MXPA06008312A - Data detection for a hierarchical coded data transmission - Google Patents

Abstract

Techniques for performing data detection for a hierarchical coded data transmission are described. In one data detection scheme, log-likelihood ratios (LLRs) for code bits of a first data stream are initially derived based on received symbols for the data transmission. The LLRs for the first data stream are decoded to obtain decoded data, which is re-encoded and remodulated to obtain remodulated symbols. Interference due to the first data stream is estimated based on the remodulated symbols. LLRs for code bits of a second data stream are derived based on the LLRs for the code bits of the first data stream and the estimated interference. The LLRs for the first data stream may be derived from the received symbols in real-time without buffering the received symbols. The LLRs for the second data stream may be derived after the first data stream has been decoded.

Description

DETECTION OF DATA FOR A TRANSMISSION OF HIERARCHICAL CODIFIED DATA I. FIELD OF THE INVENTION The present invention relates generally to communication, and more specifically to techniques for performing data detection for transmission of hierarchical encoded data in a wireless communication system.

II. BACKGROUND OF THE INVENTION Hierarchical coding is a data transmission technique in which multiple (eg, two) data streams overlap (e.g., aggregated) together and are transmitted simultaneously. "Coding" in this context refers to channel coding instead of data encoding in a transmitter. Hierarchical coding can advantageously be used, for example, to transmit broadcast services to users within a designated broadcast area. These users can experience different channel conditions and achieve different signal-to-noise-e-interference ratios (SNRs). Consequently, these users are able to receive data at different data rates. With hierarchical coding, the broadcast data can be divided into a "base current" and an "improved stream". The base current is processed and transmitted in such a way that all users in the diffusion area can recover the current. The improved current is processed and transmitted in such a way that users with better channel conditions can recover the current. To recover a transmission of hierarchical encoded data, a receiver first detects and recovers the base current by treating the enhanced current as noise. The receiver then estimates and cancels the interference due to the base current. The receiver then detects and recovers the improved current with interference from the canceled base current. For better performance, the base current and the improved current are typically recovered sequentially, one stream at a time, in the order described above. Typically, a large amount of processing is required to recover each current. In addition, a large amount of intermediate storage may also be required depending on the manner and speed at which each stream can be detected and recovered. Large amounts of processing and buffering can affect the performance and cost of the system. Therefore, there is a need in the art for techniques that effectively perform data detection for a transmission of hierarchical encoded data.

SUMMARY OF THE INVENTION The techniques for performing data detection for a transmission of hierarchical encoded data is described therein. These techniques can be used for single carrier wireless communication systems as well as multiple carrier (for example, OFDM). In a data detection scheme, the received symbols are initially obtained for a transmission of hierarchical encoded data with multiple data streams (for example, two) and logarithmic probability relationships (LLRs) for code bits of a first stream data (the base current) are derived based on the received symbols. The LLRs for the first data stream are decoded to obtain decoded data which, additionally, is recoded and re-modulated to obtain demodulated symbols for the first data stream. Interference due to the first data stream is estimated based on the demodulated symbols. The LLRs for code bits of a second data stream (the enhanced stream) are then derived based on the LLRs for the code bits of the first data stream and the estimated interference. The LLRs for the first data stream may be (1) derived from the received symbols in real time without buffering the received symbols and (2) stored in a buffer for decoding. The LLRs for the second data stream can be (1) derived after the first data stream has been decoded and (2) stored in the same buffer by overwriting the LLRs for the first data stream. The received symbols are not used to derive the LLRs for the second data stream and, therefore, they do not need to be stored in the buffer. In another data detection scheme, the LLRs for the code bits of the first data stream are initially derived based on the received symbols. Estimates of data symbols (or uncoded firm decision symbols) for the first data stream are then derived based on the received symbols or the LLRs for the first data stream. The interference due to the first data stream is estimated based on the data symbol estimates and is canceled from the received symbols to obtain symbols canceled by interference. The LLRs for the code bits of the second data stream are then derived based on the symbols canceled by interference. The LLRs for both the first and second data streams can be calculated from the symbols received in real time without buffering the received symbols. The LLRs for the second data stream can be adjusted / updated after the first data stream has been decoded by (1) detecting errors in the data symbol estimates based on the demodulated symbols for the first data stream and both ( 2a) set the LLRs for the code bits of the data symbol estimates that are in error for cancellations or (2b) modify the LLRs for the code bits of the data symbol estimates that are in error with correction factors derivatives based on the remodulated symbols and the data symbol estimates. Various aspects and embodiments of the invention are described in greater detail below.

BRIEF DESCRIPTION OF THE DRAWINGS The characteristics and nature of the present invention will become more evident from the detailed description set forth below when taken together with the drawings in which similar reference characters identify correspondingly to the end and where: FIGURE 1 shows a transmitter and a receiver in a wireless communication system; FIGURE 2A shows a signal constellation for the QPSK; FIGURE 2B shows a signal constellation for the hierarchical coding with the QPSK for both the base current and the improved current; FIGURE 3 shows a reception processor (RX) for a first data detection scheme; FIGURE 4 shows an RX processor for a second data detection scheme; FIGURE 5 shows an RX processor for a third data detection scheme; and FIGURE 6 shows an RX processor for the second data detection scheme with a higher order modulation scheme for the base current.

