Bad frame detection

(54)

PATENT INVENTOR INVENTOR'S OFFICE NAME

Telefonaktiebolaget LM Ericsson, 126 25 Stockholm SE Karim Jamal, Stockholm SE, Fredrik Jansson, Sundbyberg SE Dr Ludwig Brann Patentbyrå AB Detection of incorrectly received communication systems

CALLED PUBLICATIONS:

US A 5

US A 5

SUMMARY:

In a digital information piece in bursts, a first portion of the bits belonging to frames intertwined across data frames in one (56) (57)

255 343 (395: 2.51), US A 4

220 568 (371: 37.1), US A 4

829

541

526 (371:39.1),

091 (371: 37.5) radio communication systems are transmitted and received multiple bursts, while a second portion of the information bits does not belong to the frames and is not interleaved across multiple bursts but is included in each burst. A possible bit error in a redundancy containing bit sequence in the second part of information bits is detected (110). A bad frame is indicated if a bit error is detected in this bit sequence.

<img file="SE504396C2_D0001.tif" />

The numbers in brackets indicate international identification code, INID code. Letters in clamps indicate international document code.

/

504 396

FIELD

The present invention relates to a device and method for detecting bad frames in a digital communication system.

BACKGROUND OF THE INVENTION

Methods for encoding / decoding voice signals in a radio communication system are well known and even standardized (eg IS-54 in the US and GSM in Europe). Furthermore, methods for improving the encoding / decoding of background sounds, primarily intended for digital cellular telephone systems, have been described in Swedish patent application 93 00290-5. Both of these types of methods are primarily designed to handle a situation in which the connection between the speech encoder and the speech decoder is near enough ideal, in the sense that only a small amount of bit or transmission error remains after channel decoding. However, since the connection is a radio channel, the received signal may contain some bits of transmission error. In such cases, it may be necessary to modify the above methods.

Thus, it is essential to be able to reliably detect poorly received frames to perform these modified methods. Usually, a Cyclic Redundancy Check (CRC) is used at the receiver as a quality measure. If CRC fails, this is an indication that the received frame may be poor. Another quality measure is so-called soft information from the detector. This soft information essentially indicates the probability that the received speech frame (or parts thereof) is correct. One problem with the first type of measurement (CRC) is that a speech frame is interleaved over a number of time slots. Since the CRC requires bits from all these time slots before the check can be performed, this means a delay before the decision is made and error-taking measures can be taken. One problem with soft information is that it is difficult to set up

504 396 • the thresholds correctly. If the thresholds are set too low, a high frequency of false alarms will occur (acceptable frames are treated as bad frames).

An object of the present invention is a device and method in which poor frames are detected more reliably, in order to enable concealment of poor frames in accordance with the above-mentioned modified methods.

SUMMARY OF THE INVENTION

In accordance with the present invention, this object is solved by a method according to claim 1.

Further, in accordance with the invention, the object of the invention is also solved by a device according to claim 11.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention, further objects and advantages thereof, are best understood by reference to the following description and the accompanying drawings, in which

Fig. 1 is a schematic block diagram of the relevant portions of a receiver in a radio communication system containing a device in accordance with the present invention;

Fig. 2 is a timing diagram of a received signal strength and corresponding time slots; and

Fig. 3 is a flowchart of the method according to the present invention.

504 396

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

In order to understand the function of the invention, it is expedient to briefly recapitulate the function of a typical digital cellular radio connection and typical methods for concealing poor frames.

