TW561711B - RAKE receiver with embedded decision feedback equalizer - Google Patents

Abstract

The performance of a RAKE receiver for indoor multipath WLAN applications on direct sequence spread spectrum signals having relatively short codeword lengths is enhanced by embedding a decision feedback equalizer structure in the signal processing path through the receiver's channel matched filter and codeword correlator. The decision feedback equalizer serves to cancel both inter-codeword interference (ISI) or \"bleed-over\" between codewords, and intra-codeword chip interference (ICI) or smearing of the energy within the chips of a respective codeword.

Description

561711 A7 __ ___ B7 V. Description of the Invention (i) Related Applications This patent application was filed on June 29, 1999, and is a continuation of US Application Serial No. 09 / 342,583 entitled "Rake Receiver With Embedded Decision Feedback Equalizer" Applications are hereby incorporated by reference. Field of the Invention: The present invention is generally related to wireless communication systems, such as, but not limited to, wireless local area network (WLAN), and especially directly to the newly improved raKE Receiver related 'The receiver includes an embedded decision feedback equalizer (DFE), which can increase the tolerance to (indoor WLAN) multipath distortion without losing the durability to thermal noise. BACKGROUND OF THE INVENTION: The current demand for faster (higher data rate) wireless communication products is the goal of many proposals before the IEEE 802.11 committee that uses the new standard for the 2.4 GHz portion of the spectrum. FCC Part 15.247 requires the use of spread spectrum technology to implement the new standard. Allows internal packet data rates to exceed 10 Mbps Ethernet speed. Current 802.11 standards use frequency hopping (FH) or direct sequence (DS) Frequency (SS) technology currently only covers one or two μ pbs data rates. The FCC requires the use of spread-spectrum signals derived from the advantages of the original SS characteristics, and reduces the average transmission power space density to make the signal more tolerant of careless interference, and Through the use of spatially redundant receiver technology and thereby counteracting self-interference caused by multipath distortion. As shown in Figure 1, the power delay diagram (PDF) of the transmitted signal 10 will be due to multipath distortion in the indoor WLAN system ( Such as the complexity reduction illustrated in Figure 2 _-4 -___ This paper size applies the Chinese National Standard (CNS) A4 specification (210X297 mm) 5. Description of the invention (2 low example), and shows a large number of power reduction points The old and young attenuation characteristics. The actual situation of the inward transmission environment driving this behavior is that there are a large number of reflectors (such as walls) between the transmitter position 1 4 / and the receiving position 丨 5 in the building, as shown by 疋Nodes 1 2 and 1 j, and the propagation losses associated with individual later arrival time propagation paths k 11, h, and h, include energy that exhibits a logarithmic decrease. The power delay diagram of a signal is related to its power as a function of The average power level of the signal dissipates from time to time, and the average power level of the signal constitutes the inconsistency of its corresponding component. The main principle aspect of the power reduction multipath effect is that, due to the fact that the propagation delay of the signal and the total distance traveled Proportionally, so on average, the strongest (with the fewest obstacles) transmission path is the path where the signal reaches the receiver only. In the known random occurrence process, the first from transmitter position 14 to receiver position 15 Reaching, direct, or sight paths may encounter attenuating media (such as one or more building walls, etc.), and later arriving signals that are reflected by highly reflective surfaces and do not encounter attenuating media may be better than the first arriving signal Even greater impulse response (CIR). On average, however, this rarely happens for the number of echo signals following a CIR peak. In terms of practical application, the root mean square (RMS) delay of the multi-path channel of the small office and the home office (SOHO) ring extension range is 20-50ns, while the business environment is 5 (MO). nsec, and the factory environment is 1⑼ · 200nsec. For reduced-power attenuation channels, krms delay expansion such as reduced-power decay constant. For relatively low signal bandwidth (less than i MHz), multipath caused Attenuation is usually "very average". However, when the bandwidth is greater than 1 ^ 7 (for example, a 10 MHz bandwidth required by a DSSS system to maintain the higher data rate of 10 Mbps described above) It will vary according to the frequency, and the Chinese National Standard (CNS) 8-4 specification (21 × 297 mm) is applied to the paper size. 3 V. Description of the invention (Production ... Reliable communication constitutes a serious obstacle. In this way, the path to the second night Distortion can cause severe propagation loss in the chirp band. The preferred mechanism for this severe frequency selection multipath distortion problem is = positive receiver .: Commonly known as a "rake" receiver. In order to be successful, KE receives read Operation, you must use the transmission frequency Greater than; m frequency _SS structure. In the Dsss signal structure, individual code words are formed by a series of PN code "chip codes". Here, the word "shima" is used instead of the "symbol" to avoid "There is a confusion between the L and Ma characters. The chip code can be transmitted using a fairly simple modulation rule like 疋 QPSK & and the chip chip code can be set in the order of the marks, or it can be pseudo-random. Word phase modulation can be used to convey information. In other words, if additional code bits are added to each codeword, the phase of the codeword is shifted. For example, two extra bits can be used to provide a quadrature (90 degree) phase Transfer increase: 0, 90, 180, and 270 ... The chip code of the codeword can be selected from the multicodeword set, and the M bit can select a specific codeword from the N codewords constituting the multicodeword set. An example of the rule is the use of Walsh * HadamarcU4 for the codeword set. For the 2.4 01 ^ spectrum of the above-mentioned reference, 8 £ 8,021 The Standards Committee proposes the use of an 8-bit coding rule, where the position bits are from N = Choose one of 6 4 multi-chip code words, and the remaining two bits define One of the four possible (orthogonal) phases of the codeword. As shown in the illustration in Figure 3, the received (extended) signal 'in the channel matching correction or rake receiver' will be consumed by the codeword corrector 31, The output (shown as a sequence of time-of-arrival pulses 3 2-1, 32-2, 32-3) is supplied to a continuous multipath combiner 3 3. The codeword corrector 3 1 contains multiple corrections. a Home Standard (CNS) A4 Specification (210X 297 Male I) 561711 A7