DETAILED DESCRIPTION OF THE INVENTION The word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any modality or design described here as "exemplary" does not necessarily have to be interpreted as preferred or advantageous over other modalities or designs. FIGURE 1 shows a block diagram of a transmitter 110 and a receiver 150 in a wireless communication system 100. In the transmitter 110, a coder / modulator 122a within a data transmission (TX) processor 120 receives, encodes, interpolates, and modulates (i.e., symbol maps) a base data stream (indicated as. ,.}.) and provides a corresponding base symbol stream (indicated as { s¿.}.). An encoder / modulator 122b, similarly, receives, encodes, interpolates, and modulates an improved data stream (indicated as {e.}.) And provides an improved symbol stream (indicated as {se. . ) correspondent. The data for each stream is typically encoded in packets, each packet being encoded separately in the transmitter and separately decoded in the receiver. Each of the streams of symbols. { sb} Y . { HE} contains "data symbols", which are modulation symbols for data. A combiner 130 receives and combines the streams of base and enhanced symbols. Within the combiner 130, a multiplier 132a receives and multiplies the base symbol stream. { sb} with a scale factor Kb? and a multiplier 132b receives and multiplies the improved symbol stream. { HE} with a scale factor Ke. The scale factors Kb and Ke determine the amount of transmission power to be used for the base current and the enhanced current, respectively. A larger fraction of the total Potai transmitted power is typically assigned to the base current. An adder 134 receives and adds the scale data symbols from the multiplier 132a with the scale data symbols from the multiplier 132b and provides combined or compound symbols, which may be expressed as: a. x. = Kb ..sb + Ke .se, Ec (1) where sb is a data symbol for the base current, it is a data symbol for the enhanced current, and x is a combined symbol. The climbing and the combination are done based on symbol by symbol. A transmitting unit 138 (TMTR) receives a stream of combined symbols (indicated as { X.}.) From the combiner 130 and the pilot symbols, processes the combined symbols and pilots based on the system design, and generates a more modulated signals. A pilot symbol is a modulation symbol for a pilot, which is known a priori by both the transmitter and the receiver and can be used by the receiver for channel estimation and other purposes. The transmitter unit 138 can perform an orthogonal frequency division multiplexing (OFDM) modulation to transmit the combined and pilot symbols over multiple subbands, a spatial processing for transmitting the combined symbols and pilots from multiple antennas, etc. The modulated signal (s) are transmitted through a wireless channel to a receiver 150. At the receiver 150, a receiving unit (RCVR) 160 receives a or more signals through a wireless channel, processes the received signal or signals in a manner complementary to the processing performed by the transmitting unit 138, provides the received pilot symbols (indicated as { yP.}.) to a channel estimator 162, and provides a stream of received symbols (indicated as { and.}.) to an RX processor 170. The received symbols can be expressed as: Y = h -x + n = h - [Kb - sb + Ke - s + n, Ec (2) where h is a gain of the complex channel for the combined symbol x, n is the noise observed by the combined symbol x, ey is the symbol received for the combined x symbol. The noise n includes interference and channel noise, receiver circuitry noise, etc. The channel estimator 162 estimates the response of the wireless channel based on the received pilot symbols and provides gain estimates of the channel. { ñ} . To simplify it, the description in it assumes the estimation of the channel without error, that is, fi = h. The RX processor 170 includes detectors 172 and 176, an interference canceller 174, decoders 182 and 186, and an encoder / modulator 184. The detector 172 performs the detection of data in the stream of received symbols. { Y} for the base current and provides the symbols detected for the base current (indicated as { sjb.}.). Each symbol §i, detected is an estimate of the data symbol sb and can be represented, for example, by a set of logarithmic probability relationships (LLR), as described below. The decoder 182 decodes the detected symbols for the base current and provides a decoded base current (indicated as { ¿T.}.). The encoder / modulator 184 then recodes and reorders the decoded base current in the same manner as that performed by the transmitter 110 and provides a re-modulated base current (indicated as {sb.}.), Which is an estimate of the current of base symbols. { s} . The interference canceller 174 receives the re-modulated base current, estimates and cancels the interference due to the base current from the stream of received symbols, and provides a stream of canceled interference symbols (indicated as {e.}.).) To the detector 176. The detector 176 performs the detection of data on the stream of canceled interference symbols. { ye } for the improved current and provides symbols detected for the enhanced current (indicated as {se se.}.). The decoder 186 decodes the detected symbols for the enhanced current and provides an improved decoded current (indicated as,.}.).

The controllers 140 and 190 direct the operation to the transmitter 110 and receiver 150, respectively. The memory units 142 and 192 provide storage for the data and programming codes used by the controllers 140 and 190, respectively. The data detection performed by the receiver is influenced by several factors such as the modulation scheme used by each data stream, the specific form used to represent the detected symbols, the technique used to perform the data detection, etc. clarity, the data detection for a transmission of hierarchical encoded data with quadrature phase shift modulation (QPSK) for both streams and using the LLRs to represent the detected symbols is specifically described below. FIGURE 2A shows a signal constellation 200 for the QPSK, which includes four signal points 210a to 210d in a complex two-dimensional plane. These four signal points are located in coordinates of 1 + jl, 1 - jl, -1 + jl, and -1 - jl and are given the label of ll ',? 10', 01 ', 00', respectively. For the modulation of the QPSK, each pair of code bits (indicated as bx and b2) is mapped to one of the four possible signal points, and the complex value for the correlated signal point is the modulation symbol for the pair of code bits.

