Dynamic iterative decoding for balancing quality of service parameters

[0002] This invention relates to decoding of a coded information signal in a telecommunications system and, more particularly, to a method and apparatus for iterative decoding of a coded information signal that allows quality of service parameters to be dynamically balanced in a telecommunications system.

[0003] Multidimensional coding, or turbo coding, and associated iterative decoding methods have increasingly important application as data transmission services grow in sophistication. For example, next generation wireless communications systems currently under development are expected to support internet-based services, such as e-mail and web browsing. It has been proposed for at least some of these wireless communications systems to utilize turbo coding for the higher speed data transmission of the system.

[0004] A turbo decoder decodes an encoded signal by recursively processing frames of the encoded signal using more than one iteration through the decoder, as described, for example, by Berrou in “New Shannon Limit Error-correcting and Decoding: Turbo-codes (1)” Proceedings ICC 1993, pp. 1064-1070, and by Berrou and Glavieux in “Turbo-codes: General Principles and Applications,” Audio and Video Digital Radio Broadcasting Systems and Techniques,1993, pp. 215-226.

[0005] Several schemes have been presented in the literature for improving the performance of iterative decoders. U.S. Pat. No. 5,761,248, “Method and Arrangement for Determining an Adaptive Abort Criterion in Iterative Decoding of Multi-Dimensionally Coded Information,” Hagenover, et al., discloses a method and apparatus for comparing a weighted decision on a soft output of a substep of iterative decoding to a weighted decision on a soft output and a combination of previous subsets of iterating and aborting the iterative decoding dependent on the comparison result. The publication, Variable Latency Turbo Codes for Wireless Multimedia Applications,” by C. Valenti and D. Woerner, Proceedings of International Symposium on Turbo Codes and Related Topics, Brest, France, September 1997, pp. 216-219, disclosed the use of interleavers of variable sizes to achieve different quality of service, QoS, requirements.

[0006] Each of the above methods has its drawbacks. Using the first approach, there is no adaptive QoS consideration. With the second approach, using different interleavers for QoS may be difficult due to incompatibility with specifications of transmission frames per data rate. It also requires both the transmitter and receiver to change interleavers frequently, which is undesirable.

[0007] The present invention provides an iterative decoder that allows quality of service, QoS, parameters to be dynamically balanced. By dynamically adjusting the number of iterations, N, within an appropriate range, a coded information signal may be iteratively decoded so that at least one quality of service parameter remains within an acceptable performance range.

[0008] In an embodiment of the invention, the invention is implemented in an iterative decoder that performs iterative decoding on a coded information signal based on minimum, N_min, and maximum, N_max, values for the number of decoding iterations, N, to be performed for a particular data transmission, with N_minand N_maxdetermined according to QoS requirements. QoS requirements for a particular data service may be expressed in terms of a maximum delay, Tdelay, and a maximum bit error rate (BER). The BER and maximum Tdelay may then be used to determine N_minand N_maxfor decoding the coded information signal. The determination of N_minand N_maxmay be made based on data tables stored in memory that include data on BER versus N and Tdelay versus N for a particular decoder configuration to be used in the iterative decoder.

[0009] In the embodiment, the iterative decoder receives the coded information signal in the form of data frames over a data transmission channel. For each data frame, N_miniterations of decoding are performed. After N_miniterations are performed, a convolutional redundancy code (CRC) check is performed on the decoded results of N_miniterations. If the CRC check indicates that the data frame has been correctly decoded, the iterative decoder outputs the result and processes the next frame. If the CRC check indicates that the data frame has not been correctly decoded, the iterative decoder performs at least one other decoding iteration and CRC check. The decoding iterations are repeated until the CRC check indicates that the data frame has been correctly decoded or until the number of iterations is equal to N_max. If the number of iterations becomes equal to N_max, the data frame is considered in error and discarded and the decoding is begun on the next data frame. The iterative decoder may be configured to initiate transmission of a frame retransmit request to the transmitter if the data frame is discarded.

[0010] FIG. 1 is a functional schematic diagram of an iterative decoder according to an embodiment of the invention;

[0011] FIG. 2 is a flow diagram illustrating process steps performed according to an embodiment of the invention; and

[0012] FIGS. 3A and 3B are a representative bit error rate versus number of decoder iterations plot, and a decoding delay versus number of decoder iterations plot, respectively, according to an embodiment of the invention.