561711 A7 B7

V. Description of the invention (5 Codewords with automatic pulse correction function and zero-crossing correction function. In addition, for important multipaths, in order to minimize the degradation caused by intermediate codeword interference, the codeword length must be very large (For example, in the order of 128, 256, or more, such as for military use.) However, in a commercial environment, the number of chips per mark must be limited in order to increase the available data bandwidth. Because when the number of chips per codeword decreases, the content of codeword overflow will increase, so multipath distortion will be obvious, so very small codeword chip density can cause codeword energy overflow / Too many path codewords are leaking. Therefore, the question is how to optimize the output noise ratio of a RAKE receiver using such a relatively poor ideal codeword. SUMMARY OF THE INVENTION According to the present invention, the embedded RAKE receiver architecture can be successful Point out the problem, where the receiver includes a chip-coded decision feedback equalizer embedded in the signal processing path to the receiver channel matching waver and codeword corrector (D FE) structure. This decision feedback equalizer is used to reduce or eliminate the distortion that two restricted chip code length codewords will suffer when using multipath channel transmission. The first is the individual codeword CWi and other codewords cwi The energy in + j "flows out" or leaks. The second kind of distortion is the "fouling" of the energy in the chip code of the individual codeword. The final table feedback network adapter is particularly suitable for combating indoor multipath distortion in WLAN. Because this multipath distortion is mainly the smallest phase, the strongest signal components almost always arrive earliest, while the weaker components arrive more slowly. Therefore, 'Most multipath distortions show a "tail" on the channel pulse response. Degradation. In addition, the feedback tapping of DFE can ideally be used to resist the Chinese paper standard (CNS) A4 specification (210 X 297 public love) of this paper size. The forward tap is used to oppose the maximum phase component. As a result, it is used to counter indoor multipath £) {7 £; almost no positive tap is required, and most of the four processes are performed within the feedback tap. It is expected that the implementation of DFE forward transfer taps at the base frequency requires a full-composite multiplexer, and when QpsK components are used, the feedback taps only require composite addition and subtraction, and implementation is not a very important issue. Although embedding a DFE into a RAKE receiver can easily combat indoor multipath, it makes decisions on a single chip and therefore requires a relatively high snr. Feedback tapping is in addition to long degraded multi-path echo tails and can span multi-path codewords, thus eliminating intermediate chip code fouling and intermediate codeword overflow. The codeword modifier uses a soft decision chip code that continuously combines codewords to improve the noise ratio. Although the DFE chip code decision error is made, the codeword corrector can still use chip codes that continuously combine all codewords to make the correct decision. For the reduction of the voice ratio, the wrong propagation of the equalizer of decision-making will result in cluster errors in the cluster. If the decision of the soft-chip code used to set the equalizer tap coefficient is incorrect, the DFE * will be weighted. The characteristics will deteriorate rapidly, avoiding multipath distortion compensation. To handle low SNR, all codeword chips received are tested before making a hard decision. This can be achieved by generating D F E-type modified detection statistics for each transmitted code. Performing feedback equalization by performing codeword chip code generation can generate predicate statistics for potentially transmitted shima words, assuming that the codeword has actually been transmitted. "The DFE embedded signal processing architecture for canceling intermediate codeword interference can" round "the codeword modifier. In