For example, bit b may be used for a component in phase (I) and bit b2 may be used for a quadrature component (Q) of the modulation symbol. In this case, the modulation symbol for each pair of code bits can be expressed as: s = bx + jb2, where bx e. { l, -l} and b2 e. { l, -1} . FIGURE 2B shows a signal constellation 250 for the hierarchical coding with the QPSK for both the base current and the enhanced current. The constellation of the QPSK for the base current is represented by four signal points 210a to 210d. The constellation of the QPSK for the enhanced current is superimposed on the constellation of the QPSK for the base current and is represented by four points 260a to 260d of signal at each signal point 210. The factors Kb and Ke of scale determine (1) the distance between the signal points 210 of the base current and the center of the complex plane and (2) the distance between the signal points 260 of the enhanced current and the points 210 of base current signal. Returning to FIGURE 2A, with the QPSK, a modulation symbol for one of only four possible signal points is transmitted for each pair of code bits. However, due to noise, interference, and distortion in the wireless channel, a received symbol (eg, symbol 212 in FIGURE 2A) may not fall directly into one of the four possible signal points. The data detection is done to remove the effect of the wireless channel (for example, to remove the gain h from the complex channel) and to determine which of the four possible signal points is the symbol s of transmitted data. The information for each detected symbol s is often represented in the form of an LLR for each of the two constituents of code bits bx and b2 for the detected symbol. Each LLR indicates the probability that the code bit b ± is a one ("1" or +1) or a zero ("0" or -1). The LLR for the i-code bit of the detected symbol s can be expressed as: where b ± is the code bit i ° for the detected symbol s; Pr (s \ b ± = 1) is the probability of the symbol s detected with bit b ± being 1; to. Pr (s \ b ± = -1) is the probability of the symbol s detected with a bit b ± being -1; and b. LLR ± is the LLR of the bit in code b ±. An LLR is a bipolar value, with a higher positive value corresponding to a higher probability that the code bit is a +1 and a larger negative value corresponding to a higher probability that the code bit is -1. A LLR of zero indicates that the code bit is equally likely to be +1 or -1. The LLR for each bit in code is typically quantized to a predetermined number of bits (or L bits, where L> 1) to facilitate storage. The number of bits to be used for the LLRs depends on several factors such as the decoder requirements, the SNR of the detected symbols, etc. FIGURE 1 shows a symbolic representation of the data detection for a transmission of hierarchical encoded data. The detection of data can be done in several ways. Below, three data detection schemes are described. FIGURE 3 shows an RX processor 170a for a first data detection scheme in which the base current and the enhanced current are both detected based on symbols. { Y} received. The RX processor 170a is an embodiment of the RX processor 170 in FIGURE 1. Within the RX processor 170a, the symbols. { Y} received are initially stored in a buffer 314. A unit 320 for calculating the LLR of the base current retrieves the received symbols from the buffer 314 and performs data detection on each symbol and received to obtain two LLRs for two bits of data. code of a symbol sb of the base current that is carried in that received symbol. The two LLRs for the base current can be expressed as: LLR »+ JLLRb2 = j,.?- 3L- *. , Ec 4) where LLRb and LLRb2 are the LLRs for the two bits of the symbol s¿ of the base current within the symbol and received; to. ii is a channel gain estimate for the symbol and received; b. "*" indicates a complex conjugate; c. Eb is the energy of the sb symbol of the base current; and d. N0, b is the power of the interference and the noise observed by the symbol sb of the base current. It is assumed that the symbol sb of the base current has a complex value of ± ^ jEb 12 ± j ^ b / 2. The energy of the base current symbol is Eb = EloUl -Kbl / (Kl + Ke), and the energy of the improved current symbol is Et = Elolal -Ke l (Kb +? ß), where Etotai is the total energy for the combined x symbol. The power N0 / jb of the interference and noise includes the channel noise N0 and the interference from the enhanced current. The calculation unit 320 provides the LLR of the base current (indicated as { LLRb.}.) Through a multiplexer 322 (Mux) to a buffer 324 for storage. The decoder 182 receives and decodes the LLRs of the base current from a buffer 324 and provides data. { S} decoded for the base current. The decoder 182 can implement a Turbo decoder or a Viterbi decoder if the Turbo coding or the convolution coding, respectively, was performed on the transmitter. A Turbo decoder performs the decoding in the LLRs for multiple iterations to obtain ever better estimates of the transmitted data bits. The decoding process typically requires some amount of time to complete and may require additional storage for the base current LLRs for the duration of the decoding process (eg, for a Turbo decoder). After the LLR of the base current has been decoded, the data. { dbJ decoded are recoded and demodulated by an encoder / modulator 184 to obtain sb symbols} remodulated An interference estimator 330 receives and multiplies the symbols . { sb ".}. demodulated with the estimate { h.}. of channel gain and provides estimates { ib.}. interference due to the base current. An adder 332 receives and subtracts the estimates. .} interference from the received { and.}. symbols obtained from the buffer 314 and provides the symbols {and.}. canceled by interference, which may be expressed as: a. yh-sb ^ y ~ h'Sb.Ec (5) The symbol . { sb} remodulated is equal to the sb symbol of the base current if the base current is decoded correctly. If the base current is decoded correctly or in error, it can be determined based on a CRC (cyclic redundancy check) or some other error detection scheme. A LLR calculation unit 340 of the improved stream performs data detection on the symbols. { ye } canceled by interference to obtain two LLRs for the two code bits of each symbol will be of the enhanced current. The two LLRs for the enhanced current can be expressed as: LLRel + jLLRe2, Ec (6) where LLRe and LLRe are the LLRs for the two bits of the symbol of the derived enhanced current based on the symbol e canceled by interference; to. Ee is the power of the symbol of the improved current; and b. No, e is the power of the interference and the noise observed by the symbol of the improved current.