[0013] Iterative decoding according to the invention may be implemented into a receiver or a telecommunications system in which available system data services may each have different quality of service (QoS) requirements. The telecommunications system could be any type of data transmission system in which transmitted data is iteratively decoded. For example, a turbo decoder according an embodiment of the invention may be implemented into the receiver of a mobile station or base station of a cellular telecommunications system operating according to any one of a number of known standards or standards to be developed, such as the cdma2000 ITU-R RTT candidate submission Vol. 18/Jul. 27, 1998, published by the Telecommunications Industry Association (cdma2000) standard or the “Specification of Air-Interface for 3G Mobile System, Version 0.5,” Jul. 21, 1998, Association of Radio Industries and Businesses (ARIB), Japan, which specify a turbo coding and decoding scheme for high-speed data transmission. The iterative decoder could also be implemented into other systems using wireless transmission, such as satellite systems, high definition television systems, or data transmission methods other than wireless, for example, landline or optical fibers. Referring now to FIG. 1, therein is a functional block diagram of an iterative decoder 100 according to an embodiment of the invention. Iterative decoder 100 comprises input buffer 102, output buffer 116, N_min/N_maxcounter 120, decision circuitry 118, control processor 132, and decoder circuitry 134. Decoder circuitry 134 comprises first constituent decoder 104, interleaver 106, interleaver 108, second constituent decoder 112, de-interleaver 114, and de-interleaver 110.

[0014] Iterative decoder 100 provides iterative decoding of a coded information signal that may be dynamically adjusted to balance quality of service (QoS) requirements for a particular data service. Iterative decoder 100 may decode a coded information signal that is encoded according to any type of iterative coding scheme, such as, for example, recursive convolutional systematic turbo coding. The interleaver size and code rate of the particular code may be of any value acceptable to provide data rates within the range of acceptable values for a particular data service. Iterative decoder 100 is then controlled, according to the invention, to allow QoS parameters to be dynamically adjusted within the range of acceptable values for the particular data service to provide enhanced performance.

[0015] QoS requirements for a particular data transmission service may be expressed in terms of a maximum delay, Tdelay, and a maximum bit error rate (BER). In the embodiment of the invention, Tdelay and BER are both taken into consideration when dynamically determining upper and/or lower limits for the number of iterations of turbo decoding, N, to be performed in the decoding of a received data frame. The determination of upper and/or lower limits for N, using both Tdelay and BER, allows turbo coding of a data frame to be performed in a manner that balances the effects of the QoS requirements on the decoding. By choosing an N_maxand N_minbased on Tdelay and BER, respectively, the chances of decoding a data frame in a manner that meets both of these requirements are increased. Additionally, actual delay may be reduced as much as possible. If BER requirements can be met with a minimum number of decoder iterations on a good data transmission link, the delay will be minimal for that BER. This maximizes the quality of the data service.

[0016] Referring now to FIG. 2, therein is a flow diagram showing process steps performed by iterative decoder 100 according to the embodiment of the invention. The process of FIG. 2 begins at step 200 when one or more data frames are received to be decoded in iterative decoder 100. Next, at step 202, control processor 132 determines a value for a minimum number of iterations, N_min, and a maximum number of iterations, N_max, to be performed in the turbo decoding on each data frame. N_minand N_maxmay be determined from QoS information that is transmitted to control processor 132 at the setup or initialization of the transmission link. In the embodiment of FIG. 1, the QoS information in the form of a bit error rate (BER) and maximum delay, Tdelay, allowed for the turbo decoder processing is received at data transmission link 122 and transferred to control processor 132 through buffer 102. Control processor 132 may then determine N_minand N_maxfrom look-up tables stored in a memory in control processor 132. The look-up tables may be created from data for values of BER and Tdelay over the data link versus number of iterative decoder iterations, N, in average channel conditions for a particular turbo-decoding configuration.

