JP3251591B2 - Apparatus and method for rate determination in on / off variable rate communication system - Google Patents

Description

Description: FIELD OF THE INVENTION The present invention relates to rate determination in variable rate communication systems, and more particularly to rate determination in communication systems where on / off or amplitude modulation keying is used to implement a variable rate transmission scheme. Regarding judgment.

Background of the Invention Efficient use of limited resources has long been an important goal of communication system designers. For example, in a point-to-point communication system, such as a telephone modem, this means maximizing the rate of information transmission (in terms of bits of information per second) for limited transmitter power and channel bandwidth. I do. In a multiple access system such as a cellular wireless network, the available resources are shared among many users, and the goal of the system designer is to maximize the capacity of the system. This is often expressed as the number of users that the system can support simultaneously at a particular information transmission rate per user.

In a multiple access system that supports voice transmission, the efficiency of the communication system can be further improved by utilizing the nature of each user's voice activity. For example, a well-known method called Digital Speech Interpolation (summarized in Advanced Digital Communications by K. Feher) digitally encodes and packetizes the speech of each N user and then converts the packet stream. Multiplexing on a single communication channel has long been used to increase the capacity of trunked telephone systems. Although it is not impossible that some packets will be dropped, the available capacity of the link is increased.

European Telecommunications Standards Institute (ETSI: European Telecommun
GS issued by the Communications Standards Institute)
Related cases arise in cellular wireless systems such as the M (Groupe Special Mobile) system specification. In this case, a voice activity detector in the digital voice encoder used to encode each user's voice disables the user's transmitter when the user is not talking. "Discontinuous Transm
ission) "or" TDX "
It has the important effect of saving mobile station battery consumption,
Nominally, it allows shorter radio frequency (RF) channel reuse distances (since the average co-channel interference power is reduced), thereby improving system capacity.

North American Telecommunications Industry Associati
on) Standard IS-95-A Mobile Station Base Station-C
ompatibility Standard for Dual-Mode Wideband Spre
A similar method is employed in the cellular communication system described in ad Spectrum Cellular System. This system is based on code division multiple access (CDMA).
e Access) based air interface. Because the capacity of such systems is limited by the mutual interference between users, the system seeks to minimize the power transmitted by each user by utilizing voice activity.

The TIA standard IS-95-A describes different ways to achieve this on each forward link (base station to mobile station) and reverse link (mobile station to base station),
In the description herein, only the methods used on the reverse link are relevant. As shown in FIG. 1 (enlarged summary of drawing 6.1.3.1-2 of TIA standard IS-95-A), the mobile station (M
S: mobile station) divides the 8 kHz pulse code modulation (PCM) user voice signal (100) into 20 ms segments or frames and then encodes these frames into information packets using a digital voice encoder (101). Become
The strict specification of the digital voice encoder (101) is TIA standard IS-96-A Speech Service Option Standard for W
ideband Spread Spectrum Described in the Digital Cellular System.

During encoding, each frame is classified by the digital audio encoder (101) and the associated audio activity detector (107) as belonging to one of four distinct transmission rates. Here, these are "rate 1 /
1 ”,“ Rate 1/2 ”,“ Rate 1/4 ”, and“ Rate 1 ”
/ 8 ". Again, the exact description of the voice activity detector (107) is described in the TIA standard IS-96-A. The speech encoder (101) uses more information bits to encode a frame, which occurs during active talk spurts, and uses fewer bits during silence periods. Rate 1/1 uses the most information bits and rate 1/8 uses the least information bits. Bit utilization of frames encoded at rate 1/2 and rate 1/4, which generally occurs during the transition between active utterance and silence, is between these limits.

The result of this frame classification process is the rate indicator (10) generated by the voice activity detector (107).
8) is displayed on the digital audio encoder (101). Next, rate 1/1 and rate 1/2 packets are
Standard IS-95-A Section 6.1.3.3.2.1 Revere Traffi
Block coding or cyclic coding (102) is performed as specified in c Channel Frame Quality Indicator. This is followed by TIA standard IS-95-A section 6.1.
In 3.1.3 Convolutional Encoding, channel coding (103) specified as a rate 1/3 convolutional code is performed. At this point, Table 1 shows the number of channel coded bits that make up the coded packet (104).