In a communication link in a digital cellular telephone system, the audio signal is first digitized, and then a speech coding algorithm is applied (see, for example, Applications of voice processing to telecommunications, Lawrence R. Rabiner, Proc. IEEE, vol. 82, no. 2, pp. 199-228) . This algorithm compresses the speech signal and converts it into a number of quantized parameters (usually in a frame-based way). The resulting bits are then protected by adding coding redundancy using channel coding methods (see, for example, GC Clark and JB Cain, Error Correction Coding for Digital Communication, Plenum Press, 1981). In addition, some bits in a frame to be transmitted are stored over several time slots. A time slot will therefore contain bits from several frames. The resulting bitstream is then modulated (see, for example, JG Proakis, Digital Communication, 2nd ed., McGraw-Hill, 1989) and broadcast, e.g. by using TDMA (TDMA = Time Division Multiple Access) methods. At the receiver, the signal is demodulated. Any time or multi-way dispersion can be counteracted by various equalization methods, e.g. Viterbi equalization or decision feedback equalization (see for example the reference by JG Proakis above). After de-interleaving, channel decoding is used (see, for example, the reference by GC Clark and JB Cain above) in order to decode the received bits. The bits that form the quantized parameters of a frame that the speech decoder needs to reconstruct the transmitted speech signal are therefore obtained by separating information from multiple time slots. From the above discussion, it is obvious that interference on the transmission channel can affect the reconstructed speech signal and thereby reduce the quality of this signal.

504 396

Although channel coding / decoding methods can significantly reduce the sensitivity to interference, it is usually not enough to apply channel coding alone in a digital cellular system. In fact, it is quite common to use so-called error-tracking methods to further mask the noticeable effects of bit errors remaining at the input of the speech decoder. These methods are mostly based on information regarding the quality of the transmission channel, such as information available or estimated on the receiver side. When such information indicates that the quality of the transmission channel is poor, the error tracking methods initiate special measures in the speech decoder with the aim of reducing the negative effects of bit errors on the reconstructed speech signal. The degree of sophistication of the error tracking methods depends on the type of information about the quality of the transmission channel. Some ways to obtain such information will now be described.

Direct information regarding channel quality can be obtained by measuring the signal strength. A low value would then indicate a low signal-to-noise ratio, which means that channel quality can be expected to be poor. Channel coding methods offer an additional level of sophistication. One method type is to use redundant channel coding, e.g. cyclic redundancy control (CRC) (see, for example, the reference by GC Clark and JB Cain above), especially when the code is used for error detection. Furthermore, soft (non-binary quantized) information can be obtained from the convolution decoder (if a convolution code is used), the demodulator, the equalizer and / or the block code decoder (see, for example, the reference by JG Proakis above). One method that is often applied is to divide the bits of information from the speech coder into different classes, each with different error correction / detection schemes, according to the significance of the different bits (see for example EIA / TIA IS-54B). Parts of the information with error detection / correction codes can be used as indicators of any bit errors that occur in the speech frame.

Some methods for introducing error tracking in conventional speech decoders for the purpose of masking frames that are considered to contain bit errors will now be briefly described. When a bad frame is detected, it is common to use the information from the previous, accepted frame. Often this method is combined with attenuation of the output level in case the situation with poor frames would remain under several frames (see for example EIA / TIA IS-54B). This situation is not uncommon in mobile phone systems, where fading valleys can persist for fairly long periods of time in case the mobile speed is low. The result of the attenuation is that the interference is masked in the reconstructed signal. In particular, strong click sounds are avoided. If more detailed information is available regarding the quality of each received portion of incoming bits, it is possible to trace any transmission errors to some of the speech decoder parameters. Since these parameters are a model of different phenomena in the speech, error-tracking methods can be developed that are optimized for the physical significance of each individual parameter. One example is the so-called pitch gain parameter (see, for example, TB Minde et al., Techniques for low bit rate speech coding using long analysis frames, ICASSP, Minneapolis, USA, 1993). A pitch gain value greater than one is sometimes required for this parameter during transient speech periods. However, such a value corresponds to an unstable filter model, which means that it can be somewhat risky to use one. In particular, it is appropriate to introduce error-tracking methods that limit pitch gain to values ​​less than one in the event that a possible bit error is detected in this parameter. Another example is the spectral filter model often used in modern speech coding algorithms (see, for example, the reference by TB Minde et al. Above). In this case, error tracking methods can be used to prevent the use of unstable filters when bit errors are indicated in the corresponding spectral information. The reverse is also relevant; when an unstable filter is detected, a poor frame can be indicated and error-tracking methods applied.