this architecture, the channel matching filter will be combined with the post-cursor representative echo difference generated when the channel pulse response is eliminated. V. Description of the invention (from 7 j onwards #results will produce a "clean" copy of the received codeword, which codeword is transferred to the codeword modifier of the RAKE receiver. The output of this modifier will be coupled to the codeword decision operator, The codeword decision operator will check that there are 7 chip codes in the received codeword to make a decision as to which codeword has actually been transmitted. The decision of the codeword, come & make a copy of the chip valley and the transmission code The phase information of the word is then combined with the channel impulse response evaluation implemented in the FIR filter in order to generate a representation of the multi-path echo of the post-cursor in the signal received by the channel match m. Right to cancel individual codewords The intermediate code word chip code of the energy in the code is defaced. The signal processing branch of each codeword corrector determines the history of all chip code contents and each different group of codes that make up each received codeword. Along with the difference between feedback and knife connection, this means that the specific codeword of the path transmitted from the transmitter through the multi-channel channel will experience the multi-path loss of the cursor after this. The upstream multi-path distortion can be removed upstream or downstream. ′ In the upstream embodiment, each successively received codeword chip set is coupled to a plurality of codeword modifier statistical branches, and each branch is Accompanying different codewords. For an unlimited example using an octet field, it has a K / 256 codeword combination (containing 26 = 64 codewords, each QpSK code has 22 = 4 possible quadrature phases (0., 90., 18 .., 27.)). Within the branch of the individual codeword modifier, the received signal path will be combined with the output difference of the FIR filter feedback tap set to synthesize its individual code. The multi-path channel impulse response of the chip code set (for example, in this example, each code is composed of eight chip codes), which can generate a representative multi-path echo of the code word chip code set. -This paper size applies to China® Home Standard (CNS) A4 (210X 297 mm) 561711 V. Description of the invention (8 The synthesized post-cursor multi-path echo is subtracted from the received codeword, and the input to the accompanying codec modifier is a valid "clear" version of the received codeword chip set. It has removed the multi-path chip code fouling. The output of each codeword corrector is combined with the peak integrator, which will select the largest output as the transmission codeword. There is a more effective rule to synthesize The multipath channel impulse response is to subtract individual fir feedback filter tap stages from the signal received downstream of each word correction n. To simplify complexity, the code word corrector can be used as a fast Walsh (Hadamard) The structure is implemented, because the combination of different received signal processing paths and feedback taps is a post-correction operation, so it is not necessary to re-generate the feedback taps every time when the received codeword chip stone horse timing enters the corrector. . #This will reduce implementation complexity by allowing the remaining functionalities of the synthetic tap path to be stored in a lookup table. Brief description of the drawings: Figure 1 shows a power delay diagram with multipath distortion of an indoor WLAN system; Figure 2 illustrates a simplified example of an indoor WLAN system with a plurality of reflectors between a transmitter position and a receiver position; Figure 3 illustrates a conventional RAKE receiver; Figure 4 shows the MRAKE receiver, in which the sequence of operations performed by the channel matching waver (coherent multipath combiner) and the codeword corrector becomes reversed; Figure 5 shows the impact of transmitting too many paths on the WLAN channel Limit code length codeword