Interleave (105) (TIA standard IS-95-A section
6.1.3.1.5 Block Interleaving)
After that, each coded packet is further prepared for transmission using 64-hex quadrature modulation followed by direct sequence spreading using a 1.2288 Mc / s user-specific pseudo-noise (PN) code (105). (TIA Standard IS-95-A Section 6.1.
3.1.6 Orthogonal Modulation and 6.1.3.1.9 Quadr
ature Spreading). For this description only the following details of this process are required. From Table 1, since the modulation scheme is a 64-hex orthogonal scheme, 96 symbols (generally Walsh symbols) are required to transmit a rate 1/1 packet. Similarly, a rate 1/2 transmission requires 48 symbols, a rate 1/4 transmission requires 24 symbols, and a rate 1/8 transmission requires 12 symbols. The TIA standard IS-95-A has 6 symbols.
Specifies transmission in bursts of a number of consecutive Walsh symbols. To support this, TIA standard I
S-95-A transmits a 20 ms traffic channel frame to a “power control group” (PCG).
ps), where each group can transmit one group of six Walsh symbols (TIA Standard IS-95-A Section 6.
See 1.3.1.7 Variable Data Rate Transmission).
The MS is actively transmitting according to the selected rate.
The number of PCGs is 16, 8, 4 or 2 in total from rate 1/1 to rate 1/8. Note that the MS transmitter is disabled during inactive PCG.

This process is described in more detail in FIG. 2 which shows the transmitter activity during a 20 ms traffic channel frame (200) for each packet size or rate. In Fig. 2, the PCG period (2
The hatched portion of 02) means that a burst of 6 direct sequence spreading Walsh symbols was transmitted during the PCG. Therefore, the selected packet rate is
If 1/1, all 16 PCGs in the 20 ms traffic channel frame become active as shown by case (202) in FIG. If the selected packet rate is rate 1/2, the MS transmitter will only be active during 8PCG as shown in case (203). Similarly, rate 1/4 packets generate an active PCG as shown in case (204), and a rate 1/8 example is shown in case (205).

Note that for rates 1/2, 1/4 and 1/8 (sometimes referred to collectively as "sub-rates"), the PCG active during any 20 ms frame is based on the TIA standard IS-95-A section 6.1.3.1. 7.2 Data Burst Randomizing Algorithm
Pseudorandom PCG driven by monitoring user-specific PN codes in traffic channel frames prior to the frame to be analyzed, as described in
It is determined from the selection procedure. Specifically, the last 14 bits of the generated user-specific PN code are stored in the penultimate PCG of the previous frame, and this bit is used to activate the active PN code for each transmission rate in the subsequent frames. Select PCG. Since the user-specific code has a repetition period of several days and is shifted for each user,
The position of the active PCG in a particular traffic channel frame depends on the time at which the frame is transmitted, the identity of the user and the rate of the packets transmitted during the frame. Of course, the number of active PCGs remains constant at each rate, and only the location of the active PCG changes pseudo-randomly with time and user identification.

It is important to understand that to perform de-spreading, the base station (BS) receiver regenerates the user-specific PN code used by the MS to perform direct sequence spreading. That is something that must be done. Thus, the BS receiver can clearly identify the PCG used to transmit packets at any of the four possible rates. Furthermore, since the PCG selection procedure only depends on the user-specific PN code sequence observed during the previous frame, the BS receiver can identify the active PCG of the current frame at the start of the current frame.

By utilizing this variable rate transmission method, TIA
Standard IS-95-A mobile stations can reduce battery consumption and the average amount of interference power provided to other IS-95-A mobile stations using the same carrier frequency. However, TIA standard IS-96
-A and the traffic channel physical layer structure defined in the TIA standard IS-95-A are defined as 4 for a given packet transmission.
Note that no side information is provided to indicate to the base station receiver which of the two rates was selected. Therefore, the base station receiver needs to estimate the rate of transmission, and this process is called "rate determination."

Conventional methods for rate determination include TIA standard IS-94-
Qualcomm In used in A-compliant base station receivers
c, of San Diego, CA Base station modem (B
SM: Base Station Modem). FIG. 3 shows a simplified block diagram of this method.

