoft0 is greater than said value of SumSoft1;
deciding that said value of said bit is equal to 1 if said value of SumSoft0 is less than said value of SumSoft1.
deciding that said value of said bit is equal to 0 if said value of SumSoft0 is greater than said value of SumSoft1;
deciding that said value of said bit is equal to 1 if said value of SumSoft0 is less than said value of SumSoft1.
3. The method of claim 2, further characterized in that.
deciding that said value of said bit is equal to the bit value of the majority of the said bit and its said repetitions if said value of SumSoft0 is equal to said value of SumSoft1.
deciding that said value of said bit is equal to the bit value of the majority of the said bit and its said repetitions if said value of SumSoft0 is equal to said value of SumSoft1.
4. The method of claim 2, further characterized in that;
if said value of SumSoft0 is equal to said value of SumSoft1, perform the further steps of;
selecting the largest of said soft values associated with a majority of bit values equal to either 0 or 1;
deciding that the value of said bit is equal to the bit value represented by said majority of bit values.
if said value of SumSoft0 is equal to said value of SumSoft1, perform the further steps of;
selecting the largest of said soft values associated with a majority of bit values equal to either 0 or 1;
deciding that the value of said bit is equal to the bit value represented by said majority of bit values.
5. The method of claims 1-4 further characterized in that:
said radio communication system is a GSM mobile communications system;
said bit is a pulses bit being repeated twice.
said radio communication system is a GSM mobile communications system;
said bit is a pulses bit being repeated twice.
6. An apparatus for the determination of the value of a received bit in a radio communication system, said bit being repeated at least once, said bit and its said repetitions having associated soft values, said apparatus characterized by:
means for generating a value SumSoftO by adding together the associated soft values of those of said bit and its said repetitions equal to 0;
means for generating a value SumSoftl by adding together the associated soft values of those of said bit and its said repetitions equal to 1;
means for comparing said values of SumSoft1 and SumSoftO;
means for determining the value of said bit based on the outcome of said comparison.
means for generating a value SumSoftO by adding together the associated soft values of those of said bit and its said repetitions equal to 0;
means for generating a value SumSoftl by adding together the associated soft values of those of said bit and its said repetitions equal to 1;
means for comparing said values of SumSoft1 and SumSoftO;
means for determining the value of said bit based on the outcome of said comparison.
7. The apparatus of Claim 6 further characterized by:
means for selecting the largest of said soft values associated with a majority of bit values equal to either 0 or 1;
means for deciding that the value of said bit is equal to the bit value represented by said majority of bit values.
means for selecting the largest of said soft values associated with a majority of bit values equal to either 0 or 1;
means for deciding that the value of said bit is equal to the bit value represented by said majority of bit values.
8. The apparatus of Claim 7 further characterized by:
there being two bits among said bit and its repetitions, one equal to 0 and one equal to 1, both having identical associated soft values equal to said largest soft value;
means for selecting the second largest of said associated soft values;
means for deciding that the value of said bit is equal to the value of the bit having the second largest associated soft value among said associated soft values.
there being two bits among said bit and its repetitions, one equal to 0 and one equal to 1, both having identical associated soft values equal to said largest soft value;
means for selecting the second largest of said associated soft values;
means for deciding that the value of said bit is equal to the value of the bit having the second largest associated soft value among said associated soft values.
9. An apparatus for detection of the value of a pulses bit in a GSM mobile communications system, said bit being repeated twice, said bit and its said repetitions having associated soft values, said apparatus characterized by:
means for adding together said associated soft values of those said values equal to 0, thereby yielding a value SumSoftO;
means for adding together said associated soft values of those said values equal to 1, thereby yielding a value SumSoft1;
means for determining whether said value of SumSoftO is greater than said value of SumSoft1;
means for determining if said value of SumSoftO is greater than said value of SumSoftl then deciding that said value of said pulses bit is equal to 0;
means for determining whether said value of SumSoftO is less than said value of SumSoft1;
means for determining if said value of SumSoftO is less than said value of SumSoftl then deciding that said value of said pulses bit is equal to 1;
means for determining said value of said pulses bit based on a majority decision.