Sequence part; X Figure 6 illustrates {1 八 仄 having a decision feedback equalizer embedded in the signal processing path of the invention description (9) in the signal processing path to the coherent multipath combiner waver and codeword corrector 561711 A7 ______B7 £ Receiver; Figure 7 illustrates the decision feedback equalizer; Figure 8 illustrates the RAKE receiver of Figure 6, which includes the decision feedback equalizer structure of Figure 7; Figure 9 shows the multipath pulse response characteristics; Figure 1 0 illustrates the DFE embedded signal processing architecture for canceling multi-path intermediate codeword interference (丨 s 丨); Figure 11 illustrates the traditional codeword corrector for RAKE receivers; Figure 1 2 # Ming for removing codewords Signal processing architecture for post-cursor multipath distortion upstream of the corrector; and FIG. 13 illustrates a signal processing architecture for removing post-cursor multipath distortion downstream of the codeword corrector. FIG. 14 illustrates a more general embodiment of the signal processing architecture of FIG. 13. Detailed description of the invention: Before explaining the new and improved embedded DFE RAKE receiver architecture of the present invention in detail, we can observe that the receiver architecture of the present invention mainly uses traditional digital communication circuits and modulators with digital signal processing elements. The configuration, and the accompanying supervisory control circuits, control the operation of such circuits and components. I can also understand that the present invention is applicable to direct sequence spread spectrum (DSSS) signals with any codeword modulation, including complementary codes. In the implementation of actually incorporating a printed circuit card of an existing wireless communication device, the present invention can be implemented as a field programmable gate array (FPGA) or an application-determined integrated circuit (ASIC) chipset. Therefore, the configuration structure of this circuit and components and its communication with other -12-I paper standards are applicable to China National Standards (CNS) A4 specifications (210 X 297 mm) ~ '— 561711 A7 B7 V. Description of the invention (1 〇 The interface of the device is described in an easily understandable block diagram (for the most part), which only shows the part of the invention in detail, so it is not ambiguous, and it can be easily understood by those skilled in the art. The benefits are described here. Therefore, the block diagram will show the main components of the present invention in a traditional functional grouping, so that the present invention can be more easily understood. In order to understand the improvements provided by the DFE embedded RAKE receiver of the present invention, It is easy to illustrate the distortion effect of the direct chip spread spectrum codeword with finite chip length encountered in the present invention. Fig. 5 shows the direct chip spread spectrum codewords (CW) with limited chip code length (CW) that transmit too many paths WLAN channels .., 50 -1, 50-2, 50-3, ..... During the transmission, the chip code transmitted by the channel will produce two kinds of distortion ... The intermediate codeword overflows and the internal codeword chip code is dirty. The former Knot If the result is shown as 51, it is the energy “outflow” or leakage in the individual codewords CWi and other codewords cwi + j. The second distortion form is the “fouling” of the energy in the chip code of the individual codeword. The solution $, as shown in Figure 6, is to use chip-type decision feedback on the signal processing path between the coherent multi-path combiner (channel matching filter) 3 3 and the code word corrector 3] [ Modifier (DFE) 36 to modify the Rake receiver. The basic configuration of the decision feedback equalizer is shown in Figure 7 (including the structure of a finite impulse response (FIR) filter with multiple tapped delay line settings), including a plurality of Positive-feed tap 71, the signal transmission path is coupled to the first input 72 of the differential coupler 73 and has a first input 74 coupled to receive the output of a set of feedback taps. The output of the differential coupler 73 is coupled To decision unit 76, and the output of this unit is transferred to feedback tap 75. Figure 8 illustrates the paperless standard of the financial system in Figure 6® S Home Standard (CNS) A4 specifications (〖1G X 29 ^ 董 Γ-- ------- 561711