The calculation unit 340 provides the LLRs of the improved current (indicated as { LL.Re.}.) Through a Mux 322 to the buffer 324 for storage. The decoder 182 then decodes the LLRs of the enhanced current to obtain decoded data for the enhanced current. { from} . For the first data detection scheme, the RX processor 170a needs to store the symbols. { Y} received in a buffer 314 and the LLR of the base current in a buffer 324 while the base current is decoded by a decoder 182. The size of the buffers 314 and 324 depends on the size of the data packet, delays in the decoding, and possibly other factors. The same buffer 324 can be used to store both the LLR of the base current and the LLR of the improved current since these currents are decoded sequentially. FIGURE 4 shows an RX processor 170b for a second data detection scheme in which the base current is detected based on the symbols. { Y} received and the improved current is detected based on the LLR of the base current. The RX processor 170b is another embodiment of the RX processor 170 in FIGURE 1. Within the RX processor 170b, a unit 420 for calculating the LLR of the base stream performs a data detection on the symbols. { Y} received to obtain the LLR. { LLRb} of the base current, as shown in equation (4). The calculation unit 420 provides the LLRs of the base current through a multiplexer 422 to a buffer 424 for storage. The decoder 182 receives and decodes the LLRs of the base current from a buffer 424 and provides data. { dbJ decoded for the base current. After the LLRs of the base current have been decoded, the encoder / modulator 184 recodes and re-modulates the data. { dbJ decoded to obtain remodulated symbols sb} for the base current. The LLRs of the base current are derived from, and are closely related to, the received symbols. The LLRs of the enhanced current can thus be calculated directly from the LLRs of the base current instead of the received symbols. The LLR of the improved current can be represented as: LLRe? + jLLRe2 where h indicates the gain estimate of the channel power for the symbol and received. The first equality in equation (7) is obtained by substituting equation (5) for equation (6). The amount within the parentheses in the third equality is for the LLR of the base current. Equation (7) indicates that the LLRs of the enhanced current can be derived from the LLRs of the base current and the demodulated symbols. Inside the RX 170b processor, a multiplier 426 receives and scales the LLRs of the base current with the gain G and provides the LLRs of the scaled base current. An interference estimator 430 receives and multiplies each symbol. { } remodulated with both, the estimate \ h \ 2 of power gain and the gain G2 to obtain an interference estimate = sb '\ h \ O2 due to the base current. The processing by the interference estimator 430 is different from the processing by the interference estimator 330 in FIGURE 3. An adder 432 receives and subtracts the interference estimate h from the scaled base current LLR and provides the LLRs of the enhanced current , which are sent through a Mux 422 and the buffer 424 for storage. The decoder 182 then decodes the LLRs of the enhanced stream to obtain data. { from} decoded for the improved current. As shown in equation (7), the symbols. { Y} received are not used to derive LLRs from the enhanced current. For the second data detection scheme, the RX processor 170b does not need to store the received symbols and only a buffer 424 can be used to store both the LLRs of the base current and the LLRs of the enhanced current. This can greatly reduce the buffering requirements for the receiver. The LLRs of the base current are quantized and stored with a sufficient number of bits such that these LLRs provide a good decoding performance for the base current and can be subsequently used to derive the LLRs from the enhanced current. For the second data detection scheme, the number of bits to be used for the LLRs of the base current affects the precision and margin of the LLRs of both currents. In a specific implementation of Turbo decoder, LLRs are quantized in six bits with a range of [-8, 8] and an accuracy of 0.25. The precision indicates the maximum possible quantization error. Arates, the range and precision are typically selected based on the decoding performance and are only indirectly related to the quantization signal-to-noise ratio (SQNR). In addition, the range and precision are typically not changed based on factors such as code rate or SNR operation. The accuracy of LLRs of the enhanced current is affected by the gain G used to scale the LLR of the base current in equation (7). If the power N0 / £, of the interference and the noise observed by the base current is dominated by the noise N0 of the channel and not the interference of the improved current, then N0, ±, is approximately equal to N0 and the improved current will have an S? R smaller than that of the base current because a lower power is typically used for the improved current. In this case, the gain G will be less than one and, because the LLRs of the base current are scaled by GX / the accuracy of the LLRs of the enhanced current is not affected by the accuracy of the LLRs of the base current. However, if the gain Gx is greater than one, then one or more additional lower order / less significant bits may be used for the LLRs of the base current. The quantification interval should be wide enough so that the LLRs of the base current are not saturated or reduced to a too low value, which can degrade performance. The saturation of the LLRs of the base current typically does not pose a serious problem for the Turbo decoder but can severely affect the quality of the LLRs of the improved current that derive from the LLRs of the base current. To determine how many additional higher order bits are needed to avoid LLR saturation, the symbol and received in equation (4) can be replaced by h - sb + t ?, as follows: LLRM + jLLR? where nb represents the noise and interference observed by the base current, which includes interference from the enhanced current. If a symbol of the base current sb = JEb / 2 + j-¡Eb ¡2 is transmitted, then each of the two LLRs for the symbol sb would have a mean of flb = 2 b '\ h \ / N? J , = 2'SNRb and one standard deviation of sb = The SNR for the base current is then SNRb ~ Eb - \ h \ 2 / N0 / b. Assuming a "reasonable" variation of the average term plus three times the standard deviation for the symbol and received, the magnitude of the LLR should be able to take values up to μi + $ sb -2-SNRb + 6tjSNRb. This number increases with a SNR for the increasing base current. Thus, the worst case is when the channel noise is zero and the SRN of the base current hits a background noise caused by interference from the enhanced current. At this point, the SNR of the base current is SNRb = Eb / Ee, and the maximum LLR magnitude to accommodate is 2-Eb! Ee + 6 Eb / Ee. It should be noted that this is a conservative interval because when N0lb is dominated by interference from the enhanced current, the noise is no more Gaussian noise but noise of the QPSK, which does not vary beyond the average term plus a single standard deviation. The number of bits to be used for the base current LLRs can be selected based on the energy ratio of the base current symbol for the power of the enhanced current symbol. For example, if the power of the base current is four times stronger than that of the improved current (or Eb / Ee = 4), then the LLR of the base current should be quantified with a magnitude of up to the implementation of the decoder Exemplary turbo described above with a range of [-8, 8], the LLRs of the base current can be quantized and stored with two additional bits of higher order, or 8 bits in total. As another example, if the power of the base current is nine times stronger than that of the improved current (or Eb / Ee = 9), then the LLR of the base current should be quantified with a magnitude of up to 2-9 + 6V9 = 36, and three bits of higher order can be used for LLRs. Although the LLRs of the base current may be stored with additional bits for the second data detection scheme, the total memory required may still be significantly less than that of the first data detection scheme, which stores both the received symbols and the LLRs. of the base current. This is especially true since the received symbols probably also require a larger bit width in the presence of the enhanced current. FIGURE 5 shows an RX processor 170c for a third data detection scheme in which the base current is detected based on the symbols. { Y} received and the enhanced current is detected using an uncoded interference cancellation. However, the RX processor 170c is another embodiment of the RX processor 170 of FIGURE 1. Within the RX processor 170c, a LLR calculation unit 520 of the base stream performs data detection on the symbols. { Y} received to obtain the LLR of the base current, as shown in equation (4). The calculation unit 520 provides the LLRs of the base current to a buffer 524 for storage. The decoder 182 receives the LLRs of the base current from a buffer 524 through a multiplexer 526, decodes these LLRs, and provides data. { db decoded for the base current. For the third data detection scheme, the LLRs of the enhanced current are calculated from the symbols. { Y} received, similar to the first data detection scheme. However, the interference due to the base current is estimated based on the estimates of uncoded data symbols (instead of the demodulated symbols) for the base current. Thus, the LLR of the improved current can be calculated at the same time with the LLR of the base current, instead of having to wait for the decoding of the base current to be completed. An estimate sb 'of uncoded data symbols (or simply, an estimate of data symbols) is an estimate of the sb symbol of the base current obtained by making a firm decision on either a symbol and received or the LLRs of the base current for the symbol and received. For example, to refer to FIGURE 2A, an estimate of data symbols for a received symbol 212 may be the signal point at 1 + j'l, which is the signal point closest to the symbol 212 received. The data symbol estimates are derived based on the symbols received without the benefit of the error correction capability of the code used for the base current. In this way, the data symbol estimates are more prone to errors than the demodulated symbols, which benefit from the error correction capability of the base current code. Consequently, the estimates # í} Uncoded interference from derived data symbol estimates are less reliable, and LLRs from the enhanced current derived from the symbols. { ye '} canceled by uncoded interference are also less reliable than those derived by the first data detection scheme. The performance of the decoding for the enhanced current may be degraded if the LLRs for the data symbol estimates that are in error are given high reliability values (or greater weight) in the decoding process. Several schemes can be used to mitigate the deleterious effects of data symbol errors (or firm decision symbol errors) in the decoding of the enhanced stream. Data symbol errors can be detected by comparing each demodulated symbol against the corresponding data symbol estimate and declaring an error if two are not equal.