means for adding together said associated soft values of those said values equal to 0, thereby yielding a value SumSoftO;
means for adding together said associated soft values of those said values equal to 1, thereby yielding a value SumSoft1;
means for determining whether said value of SumSoftO is greater than said value of SumSoft1;
means for determining if said value of SumSoftO is greater than said value of SumSoftl then deciding that said value of said pulses bit is equal to 0;
means for determining whether said value of SumSoftO is less than said value of SumSoft1;
means for determining if said value of SumSoftO is less than said value of SumSoftl then deciding that said value of said pulses bit is equal to 1;
means for determining said value of said pulses bit based on a majority decision.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| SE9701866-7 | 1997-05-20 | ||
| SE9701866A SE511310C2 (en) | 1997-05-20 | 1997-05-20 | Method for bit detection in a radio communication system |
| PCT/SE1998/000915 WO1998053574A1 (en) | 1997-05-20 | 1998-05-15 | Bit detection method in a radio communications system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA2290342A1 true CA2290342A1 (en) | 1998-11-26 |
Family
ID=20406999
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002290342A Abandoned CA2290342A1 (en) | 1997-05-20 | 1998-05-15 | Bit detection method in a radio communications system |
Country Status (10)
| Country | Link |
|---|---|
| US (1) | US6519740B1 (en) |
| EP (1) | EP0983655B1 (en) |
| JP (1) | JP2001526017A (en) |
| CN (1) | CN1153397C (en) |
| AU (1) | AU7560998A (en) |
| CA (1) | CA2290342A1 (en) |
| DE (1) | DE69825274T2 (en) |
| SE (1) | SE511310C2 (en) |
| TR (1) | TR199902533T2 (en) |
| WO (1) | WO1998053574A1 (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000026783A1 (en) * | 1998-10-30 | 2000-05-11 | Infineon Technologies Ag | Storage device for storing data and method for operating storage devices for data storage |
| US7191385B2 (en) * | 2003-04-24 | 2007-03-13 | Locus Location Systems, Llc | Error correction in a locating method and system |
| JP4287778B2 (en) * | 2004-03-31 | 2009-07-01 | 株式会社ケンウッド | Communication quality determination apparatus and communication quality determination method |
| US7900125B1 (en) | 2004-05-24 | 2011-03-01 | Seagate Technology Llc | Majority detection in error recovery |
| US20050262423A1 (en) * | 2004-05-24 | 2005-11-24 | Maxtor Corporation | Majority detection in error recovery |
| US7397843B2 (en) * | 2004-08-04 | 2008-07-08 | Telefonaktiebolaget L L M Ericsson (Publ) | Reduced complexity soft value generation for multiple-input multiple-output (MIMO) joint detection generalized RAKE (JD-GRAKE) receivers |
| US7634706B1 (en) | 2005-11-22 | 2009-12-15 | Seagate Technology Llc | Majority-detected erasure enhanced error correction |
| US8687744B2 (en) * | 2006-01-18 | 2014-04-01 | Broadcom Corporation | Method and system for an improved cellular diversity receiver |
| US7885032B1 (en) | 2006-03-06 | 2011-02-08 | Seagate Technology Llc | Apparatus and method for bulk erasure of disk drives |
| TWI400923B (en) * | 2008-09-26 | 2013-07-01 | Realtek Semiconductor Corp | Signal receiving device and frequency judging circuit |
| US10784989B2 (en) | 2018-03-14 | 2020-09-22 | Cypress Semiconductor Corporation | Bit error correction for wireless retransmission communications systems |
Family Cites Families (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3863215A (en) * | 1973-07-03 | 1975-01-28 | Rca Corp | Detector for repetitive digital codes |
| JPS5910099B2 (en) * | 1976-10-27 | 1984-03-07 | 日本電気株式会社 | Majority decision circuit |
| US4400811A (en) * | 1981-04-08 | 1983-08-23 | Rockwell International Corporation | Detecting accurrences of redundant digital codewords |
| US4518947A (en) * | 1981-09-28 | 1985-05-21 | E-Systems, Inc. | Apparatus for decoding redundant interleaved data |
| JPH01151332A (en) | 1987-12-08 | 1989-06-14 | Matsushita Electric Ind Co Ltd | Digital control signal discrimination device |
| JPH01166632A (en) * | 1987-12-22 | 1989-06-30 | Mitsubishi Electric Corp | Method and circuit for digital signal decoding |
| US5241548A (en) * | 1991-05-23 | 1993-08-31 | Motorola, Inc. | Method for error correction of a transmitted data word |