[0017] Referring now to FIG. 3A and FIG. 3B, therein are illustrated plots representing BER and Tdelay, respectively, versus number of turbo decoder iterations, N, for a particular turbo-decoding configuration. In systems in which it is possible to provide data channels of varying rates or interference quality, the system may dynamically determine the QoS requirements based on the particular configuration of data channels used to transmit the information. For example, for a higher speed transmission link, the system Tdelay requirements for the iterative decoder may be less stringent, allowing a higher Tdelay. Data representing a plurality of pairs of curves, such as those illustrated in FIG. 3A and FIG. 3B, may be stored within control processor 132. Each pair of curves may represent the BER and Tdelay characteristic for a particular iterative decoding configuration that may be used in iterative decoder 100. Control processor 132 may then use the information identifying the particular iterative coding scheme being used to choose the appropriate pair of curves.

[0018] Given a BER of BER_x, based on a particular transmission rate, N_minis chosen from data stored in memory of control processor 132. This is illustrated in FIG. 3A. Tdelay_xis then used to choose N_maxfrom the data stored in memory of control processor as illustrated in FIG. 3B. If N_minis greater than N_max, the value of N_minwill be used for N_max, and N_minwill be set to zero.

[0019] Next, at step 204, a data frame received on the link is input on data transmission link 122 to input buffer 102 and processed in decoder circuitry 134 for N_miniterations. Processing for N_miniterations before any decision is made reduces unnecessary processing that would otherwise be done when the chances of meeting the BER requirements were low. Control processor 132 generates appropriate control signals on data line 124 to input buffer 102 and data line 130 to N_min/N_maxcounter 120, so that each data frame is processed through the correct number of decoding iterations according to the embodiment of the invention. Next, at step 206, a hard decision is made in decision circuitry 118 on the decoding output of de-interleaver 114 after N_mindecoding iterations. Decision circuitry 118 determines whether a CRC check for the current data frame indicates a CRC error and generates a decision result signal to control processor 132. Next, at step 208, control processor 132 determines if there is a CRC error. If it is determined at step 208 that there is no CRC error, the process moves to step 210, where control processor 132 generates a control signal to output buffer 116 causing output buffer 116 to output the decoded data frame. Next, at step 220, control processor 132 determines if another frame exists in input buffer 102 for processing. If it is determined that another frame does not exist, the process moves to step 222 and ends. If, however, at step 220 it is determined that another frame exists for processing in input buffer 102, the process moves to step 204. At step 204, the next data frame is input from input buffer 102 to first constituent decoder 104 and interleaver 106. The process then begins again for the next data frame.

[0020] If, however, at step 208, it is determined that the CRC check indicates a CRC error, the process moves to step 212. At step 212, a field indicating the total number of iterations for the current data frame being decoded is incremented by one. The process then moves to step 214. At step 214, control processor 132 determines if the total number of iterations is less than or equal to N_max. If the number of iterations is not less than or equal to N_max, the process moves to step 218. At step 218, control processor 132 generates control signals to cause output buffer 116 to discard the current frame. Control processor 132 may also generate signals to an associated transmitter (not shown) that would initiate a retransmission request in the reverse direction over data transmission link 122 for a discarded frame. Next, at step 220, control processor 132 determines if another frame exists in input buffer 102 for processing. If it is determined that another frame does not exist, the process moves to step 222 and ends. If, however, at step 220 it is determined that another frame exists for processing in input buffer 102, the process moves to step 204. At step 204, the next data frame is input from input buffer 102 to first constituent decoder 104 and interleaver 106. The process then begins again for the next data frame.

[0021] While the invention has been described in the context of particular embodiments, various other embodiments and variations of these particular embodiments may be realized. For example, the processing to determine the values of N_minand N_maxfrom QoS requirements may be performed completely by processors on the transmitting side of the data transmission link, and N_minand N_maxmay be transmitted to iterative decoder 100 directly. The various steps of the processing could also be allocated between the transmitting side of the data transmission link and turbo decoder 100 in other ways depending on the system resources and desired allocation of processing tasks. Thus, although the method and apparatus of the present invention has been illustrated and described with respect to presently preferred embodiments thereof, it will be understood that numerous modifications and substitutions may be made to the embodiments described and that numerous other embodiments of the invention may be implemented without departing from the spirit and scope of the invention as defined in the following claims.