In a first error compensation scheme, the LLRs of the enhanced current corresponding to the estimates of data symbols that are in error are not given weight in the decoding process. This can be achieved by adjusting these LLRs to cancellations, which are zero LLR values that indicate equal probability that the code bits are +1 or -1. If the symbol error ratio (SER) is relatively low, then the effects of using cancellations for the LLRs corresponding to data symbol errors may be minor. For example, in a background noise of 6dB (corresponding to the base current that has four times the power of the improved current), the SER is approximately two percent. The degradation in the performance of the encoding to declare these firm decision symbol errors as cancellations should not be meaningful. In a second error compensation scheme, the LLRs of the improved current obtained with estimates of data symbols that are in error are updated based on the demodulated symbols after the base current has been decoded. The LLR of the improved stream of equation (7) can be expressed as: LLR? + jLLRe2 = (LLRe'l + jLLRe'2) + (sb '- a)' \ \ 2 -G 2 > where LLRe'l and LLRe'2 SOn LLR initials for the two bits of a symbol are of the enhanced current. The equation (9) indicates that the LLR? ' and initial LLRe'2 can be obtained based on the symbols and received and the estimation sb 'of the data symbols. Once the base current has been decoded and the demodulated symbols are available, the initial LLRe'l and LLRe'2 can be updated with the remodulated symbol to obtain the LLR? and LLRe2, endings, which can be decoded to obtain decoded data for the enhanced stream. If the initial LLRs are saturated, then these LLRs can be adjusted in cancellations. The final LLRs are retained even if they are saturated. Within the RX processor 170c, a firm decision unit 528 receives either the LLR of the base current (as shown in FIGURE 5) or the received symbols (not shown in FIGURE 5) and makes a firm decision to derive the estimates. { sb '} of data symbols for the base current. The firm decision can be made as is known in the art. For example, each data symbol estimate can be adjusted to the nearest signal point in distance to the received symbol. Unlike remodulated symbols, data symbol estimates can be derived with a minimum delay. An interference estimator 530 receives and multiplies the estimates. { sb '} of data symbols with estimates. { h} of channel gain and provides estimates. { ib '} of uncoded interference due to the base current, which can be expressed as: ib '= h -sb'. An adder 532 receives and subtracts the estimates. { b} of interference from the symbols. { Y} received and provides iye 'symbols} > canceled by uncoded interference, which can be expressed as: e - and ~ ib '= y-h -sb'. A 540 LLR calculation unit of the improved stream performs data detection on the symbols. { ye '} canceled by uncoded interference to obtain the LLR. { LLRer} , of the initial improved current, similar to that shown in equation (6). The calculation unit 540 provides the LLRs of the initial improved stream to a buffer 544 for storage. After the LLRs of the base stream have been decoded, an encoder / modulator 184 recodes and re-modulates the data. { db} decoded to obtain symbols. { } remodulated for the base current. A symbol error detector 542 receives the symbols. { Yes! } remodulated and estimates. { sb '} , of data symbols, detects errors in the data symbol estimates, and provides an indication for each data symbol estimate detected in error. The symbol error detector 542 can additionally calculate a correction factor cb ~ (sb '-s6) - | / t | 2 -G2 for each data symbol estimate that is in error, if the second error compensation scheme is used. A LLR adjusting unit 546 receives and adjusts the LLRs of the initial improved current. { LLRe '} from the buffer 544 and provides the LLRs. { LLRe} of the final enhanced current through a multiplexer 526 to a decoder 182. The LLR setting unit 546 can (1) adjust the LLR of the enhanced current for estimates of data symbols that are in error for cancellations, for the first error compensation scheme, or (2) add the correction factor cb to the LLR of the initial improved current for each data symbol estimate that is in error, for the second error compensation scheme. For the third error detection scheme, the RX processor 170c does not need to store the received symbols, and the two buffers 524 and 544 are used to store the LLRs of the base current and the LLRs of the enhanced current, respectively. For clarity, the three data detection schemes have been described above for the QPSK. These data detection schemes can also be used for higher order modulation schemes, which are modulation schemes of higher order than the QPSK. The first and third data detection scheme can be used in the manner described above with any modulation scheme for the base current and any modulation scheme for the enhanced current. For the third data detection scheme, the correction factor cb -. { sb '~ ^ b)' \ ^ \ 2 'G2 can be used to update the initial LLRs, according to the modulation scheme used for the enhanced stream, provided that the data symbol estimates are detected in error. For the second data detection scheme, the LLRs of the base current contain all the information in the received symbols and can thus be used to estimate or reconstruct the received symbols. The LLRs of the enhanced current can then be calculated from the received symbol estimates. The estimation of the symbols received from the LLRs of the base current can be performed as described below. To simplify it, the following description assumes that a correlated modulation scheme of Gray with a higher order than the QPSK is used for the base current. With the correlation of Gray, the neighboring signal points in a constellation (both in the horizontal and vertical directions for a square constellation) have labels that differ in only one bit position. Gray's correlation reduces the number of code bit errors for most probable error events, which correspond to a received symbol that is correlated to a signal point near the correct signal point, at which only one bit in code would be in error The following description also assumes that the LLRs of the base current are calculated using a "maximum dual approach", which can be expressed as: where LLR ±, is the LLR for the bit in code i-th for the symbol and received; to. s? X is a hypothetical modulation symbol that is close to the symbol and received so that the i-th code bit for s ± X has a value of +1; and b. s ± r 0 is a hypothetical modulation symbol that is close to the symbol and received so that the code bit i-th for s ±? 0 has a value of -1. A signal constellation for an M-Aryo modulation scheme PSK or M-QAM contains M signal points. Each signal point is associated with a B-bit tag, where B = log2 M. B bits of code B are mapped to a modulation symbol, which is a complex value for a signal point whose tag is equal to the values of the code bits B. The LLR B are calculated for each symbol and received, each LLR being calculated based on a respective pair of hypothetical modulation symbols, s ±, ys ± t 0. Equation (10) provides an equation with respect to the symbol and received for each code bit of a modulation symbol s that is detected. Thus, there are three equations for each symbol received for the 8-PSK (B = 3), four equations for each symbol received for the 16-QAM (B = 4), etc. It can be shown that the B equations derived from Equation (10) for code bits B are linear equations. From these B equations, two unknown ones can be determined, particularly the real and imaginary parts of the symbol and received. However, the difficulty is that a different pair of symbols s?,? And yes, or hypothetical modulation is used for each of the equations B for the symbol and received, and these hypothetical modulation symbols are unknown. For the correlated Gray 8-PSK and 16-QAM, the hypothetical modulation symbols can be determined for at least two of the code B bits for each symbol and received using the technique described below. Two linear (independent) equations are then available to calculate the two unknown for the real and imaginary parts of the symbol and received. The values of s ±? and s ± / 0 for the two code bits for the symbol and received can be determined as follows. First, equation (10) indicates that the sign for each bit-to-code LLR is determined by the hypothetical modulation symbol that is closest to y / h. For example, if s ± f X is closer to y / h than s ±? 