RAKE receiving state, the receiver incorporates the decision feedback equalizer structure of FIG. Before the soft, equalized chip code level is supplied to the codeword corrector 31, the f architecture at this signal can be operated to DFE equalize the chip code. For effective signal processing, the channel-matching filter 33 and forward tap section 7 in the front end of the DFE can be implemented in a composite structure, which acts as a "whitening" matched filter. This is indicated by the dashed line 77. There are two main factors supporting the decision feedback equalizer that is particularly suitable for combating indoor multipath. First, the multipath distortion is mainly the smallest phase, because the strongest signal component arrives first, and the weaker component arrives later. As shown in the figure, please refer to the exponential power drop characteristic of Figure 丨. The (multipath) component arriving later will be quite weak, so most multipath distortion will appear on the channel pulse response like "tail" attenuation. Secondly, the feedback tap of the DFE is ideally used to counter the minimum phase multipath component 'and the positive tap is used to counter the maximum phase component. As a result, DFEs used to counter indoor multipathing require almost no forward-feeding taps, and most processing is performed within feedback taps. Because the implementation of DFE on the fundamental frequency requires a full compound multiplexer, and when using qPSK components, the feedback tap only requires compound addition and subtraction, so for practical applications, it is better to use the main feedback branch.接 的 应用。 Implementation. Figure 9 shows the multipath pulse response characteristics of a maximum pulse response component. Usually the decision unit of the DFE is focused on the peak value 91 of the impulse response, as it is a single component with the largest noise ratio. The feedback tap is subtracted from the "tail" 9 2 immediately after the peak 9 1, and the forward tap cancels the front-end pulse component 93 before the peak. In other words, in traditional DFE, cancel (middle symbol)

561711

Disturbance involves forcing the energy beyond the peak of the pulse to zero. When only using the feedback knife connection, there is no noise amplification during the equalization period, so the above-mentioned post multiplexer is not needed. As can be seen from Figure 1, this is almost applicable to indoor wireless channels. Although the integration of the decision feedback equalizer into the RAKE receiver is a particularly useful mechanism against indoor multipath at this moment, it is still determined by the chip code, so it is necessary to have

A relatively high SNR (e.g., 10 dB or more) is required for successful operation. The dfe feedback tap 75, except for the long degenerate multi-path tail 92, can span multi-path codewords, thus eliminating the above-mentioned intermediate chip code defacement and intermediate codeword overflow. The codeword corrector 31 uses a soft decision chip code that continuously combines codewords to provide installation. Even if a DFE chip code decision is made wrong, but the codeword is modified, you can still use the chip code that continuously combines all codewords to do Make the right decisions. Order

However, in order to reduce the noise ratio, the propagation of errors in the equalizer of decision feedback will cause chip code errors in the cluster. If the decision of the soft chip code used to set the equalizer tap coefficient is incorrect, the entire DFE score will be reduced. The connection weighting characteristics will deteriorate rapidly to avoid multipath distortion compensation. In other words, when the noise level increases, usually the code word decision will be destroyed soon after the chip code decision is destroyed. To address this low SNR problem, all codeword chip codes received are tested before making a hard decision. This can be achieved by generating DFE predicate statistics for each codeword that has been transmitted, rather than for each chip. The use of shout equalization to perform codeword chip code generation can generate a given potentially transmitted ("job") codeword = detection statistics' assuming that the relevant mark has actually been transmitted. As noted above, during transmission, the codewords transmitted over the channel will produce