| SE470372B (en) * | 1992-06-23 | 1994-01-31 | Ericsson Telefon Ab L M | Method and device for estimating the quality of frame error detection in the receiver of a radio communication system |
| ZA947317B (en) * | 1993-09-24 | 1995-05-10 | Qualcomm Inc | Multirate serial viterbi decoder for code division multiple access system applications |
| US5502713A (en) * | 1993-12-07 | 1996-03-26 | Telefonaktiebolaget Lm Ericsson | Soft error concealment in a TDMA radio system |
| GB2311699B (en) * | 1996-03-29 | 2000-09-20 | Nokia Mobile Phones Ltd | Method and apparatus for digital communication |
| US6215762B1 (en) * | 1997-07-22 | 2001-04-10 | Ericsson Inc. | Communication system and method with orthogonal block encoding |
-
1997
- 1997-05-20 SE SE9701866A patent/SE511310C2/en not_active IP Right Cessation
-
1998
- 1998-05-15 EP EP98923281A patent/EP0983655B1/en not_active Expired - Lifetime
- 1998-05-15 WO PCT/SE1998/000915 patent/WO1998053574A1/en active IP Right Grant
- 1998-05-15 DE DE69825274T patent/DE69825274T2/en not_active Expired - Lifetime
- 1998-05-15 AU AU75609/98A patent/AU7560998A/en not_active Abandoned
- 1998-05-15 CN CNB988052822A patent/CN1153397C/en not_active Expired - Lifetime
- 1998-05-15 JP JP55028798A patent/JP2001526017A/en active Pending
- 1998-05-15 US US09/079,853 patent/US6519740B1/en not_active Expired - Lifetime
- 1998-05-15 TR TR1999/02533T patent/TR199902533T2/en unknown
- 1998-05-15 CA CA002290342A patent/CA2290342A1/en not_active Abandoned
Also Published As
| Publication number | Publication date |
|---|---|
| CN1153397C (en) | 2004-06-09 |
| EP0983655A1 (en) | 2000-03-08 |
| SE9701866D0 (en) | 1997-05-20 |
| SE9701866L (en) | 1998-11-21 |
| CN1257627A (en) | 2000-06-21 |
| JP2001526017A (en) | 2001-12-11 |
| AU7560998A (en) | 1998-12-11 |
| DE69825274D1 (en) | 2004-09-02 |
| TR199902533T2 (en) | 2000-04-21 |
| DE69825274T2 (en) | 2004-12-30 |
| SE511310C2 (en) | 1999-09-06 |
| US6519740B1 (en) | 2003-02-11 |
| WO1998053574A1 (en) | 1998-11-26 |
| EP0983655B1 (en) | 2004-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU678667B2 (en) | Soft error correction in a TDMA radio system | |
| DK175993B1 (en) | Radio receiver | |
| US5673266A (en) | Subsequent frame variable data rate indication method | |
| US5432778A (en) | Method and an arrangement for frame detection quality estimation in the receiver of a radio communication system | |
| JP3681759B2 (en) | Confidence and frame signal quality detection in soft decision convolution decoder | |
| EP0669026B1 (en) | Method and apparatus for erasing bad frames of information in a communication system | |
| JP2919072B2 (en) | Error detection system | |
| EP0643493A1 (en) | High performance error control coding in channel encoders and decoders | |
| US5901186A (en) | Bad frame detection | |
| CA2341420C (en) | Data transmission method, data transmission system, sending device and receiving device | |
| EP0983655B1 (en) | Bit detection method in a radio communications system | |
| EP0985294A1 (en) | A method for estimating channel bit error ratio, and receiver | |
| EP0651523B1 (en) | Communication receiver using soft decision to discriminate between a traffic channel frame and a fast associated control channel frame | |
| US6240538B1 (en) | Method and apparatus for errors and erasures decoding | |
| KR100213876B1 (en) | Bits error rate measuring device using viterbi decoder | |
| US6004028A (en) | Device and method for receiving and reconstructing signals with improved perceived signal quality | |
| JP3285090B2 (en) | Digital data decoder | |
| JP3003277B2 (en) | Digital data communication method and apparatus | |
| US7088778B2 (en) | Method and apparatus for measurement of channel transmission accuracy | |
| JP3383561B2 (en) | Decoding quality estimation device | |
| Shepherd et al. | Digital short range radio: channel coding and synchronisation for speech using the GSM codec | |
| CZ411199A3 (en) | Process and apparatus for determining value of received bit | |
| JPH04304728A (en) | Signal transmission system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| FZDE | Discontinued | ||
| FZDE | Discontinued |
Effective date: 20040517 |