0, then I (y / h) - s ± / \ 2 will be less than I (y / h) - s ±? 0 \, Y LLR ± will be a negative value. On the contrary, if s ± / 0 is closer to y / h than s ±? , then LLR ± will be a positive value. Changing this fact around, the signs of the LLR of the bit in code B (firm decision of bit), determine the point of signal sc (firm decision of symbol) that is closer to y / h. For example, if LLRX = + a, LLR2 = + b, and LLR3 = -c for an 8-PSK symbol, where a, b, and c are all positive values, then the closest signal point to this 8-PSK It has a label of? 001 '. To simplify the notation, the nearest sc signal point can be re-tagged with an all-zero tag to perform an exclusive OR (XOR) on the label of each signal point in the constellation with the tag of the signal point plus near. In this new notation, the hypothetical symbol s ± 0 for each bit in code B is equal to the nearest signal point, sc, or s ±? 0 = sc for i = 1 ... B. The procedure for determining s ± t X depends on the signal constellation and is specifically described below for the correlated 8-PSK and 16-QAM of Gray. For the 8-PSK, the eight signal points in the constellation are uniformly separated by 45 ° in a unit circle. The signal point closest to y / h is labeled "OOO" from the procedure described above. For the 8-PSK constellation, the next two signal points closest to and / h are the two signal points neighboring '000' (ie one signal point to the left and another signal point to the right of ? 000 'along the unit circle). Because the constellation is correlated according to Gray, these two neighboring signal points differ from? 000 'in only one bit position. For example, if the two neighboring signal points are labeled '100' and '010', then s ± is? 100 'for the code bit further to the left and is? 010' for the central code bit. , the values of s ± / X and s ± / 0 for two of the three code bits are known and can be used together with the LLRs for these two code bits and the gain estimate h of the channel to resolve the symbol and received. For 16-QAM, the 16 signal points are arranged in a two-dimensional grid, and each signal point has at least one neighboring signal point along the real axis and at least one neighboring signal point along the axis imaginary Because the constellation is correlated according to Gray, these neighboring signal points differ from the original signal point in at most one bit position The nearest signal point a / i2 is labeled 0000 'from the procedure described previously, if the two neighboring signal points of the sc point of s Nearest signal are labeled as? 1000 'and 0001', then s ± X is' 1000 'for the code bit further to the left and is 0001' for the code bit further to the right. In this way, the values of s ±? and s >; 0 for two of the four code bits and can be used to solve the symbol and received. The use of a horizontal neighbor and a vertical neighbor avoids a situation with dependent equations. FIGURE 6 shows an RX processor 170d for the second data detection scheme with the base current modulated with a higher order modulation scheme. The RX processor 17Od includes most of the units in the RX processor 170a in FIGURE 3 (without the buffer 314) and additionally includes an estimator 326 of received symbols. The LLR calculation unit 320 of the base current derives the LLRs from the base current based on the symbols. { Y} received and provides the LLRs of the base current through a multiplexer 322 to the buffer 324 for storage. The estimator 326 of received symbols receives the LLRs of the base current from the buffer 324 and derives the estimates. { Y} of symbols received based on these LLR, for example, as described above. The adder 332 receives and subtracts the estimates. { ib} of interference from the estimates. { Y} of received symbols and provides symbols. { ye } Canceled by interference. The LLR calculation unit 340 of the improved current derives the LLR from the improved current based on the symbols. { ye } canceled by interference and provides the LLR of the enhanced current through a multiplexer 322 to a buffer 324 for storage. For clarity, the separate LLR calculation units are shown for the base current and the improved current in FIGURES 3, 5, and 6. The calculation of the LLR for both currents can be performed by a single LLR calculation unit, for example, in the form of time division multiplexing (TDM). All calculations for data detection can also be performed with a digital signal processor (DSP) that has one or more multiplication-accumulation units and one or more arithmetic logic units (ALUs). The block diagrams shown in FIGURES 3, 4, 5, and 6 can also be used as flowcharts for data detection processes. The data detection techniques described herein can be used for both single carrier and multiple carrier systems. Multiple carriers can be provided by OFDM or some other construction. The OFDM effectively divides the complete broadband system into multiple (N) orthogonal subbands, which are also referred to as tones, subcarriers, deposits, and frequency channels. With the OFDM, each subband is associated with a respective subcarrier that can be modulated with data. A combined x symbol can be transmitted on each subband used for data transmission. Up to N combined symbols can be transmitted in the N subbands in each symbol period of the ODFM. The transmitter performs OFDM modulation by transforming each group of N combined and pilot symbols. { x (k)} , which will be transmitted in a period of symbols of the OFDM, to the time domain using a point N of the fast inverse Fourier transform (IFFT) to obtain a "transformed" symbol containing N chips. To combat intersymbol interference (ISI), which is due to selective frequency fading, a portion (or Ncp chips) of each transformed symbol is typically repeated to form a corresponding OFDM symbol. Each symbol of the OFDM is transmitted in a symbol period of the OFDM, which is N + Ncp chip periods, where Ncp is the length of the cyclic prefix. The receiver obtains a stream of samples for a received signal and removes the cyclic prefix in each symbol received from the OFDM to obtain a corresponding received transformed symbol. The receiver then transforms each received transformed symbol into the frequency domain using a N point of the fast Fourier transform (FFT) to obtain N symbols. { and (k)} received for the N subbands. Each symbol y (k) received is for a combined x (k) symbol or a pilot symbol sent over the subband k, which is distorted by a gain h (k) of the channel and degraded by noise n (k), as shown in equation (2). The received symbols can be serialized and processed as described above for the three data detection schemes. The data detection techniques described herein may also be used for more than two data streams. The processing (eg calculation of the LLR, symbol estimate, interference estimate, etc.) used for the enhanced current can be repeated for each additional data stream.