504 396

For example, in US digital cellular systems in accordance with IS-54B standard, a stolen frame for FACCH would result in a lost speech frame at the speech decoder in the receiver. The speech decoder solves this problem by filling in the appropriate information. Usually, corresponding information from the previous frame is used instead of the lost frame.

If the decoder provides so-called anti-swirling for background noise, the quality of the resulting audio signal may be unacceptable if the above-mentioned concealable speech methods are used. As described in Swedish patent application 93 00290-5, this anti-porous action can be performed in several ways. One possible measure is a bandwidth expansion of the filter. This means that the poles of the filter are moved to the origin in the complex plane. Another possible modification is low-pass filtering of the filter parameters in the time domain. That is, rapid frame-to-frame variations of the filter parameters, or of representations thereof, are smoothed by low-pass filtering of at least some of these parameters. A special case of this method is the averaging of a representation of the filter parameters across multiple frames.

With the aid of this background information, the present invention will now be described with reference to Fig. 1. Fig. 1 shows the portions of a receiver in a mobile radio communication system necessary to describe a preferred embodiment of the present invention. An antenna receives the information from the transmitted signal and passes it over an input line 10 to a downconverter 12. The downconverter 12 converts the received signal to the baseband and passes it over a line 14 to a data detection circuit, which in the illustrated embodiment is represented by an equalizer 16, e.g. a Viterbi equalizer, which converts the received and downconverted signal into a bit stream, which in turn is passed over a line 18 to a dis-interleaver 20. If the received frame contains bits from a traffic channel, the bit stream is passed over a line 22 to a channel decoder 24.

RHA

Channel decoder 24 converts the bit stream nun into a filter parameter stream and an excitation parameter stream for speech decoding.

Equalizer 16 also feeds soft information about received bits or symbols to a decision means 28 over line 52. As mentioned above, the channel decoder converts the bit stream into a filter parameter stream and an excitation parameter stream for speech decoding in a speech decoder 40, which outputs an audio signal to an output signal 42. 24 decoding with cyclic redundancy control (CRC) on at least parts of each received frame. The results of these controls are passed to the decision means 28 over a line 26. A concealer 32 connected to the decision means 28 controls the processing of filter and excitation parameters in the speech decoder 40 via a line 46. The concealer 32 can be implemented as a state machine by a microprocessor.

In a preferred embodiment, the receiver also contains a speech detector (not shown). A suitable speech detector is described in WO 89/08910 in the name of British Telecom PLC. This speech detector determines from the filter and excitation parameters whether the received frame contains primary speech or background noise. The speech detector's decision is fed to a parameter modifier for modifying the received filter parameters (if desired, no signal discriminator for determining whether received signals representing background noise are stationary or not be included between the speech detector and the parameter modifier). This modification is described in detail in Swedish patent application 93 00290-5, which is hereby incorporated by reference. The possibly modified filter parameters and excitation parameters are fed to the speech decoder 40.

The cyclic redundancy control from channel decoder 24 is formed frame by frame and not time slot by time slot. However, it would be desirable to have an indication of a poor time slot for time slot. Thus, it would be desirable to perform error detection, e.g. a cyclic redundancy check, time slot for

504 396 time slot. Such detection would give an early warning of an incoming poor frame. For example, in the US standard IS54B, there is a signal in each time slot that identifies each channel, namely the CDVCC (Coded Digital Verification Color Code). This is a 12-bit long, coded, non-cached version of the eight-bit digital verification color code and is transmitted in each slot in both the forward and return links. The code used is a simple (15, 11) code abbreviated to a (12, 8) Hamming code and is therefore a code that corrects single errors. In the receiver, e.g. in the channel decoder 24, it is therefore possible to construct a flag DvccError which is true if at least one error has occurred in the decoded DVCC. Therefore, DvccError can be used as an indication of a bad frame. This indication is sent to the decision body 28 on line 26. There, it can be combined with other information received by decision maker 28. This process will be further described with reference to Fig. 3.