13 561711 V. Description of the invention (two kinds of distortion: intermediate codeword interference or "overflow" between codewords (in other words, intermediate symbol interference (ISI)), and internal codeword chip interference (Ici) or individual codeword chip codes CWA energy is defaced. The degree of ISI is related to the number of chips per codeword and the range of multipath distortion. For the 05§§ mechanism, as in military applications, the number of chips per codeword is usually There are many (such as 64 1 28, 256 levels or above), and the relative codeword overflow is relatively insignificant. However, as noted above, in a commercial environment, the number of chips per codeword must be limited per page (such as There are only eight characters per word) to maximize the available data bandwidth. Because when the number of chips per codeword decreases, the content of codeword overflow = will increase: so many path distortions will be obvious, so Very small code chip code density does not immediately cause overflow of consecutive codeword boundaries, but passes through multipath codewords. The DFE embedded signal processing architecture shown in Figure 10 for canceling intermediate codeword interference (ISI) can be " Around the codeword modifier, For this purpose, the output of the channel matching amplifier 33 is transferred to the first input ⑻ of the difference combiner 102, and the second input 103 of the combiner is coupled to receive the post-cursor generated by the evaluation channel impulse response. Represents the echo. The output of the difference combiner 1 02 (representing a "clean" copy of the received codeword) is transferred to the codeword modifier 31, and the output of this modifier is supplied to the codeword decision operator 1 () 5 The codeword decision operator 105 checks all M chip codes in the received codeword instead of a single chip code to make a decision as to which codeword has been transmitted.-It is known that the code obtained by the operator 1G5 Word decision, and then synthesize a copy of the chip code content and the phase information determined by the original transmission codeword in the transmission codeword synthesis benefit 106. The synthesized codeword is then compared with the concealed wave filter within -16 paper sizes. Shicai S 8 standards ⑽S) A4 specification (21 () < 297 公 #) 56171! A7 B7 V. The description of the channel pulse response evaluation implemented in the invention description, in order to be received by the channel matched filter 3 3 Multi-path echo . By supplying this post-cursor echo to the difference combiner 102, the total IS in the output of the channel matching filter 33] [contribution will be effectively canceled from the input to the codeword processor 31. Please note that the channel impulse response evaluation implemented in the FIR filter 107 has no (codeword) length limitation; whether it crosses one or more codeword boundaries, its output covers the signal received by the channel matched filter 3 3 The multi-path echo of the entire post cursor within the signal. In addition, for effective signal processing, the channel matching filter and forward tap section of the DFE can be implemented as a "whitening" matching waver, as shown above. The DFE embedded signal processing architecture used to cancel the internal codeword chip contamination of the energy in the individual codeword c Wi chip code will be described with reference to FIGS. As shown in FIG. 11, the codeword corrector 3 1 of the conventional RAKE receiver has a plurality of individual codeword correctors 3 1 -1 -3 1-N, each of which is used to detect different codeword chip codes. set. The octet field of this example defines a combination of 2 5 6 codewords (including 2-64 coders' each with 22 = 4 possible quadrature phases (0., 900, 180 °, 270 °) ). Each codeword in this example contains eight chip codes, so there are a total of 48 = 64K possible chip code combinations. These 64 codeword chip code combinations can be selected. Using a requirement group previously made based on the properties of the communication application (indoor WLAN) and manufacturing a modifier that reduces complexity, a large number of available chip code combinations can be reduced to codes with only the chip chip set structure according to experience at the beginning. Words, like the Walsh or Hadamard structure described above. Even if this selection criterion can be satisfied from a practical point of view, not all selection groups =

15 561711 V. Description of the invention (The codeword will have ideal attributes (especially the above-mentioned better self-correction, intersecting and repairing attributes), so that even if there is no multi-path chip code defacement, the output of the modifier cannot be displayed. Ideal pulse / zero characteristics. Especially its multi-path distortion is less than the ideal chip code set, and each chip code has a very low voice ratio, so more than one code can be expected without the chip code distortion compensation form. The output of the sub-correction is generally indistinguishable. To cope with such chip defacement problems and improve the detection statistics of code word correction, the signal processing branch to each Bima word corrector of the present invention will be set to The content of all the chip codes of each received Shima word is accompanied by the code points of the individual different groups. The difference is combined. This represents the specific word of the multi-path transmitted from the transmission through the multi-path channel. Will experience the multi-path distortion echo of the cursor. Processing the chip code of each received codeword together will increase the SNR of the received signal by 6dB, and improve the detection of the knife of each codeword modifier. The accuracy of the calculation is shown in Figures 丨 2 and 丨 3 for the individual specific embodiments of the cursor multipath distortion effectively subtracted or removed from the upstream and downstream of the corrector output of the codeword modifier. In the upstream implementation of FIG. 12 The receiving number path 1 2 0 1 for the continuous reception of codeword chip code sets (pulse hot news) will be displayed as coupled to the complex k number of codeword correction state statistics branches 1 2 0 3 · 1, ..., 1 2 〇3-Κ, each accompanied by an individual different one of the K codeword combinations. As explained above, in the example of the present invention using an octet field, it has K > 2 5 6 codeword combinations ( Contains 26 = 6 4 code children's mothers each have 2 = 4 possible quadrature phases (0., 90., 180., 270 °)) 〇 Each codeword corrector branch contains a differential combination Device 12 丨 〇, coupling has been 18-561711 A7