The data detection techniques described herein can be implemented by various means. For example, these techniques can be implemented in hardware, software, or a combination thereof. For a hardware implementation, the in-process units used to perform the data detection can be implemented within one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (the DSPDs). ), logic programming devices (PLDs), programmable field gate arrangements (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof. For a software implementation, data detection techniques may be implemented with modules (eg, procedures, functions, etc.) that perform the functions described herein. The software codes can be stored in a memory unit (e.g., memory unit 192 in FIGURE 1) and executed by a processor (e.g., controller 190). The memory unit may be implemented within a processor or external to the processor, in which cit may be communicatively coupled to the processor through various means as is known in the art. The foregoing descriptions of the described embodiments are provided to enable any person skilled in the art to make or use the present invention. Various modifications of these embodiments will become readily apparent to those skilled in the art, and the generic principles defined therein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but should be in accordance with the broadest scope consistent with the principles and novel features described herein.

Claims (34)

1. NOVELTY OF THE INVENTION Having described the present invention, it is considered as a novelty and therefore the property described in the following claims is claimed as property. CLAIMS 1. A method for performing data detection in a wireless communication system the method is characterized in that it comprises; derive logarithmic probability relationships (LLR) for code bits of a first data stream based on symbols received for a data transmission; estimate the interference due to the first data stream; and deriving the LLRs for the code bits of a second data stream based on the LLRs for the code bits for the first data stream and the estimated interference.
2. 2. The method of compliance with the claim 1, further characterized in that it comprises: decoding the LLRs for the code bits of the first data stream to obtain decoded data for the first data stream; and recoding and remodulating the decoded data to obtain remodulated symbols for the first data stream, wherein the interference due to the first data stream is estimated based on the demodulated symbols.
3. 3. The method of compliance with the claim 1, characterized in that the LLRs for the code bits of the first data stream are derived from the received symbols in real time without buffering the received symbols.
4. 4. The method of compliance with the claim 1, further characterized in that it comprises: storing the LLRs for the code bits of the first data stream in a buffer memory; and storing the LLRs for the code bits of the second data stream in the buffer memory by overwriting the LLRs for the code bits of the first data stream.
5. 5. The method according to claim 1, characterized in that the quadrature phase shift modulation (QPSK) is used for both the first and the second data stream.
6. 6. The method according to claim 1, characterized in that the modulation scheme with a higher order than quadrature phase shift modulation (QPSK) is used for the first data stream, the method further comprising: deriving estimates from the received symbol based on the LLRs for the code bits of the first data stream, and wherein the LLRs for the code bits of the second data stream are derived based on the estimates of received symbols and the estimated interference.
7. The method according to claim 6, characterized in that the derived received symbol estimates includes forming two equations for each received symbol based on the LLR for all the code bits of a data symbol carried in the received symbol for the first one. data stream, and where a symbol estimate received for the received symbol is derived from the two equations.
8. The method according to claim 1, characterized in that the LLRs for the code bits of the first and second data streams are derived based on a maximum dual approach.
9. The method according to claim 1, further characterized by comprising: deriving the channel gain estimates for a wireless channel used for data transmission, and wherein the LLRs for the code bits of the first and second streams Data and interference due to the first data stream are derived with the channel gain estimates.
10. The method according to claim 1, characterized in that the first data stream is a base current and the second data stream is an improved stream for a transmission of random coded data.
11. The method according to claim 1, characterized in that the wireless communication system uses orthogonal frequency division multiplexing (OFDM), and wherein the received symbols are from a plurality of subbands.
12. 12. An apparatus in a wireless communication system, characterized in that it comprises; a first operational calculation unit to derive the logarithmic probability relationships (the LLR) for code bits of a first data stream based on the symbols received for a data transmission; an operational interference estimator to estimate the interference due to the first data stream; and a second operational calculation unit for deriving the LLRs for the code bits of a second data stream based on the LLRs for the code bits of the first data stream and the estimated interference.
13. The apparatus according to claim 12, further characterized in that it comprises: an operational decoder for decoding the LLRs for the code bits of the first data stream to obtain decoded data of the first data stream; and an encoder and modulator operable to recode and remodulate the decoded data to obtain remodulated symbols for the first data stream, and wherein the interference estimator is operative to estimate the interference due to the first data stream based on the demodulated symbols.
14. The apparatus according to claim 12, further characterized in that it comprises: an operational buffer for storing the LLRs for the code bits of the first data stream and storing the LLRs for the code bits of the second data stream by overwriting the LLRs for the code bits of the first data stream.
15. The apparatus according to claim 12, further characterized in that it comprises: an operational estimator channel for deriving channel gain estimates for a wireless channel used for data transmission, and wherein the LLRs for the code bits of the First and second data stream and interference due to the first data stream are derived with the channel gain estimate.
16. 16. An apparatus in a wireless communication system, characterized in that it comprises: means for deriving the logarithmic probability ratios (LLRs) for the code bits of a first data stream based on the symbols received for a data transmission; means for estimating the interference due to the first data stream; and means for deriving the LLRs for the code bits of a second data stream based on the LLR for the code bits of a first data stream and the estimated interference.