Fig. 2 illustrates the signal strength S of a received signal as a function of time T. Fig. 2 also illustrates the corresponding time slots A, ..., F. Each time slot contains aggregated information from two speech frames (such as IS-54B). For example, time slot B contains information from frames k-2 and k-1, while time slot C contains the remaining information from frame k-1 and information from frame k. The gaps between the illustrated time slots represent the other channels (corresponding time slots have been omitted) that occupy the same frequency.

When a frame is collapsed after de-interleaving, a CRC is performed. For example, CRC (k-2) is performed when frame k-2 has been composed after flattening the time slots A and B. In the figure, the time slots C and D contain poor information, because the signal strength is too low below these slots. Therefore, CRC (k) will fail (CRC (k) = NOT OK). However, frame k-1 is considered acceptable, since CRC (k-1) did not fail (CRC (kl) = OK) despite time slot C being received with a very low signal strength and may contain errors. The result is that frame k-1 is accepted as

504 396 the last correctly received frame although it may contain errors.

According to the present invention, error detection is also performed on the DVCC field in each time slot. This error detection (DVCC = NOT OK) on the DVCC field will detect a poor time slot in slot C and give an indication that frame k-1 may contain errors, since parts of this frame are received in time slot C. Therefore, frame k 1 is declared unacceptable and frame k-2 is considered the latest acceptable frame. This will give a more reliable input to an error-tracking algorithm.

A preferred embodiment of the method according to the present invention will now be described with reference to the flow chart of Figure 3. Only the steps essential for understanding the present invention are shown. In step 100, the routine tests whether ordinary CRC for a frame has failed. If this is the case, the routine proceeds to an error-following routine in step 120. Otherwise, the DvccError flag is tested in step 110. If the flag is true, it is also tested whether SoftQual is less than a TH1 threshold. If both of these conditions are true, the routine proceeds to the error-completing routine in step 120. Otherwise, step 130 tests whether SoftQual is less than a second TH2 threshold. If this is the case, the error-following routine is performed in step 120. Otherwise, the routine ends in step 140. Without the present invention, there would have been no step 110. Because of step 110, the TH1 threshold can be set more aggressively (higher), since the test is combined with the test of the DvccError flag. Thus, if a time slot has both a fairly low soft information quality and a true DvccError flag, then error-following actions are required. If the DvccError flag is not true, the SoftQual parameter can still trigger the error-tracking algorithm in step 120, but in this case it must be lower than a second threshold TH2, which is more conservative (lower) than the first threshold TH1. Thus, in this case, the soft information must more strongly indicate a poor time slot to trigger the error-tracking algorithm in step 120.

504 396

The invention has been described with reference to the digital verification color code. However, from the above description, it is obvious that other channel-encoded parameters that exist in each time slot could be used instead of or in combination with the digital verification color code. It is an essential feature of the invention that a selected parameter should contain redundancy (e.g., channel coded) to allow easy error detection.

Furthermore, the invention has been described with reference to TDMA type digital radio communication systems. However, the same principles can also be applied to any system in which information is transmitted in signal bursts. E.g. it is not necessary that several channels share the same frequency or that the bursts are broadcast at regular time intervals. The invention can also be used in e.g. wired communication systems.

It is also possible to generalize the ideas described above by observing that it is not in fact necessary to require that the tested, redundancy-containing parameter must be non-interleaved. Similar improvements can be obtained if the parameter has a merging depth less than the merging depth of a speech frame (here the merging depth is defined as the number of bursts over which the information is merged or merged). An earlier indication of a poor frame is still obtained, since the parameter is interleaved earlier than the complete speech frame.

Those skilled in the art will appreciate that various modifications and changes can be made to the invention without departing from its basic idea and framework, as defined in the appended claims.

504 396