The procedure begins by first converting a received radio frequency (RF) waveform consisting of each direct sequence spread Walsh symbol from an RF signal to a baseband signal sampled at a chip rate. This includes frequency conversion, automatic gain control, symbol sampling, etc.
Various well-known RF, intermediate frequency (IF) and baseband functions are required, but need not be specified here in detail. Each transmitted Walsh symbol waveform is corrupted by noise and, after distortion by the communication channel, recovered by despreading (302), which is used to spread the transmitted Walsh symbols. Requires a correlation with the user-specific PN sequence (303). Next, Walsh symbol detection is performed in a hexadecimal correlator (312) using a well-known method of correlating the received Walsh symbol waveform with the set of 64 waveforms that make up the symbol alphabet (30).
Four). The exact details of this method and its performance are well known and are described in standard literature, including Digital Communications by JG Proakis. The maximum likelihood method of identifying a transmitted Walsh symbol in the absence of a received signal phase reference (ie, a "non-coherent" state) is based on the complex-valued cross-correlation between the 64 possible symbol waveforms and the received waveform. Correlator to maximize size and corresponding symbol
Select an index. This selection process is performed in the block labeled "MAX" (313) in FIG. The magnitude square of this maximum magnitude cross-correlation result (ma
gnitude-square) values are referred to here as "the maximum Walsh
Symbol energy or MWSE, output (31
5) is shown to be generated.

Next, the index (314) corresponding to the correlator that provided the largest magnitude correlator output is sent to the block deinterleaver (305), where the channel symbols that make up each Walsh symbol are deinterleaved. Convolutional decoding using the well-known Viterbi algorithm (30
6) Yes. The deinterleaving process (305) and the Viterbi decoding process (306) are performed four times,
It is performed once on the assumption that each possible packet transmission rate is used. During Viterbi decoding (306), the estimate of the number of channel symbol errors present in the received packet is compared to the received channel coded symbols with the symbols obtained by re-convolution coding the Viterbi decoder output for each rate. Calculated for each of the four possible rates by known methods. 4 to accommodate the estimated number of symbol errors for each rate
The hexadecimal vector (308) is sent to the rate determination function (309). Finally, the blocks or cyclic codes (referred to as Frame Quality Indicators in TIA standard IS-95-A) associated with rate 1/1 and rate 1/2 packets are decoded (307) and their syndromes or checks are performed. Sam (Error Contr by S.Lin and DJCostello
ol Coding) is a binary vector (31
0) is given to the rate judgment function (309).

Next, the number of quaternary symbol errors and the binary checksum
By dividing the hexadecimal decision space formed by the vectors (309) and (310) and then estimating the transmitted rate by identifying the region of the decision space where the hexadecimal vector corresponding to the received packet is located , A rate determination is performed. If the rate of the transmitted packet cannot be reliably determined (i.e., is out of the decision window for all four rates), the packet is declared "deleted." No further processing such as voice decoding is performed on such a packet.

This method has the disadvantage of requiring four separate Viterbi decoding operations (one for each transmission rate hypothesis). This is computationally expensive and inefficient in terms of power consumption. In addition, audio delays caused by the need to perform multiple Viterbi decodings before sending the determined rate to the audio decoder can degrade perceptual sound quality. Therefore, there is a need for a rate determining apparatus and method in a variable rate communication system that is computationally and power efficient and does not degrade perceptual sound quality.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 shows a conventional method for performing voice and channel coding and modulation and direct sequence spreading of reverse link traffic channel frames in a communication system having a variable rate voice service option. .

FIG. 2 shows the reverse link traffic channel channel when using the variable rate voice service option.
4 shows an example of a conventional power control group (PCG) for a particular frame of a frame.

FIG. 3 shows a conventional rate determination method for reverse link traffic channel packets.

FIG. 4 is a block diagram showing a preferred embodiment of the rate judgment device according to the present invention.

FIG. 5 is a block diagram showing another preferred embodiment of the rate judgment device according to the present invention.