17. The apparatus according to claim 16, further characterized in that it comprises: means for decoding the LLRs for the code bits of the first data stream to obtain decoded data for the first data stream; and means for recoding and remodulating the decoded data to obtain remodulated symbols for the first data stream, wherein the interference due to the first data stream is estimated based on the demodulated symbols.
18. 18. The apparatus according to claim 16, characterized in that the LLRs for the code bits of the first data stream are derived from the received symbols in real time without buffering the received symbols.
19. 19. The compliance apparatus with claim 16, further characterized in that it comprises: means for storing the LLRs for the code bits of the first and second data stream, wherein the LLRs for the code bits of the second stream of Data is stored by overwriting the LLRs for the code bits of the first data stream.
20. 20. A method for performing data detection in a wireless communication system, characterized in that it comprises: deriving logarithmic probability relationships (LLRs) for code bits of a first base current based on the symbols received for a data transmission; derive the data symbol estimates for the first data stream based on either the received symbols or the LLRs for the code bits of the first data stream; estimate the interference due to the first data stream based on the data symbol estimates; and deriving the LLRs for the code bits of a second data stream based on the received symbols and the estimated interference.
21. The method according to claim 20, characterized in that the data symbol estimates are derived by making firm decisions either on the received symbols or on the LLRs for the code bits of the first data stream.
22. The method according to claim 20, further characterized in that it comprises: decoding the LLRs for the code bits of a first data stream to obtain decoded data for the first data stream; recoding and remodulating the decoded data to obtain remodulated symbols for the first data stream; and adjusting the LLRs for the code bits of the second data stream based on the demodulated symbols and the data symbol estimates for the first data stream.
23. 23. The method according to claim 22, characterized in that the adjustment of the LLRs includes detecting errors in the data symbol estimates based on the demodulated symbols; and adjusting the LLRs for the code bits of the data symbol estimates detected in error for cancellations for decoding.
24. The method according to claim 22, characterized in that the adjustment of the LLRs includes detecting errors in the data symbol estimates based on the demodulated symbols, deriving correction factors for the estimates of data symbols detected in error, and update the LLRs for the code bits of the data symbol estimates detected in error with the correction factors.
25. 25. The method according to claim 20, characterized in that the LLRs of the code bits of the first and second data streams are derived from the received symbols in real time without buffering the received symbols.
26. 26. The method according to claim 20, further characterized in that it comprises: buffering the LLRs for the code bits of the first and second data streams for subsequent decoding.
27. 27. An apparatus in a wireless communication system, characterized in that it comprises; a first operational calculation unit to derive the logarithmic probability relationships (the LLR) for code bits of a first data stream based on the symbols received for a data transmission; an operational decision unit for deriving the data symbol estimates for the first data stream based on the received symbols; an operational interference estimator for estimating the interference due to the first data stream based on the data symbol estimates; and a second operational calculation unit for deriving the LLRs for the code bits of a second data stream based on the received symbols and the estimated interference.
28. 28. The apparatus according to claim 27, further characterized in that it comprises: an operating decoder for decoding the LLR of the code bits of the first data stream to obtain decoded data for the first data stream; an encoder and modulator operative to recode and remodulate the decoded data to obtain remodulated symbols for the first data stream; and an operational adjustment unit to adjust the LLR for the code bits of the second data stream based on the demodulated symbols and the data symbol estimates for the first data stream.
29. 29. The apparatus according to claim 28, further characterized in that it comprises: an operating symbol error detector for detecting errors in the data symbol estimates based on the demodulated symbols, and wherein the adjustment unit is operative to adjust LLRs for the code bits of the data symbol estimates detected in error for cancellations for decoding.
30. 30. The apparatus according to claim 28, further characterized in that it comprises: an operating symbol error detector for detecting errors in the data symbol estimates based on the demodulated symbols, and wherein the adjustment unit is operative to derive correction factors for the estimates of data symbols detected in error and for updating the LLRs for the code bits of the estimates of the data symbols detected in error with the correction factors.
31. 31. An apparatus in a wireless communication system, characterized in that it comprises: means for deriving the logarithmic probability ratios (LLRs) for the code bits of a first data stream based on the symbols received for a data transmission; means for deriving estimates of data symbols for the first data stream based on the received symbols; means for estimating the interference due to the first data stream based on the symbol estimates; and means for deriving the LLRs for the code bits of a second data stream based on the received symbols and the estimated interference.
32. 32. The apparatus according to claim 31, further characterized in that it comprises: means for decoding the LLRs for the code bits of the first data stream to obtain decoded data for the first data stream; means for recoding and remodulating the decoded data to obtain remodulated symbols for the first data stream; and means for adjusting the LLRs for the code bits of the second data stream based on the demodulated symbols and the data symbol estimates for the first data stream.
33. The apparatus according to claim 32, characterized in that the means for adjusting the LLRs includes means for detecting errors in the data symbol estimates based on the demodulated symbols, and means for adjusting the LLRs for the code bits of the Estimates of data symbols detected in error for cancellations for decoding.
34. 34. The apparatus according to claim 32, characterized in that the means for adjusting the LLRs includes means for detecting errors in the data symbol estimates based on the remodulated symbols, means for deriving the correction factors for the symbol estimations of data detected in error, and means for updating the LLRs for the code bits of the data symbol estimates detected in error with the correction factors.