Detailed Description of the Preferred Embodiment A block diagram of the preferred embodiment of the rate determination device is shown in FIG. FIG. 4 shows the RF conversion (300, 300) used to restore the noisy quadrature waveform corresponding to each transmitted Walsh symbol making up the traffic channel frame.
301) and despreading (302) processes. Also user-specific used to generate the despreading sequence
A PN code generator (303) and a Walsh symbol detector (304) constituted by a hexadecimal correlator (312) and a selector function (313) are shown. , The largest Walsh
A Walsh symbol detector that produces a symbol energy (MWSE) value (315) is shown.

The MWSE for each received Walsh symbol is "Rate 1/1
"R-1 / 1Acc." (409) ~
Sent to four accumulators (409-412), abbreviated as "R-1 / 8Acc." (412). The contents of these accumulators are
Set uniformly to zero at the beginning of each traffic channel frame. Adders associated with each accumulator (405-408)
Is gated by a control signal from the PCG selector function (400), and accumulates the MWSE only when the corresponding PCG flag has a logical value “1”. PCG selector function (400)
Monitors the output of the user-specific PN code on the previous traffic channel frame and observes the TIA standard IS-95-A section 6.1.3.1.7.2 Data Burst Randomizing Algorithm.
For each possible transmission rate, identify the active PCG in the current traffic channel frame based on the method defined by. The PCG selector (400) outputs four binary flags (401 to 404) marked "rate 1/1 PCG flag" to "rate 1/8 PCG flag". The Walsh symbol demodulated by the detector (304) is the active PCG for the rate associated with this flag.
Only if they are part of a logical "1"
Becomes That is, referring to the example of FIG.
The CG flag becomes logic “1” only in the shaded portion of the traffic channel frame corresponding to the case (203). Similarly, the rate 1/4 PCG flag takes the value "1" only during the period of the hatched part in the case (204). This process continues until all 96 Walsh symbols making up the frame have been received. Therefore, at the end of the frame, "R
The accumulator labeled "-1 / 1Acc." Accumulates 96 MWSE values, "R-1 / 2Acc." Accumulates 48 MWSE values, and "R-1
/ 4Acc. "Accumulates 24 MWSE values, and" R-1 / 8Acc. "
Accumulates two MWSE values.

After all frames have been received, the contents of the accumulators (409-412) are sent to the subtractors (417-420). Generally different for each accumulator. In FIG. 4, M-1 / 1 to M-1 / 8 (413 to 41
The scalar value marked 6) is subtracted from the output of each accumulator (409-412). Next, the output of each subtractor (417-420) is sent to a selector (421), and this selector (421) selects the maximum of the outputs of the four subtractors (417-420). This uniquely identifies which of the accumulators (409-412) has the maximum value at selector (421). Next, the estimated transmission rate (422) is identified as the rate corresponding to the accumulator (409-412) that generated the maximum value in the selector (421). This estimated rate (422) is
Deinterleaver (423) and Viterbi decoder (424)
It is used to control the operation of. These devices perform only once to decode the received channel symbols based on the rate predicted by the estimated rate (422).

It should be noted that the scalar values M−1 / 1 to M−1 / 8 (413 to 416) are set in advance by computer simulation or bench test. In practice, these scalar values are substantially constant throughout the duration of the traffic channel frame being analyzed.
However, the scalar value changes at each Walsh symbol boundary depending on other parameters associated with the base station receiver. One particular example of this arises from using a "rake" receiver to take advantage of the presence of multipath signal components in the communication channel. [Rake receivers are well known in the art,
There is no need to explain here. In this regard, the case of a four-element rake receiver is shown in FIG. 5, where each element (500) comprises at least a despreader (302) and a hexadecimal correlator (312). Note that four or more or less elements may be utilized depending on the application. In FIG.
Each element is assigned to a separate multipath signal component,
The observed delay difference of each multipath signal component is compensated by delay elements Δ ₁ -Δ ₄ . Next, each despread multipath signal component is converted to a hexadecimal correlator (312) of the same type as above.
To Next, the hexadecimal vector is assigned to each correlator (31
2) formed at the output of where the hexadecimal vector i
The ith element is the magnitude squared (magnit
ude-square). With this, Di by JGProakis
A typical non-coherent synthesis method called "square law" synthesis in standard literature such as gital Communications is obtained. In FIG. 5, four such vectors are combined by simple vector addition (522). Next, the maximum Walsh symbol energy (MWSE) (315) is identified as the largest element of the generated composite hexadecimal vector (523), and the MWSE (315) then determines the rate as shown in FIG. Used for

However, the relative strength of each multipath component may change over time. As such, the number of elements (500) that process multipath components that significantly contribute to the vector synthesis process (522) can vary. Assuming that the RF converter (301) has an automatic gain control stage, each element (50
The relative contribution of (0) can be estimated using a simple signal-to-noise ratio (SNR) estimator (501) as shown in FIG. The SNR estimator compares the MWSE (506-509) of the individual hexadecimal vectors at the output of the correlator (312) for each element with a threshold T (51).
It works by comparing with 0). MWS for a specific element
If E (506-509) exceeds T (510), the corresponding hexadecimal vector at the hexadecimal correlator output of this element is included in the synthesis process (522), otherwise excluded.
When a device is declared to be included in the synthesis process, the device is said to be "in-lock." The in-lock indication of each element is shown in FIG. 5 as binary lock flags L1 to L4 (515 to 518). Also, L1
A counter (519) is also shown which accumulates the number of elements currently in-locked by monitoring ~ L4 (515-518). This count (520) is stored in the 4x4 lookup table (52
Used to get the row address in 1). The columns of this table have M used according to the number of in-lock elements.
Contains values from −1/1 (413) to M−1 / 8 (416). The contents of the look-up table are preset by computer simulation or bench test.

Obviously, the estimated rate (430) need not be determined exclusively from the metrics available at the output of the subtractors (417-420) in FIG. Instead, these metrics can be used as supplementary information for rate determination, for example, based on symbol error rate or path metric information derived from a Viterbi decoder. It is also clear that this method can easily be extended to estimate the rate of variable rate transmissions derived from sources other than speech. This includes variable rate data transmission.

From the foregoing, it will be readily appreciated that the present invention provides a power and computationally efficient apparatus and method for rate determination that does not adversely affect perceptual sound quality. Still further advantages and modifications will readily occur to those skilled in the art. Therefore,
The invention in its broader aspects is not limited to the specific details, representative devices, and illustrative examples illustrated herein. Various modifications and variations may be made to the above specification without departing from the scope or spirit of the invention, and the invention is intended to cover all such modifications as falling within the scope of the appended claims and their equivalents. And variations.

Claims (6)

(57) [Claims] 1. A method for determining a transmission rate of a received signal in a variable rate communication system comprising: determining a first plurality of power control groups (PCGs) that are active for a first transmission rate; Determining a second plurality of PCGs that are active for a transmission rate of 2; calculating a first cumulative energy metric based on symbol energy for the first plurality of PCGs; Calculating a second cumulative energy metric based on symbol energy for the plurality of PCGs; and selecting a rate for the received signal based on the first and second cumulative energy metrics. A method characterized by being performed. 2. The method of claim 1, wherein calculating the cumulative energy metric includes calculating a metric based on a symbol energy that depends on the number of elements of the rake receiver in use. the method of. 3. The method of claim 1, further comprising calculating at least one additional metric based on one of a convolutional decoder distance metric, a Viterbi decoder metric, and a block coding metric. the method of. 4. An apparatus for determining a transmission rate of a received signal in a variable rate communication system, comprising: determining a first plurality of power control groups (PCGs) active for a first transmission rate; A power control group selector that determines a second plurality of PCGs that are active for a transmission rate of 2; a second control unit that calculates a first cumulative energy metric based on symbol energy for the first plurality of PCGs; 1
A second accumulator for calculating a second accumulated energy metric based on symbol energy for said second plurality of PCGs.
A selector for selecting a rate for the received signal based on the first and second accumulated energy metrics. 5. A power control group flag generator coupled to each of said accumulators, wherein said accumulator selectively selects a maximum Walsh symbol energy in response to the presence of a power control group flag. 5. The apparatus according to claim 4, wherein the accumulation is performed. 6. The apparatus of claim 4, wherein said selector includes a plurality of subtractors each coupled to an output of said plurality of accumulators for subtracting a scalar value from said output.