561711 A7

561711 A7 B7

V. Description of the invention L 18 1 J 2 1-N will be expanded by the expansion unit 1 3 5 0 into a plurality of κ followed by 2 6 6 codeword combinations (6 4 codewords, each of which is four possible positives Cross-phase (0., 90., 180., 270.) modifier output lines ^^ ", 1 3 5 1-Κ. Each unit N of the received signal path codeword corrector 1j20 determined by the expansion unit 1 3 50 outputs a composite value of 1321-1, ..., 1321-N to identify the four-phase rotation (+1, + j, -1,-j). As shown in the individual four sets of unwrapped output lines 1351-1-135 1-4, will the four-phase rotation output (+1, + j, · 1,-j) of individual codewords be lightly combined to a set of difference combinations? § 1360_1-1360-4 The first (+) input is 1361. These differential couplers have a second input 1362 coupled to receive composite correction values from the synthetic tap path and stored in lookup table memory. As explained above, it is clear that the feedback tap must be regenerated when each newly received codeword chip code set is timed to the corrector. The outputs of the differential combiner 1360-1-1360_K 1 3 6 3 are all coupled to the peak detector 1 3 8 0, which will select the largest true output as the actual transmission codeword. It is clear from the above description that by actually embedding a decision feedback equalizer structure on the channel-matched filter and codeword corrector of the signal processing path to the receiver, this can be enhanced for direct sequence development. Performance of traditional RAKE receivers for indoor WLAN multipath applications on high-frequency signals (with fairly short codeword lengths). This decision feedback equalizer is used to eliminate the intermediate codeword interference or "overflow" between the two codewords, as well as the internal codeword chip interference (ICI) or the energy defacement in individual codeword chips. Figure 14 illustrates a more general embodiment of the downstream implementation of Figure 13. Received signal path 140 1 similar to received signal path 130 1 Coupling to receive path -21-This paper size applies Chinese National Standard (CNS) A4 specification (210 X 297 mm) Binding

Line 561711 A7

V. Description of the Invention (20) The word corrector 1430 compares the output of the individual FIR filter feedback tapping set 4410 with the corresponding codeword 'and generates an evaluation error signal of the corresponding codeword. The evaluation codeword error signals of the mother-to-code children C W -1 to C w _ N are provided to the negative or related inputs of the corresponding difference combiner 1 360. The outputs of the difference combiners 146- to 1460-N are all consumed up to the peak debt measurement. This controller selects the largest true output as the actual transmission codeword in a manner similar to the peak detector 138. Although it has been shown and explained here that according to the specific embodiment of the present invention, we can understand that there are many restrictions and modifications that can be made by those skilled in this art, and we do not want to be limited to The details shown and explained here, but it is intended to cover all similar changes and modifications known to those skilled in the art. ~ 23-This paper is suitable for film I China National Standard (CNS) A4 size (210 X ^ 97 mm)

Claims (1)

561711 A8 B8 C8 Hold 561711 AS B8 σ is one of the second inputs, and each can synthesize the multi-path channel impulse response of the chip code set corresponding to the code word to provide a representative of the post-cursor multi-path echo of the corresponding code word; ^ human T number of code words Correctors, each coupled to the output of one of a plurality of individual difference combiners to detect a corresponding chip code set corresponding to one of the plurality of codewords; and a peak detector 'coupled to each of the plurality of codeword modifiers Output of one. J 3. As for the signal processing architecture in the second item of the patent application scope, each of the finite impulse response filters includes a separate filter and a wave filter feedback tap set. 4. The signal processing architecture of item 3 of the patent application, wherein each of the plurality of filter feedback tap sets contains a weighting coefficient established during the training interval. 5. If the signal processing architecture of item 4 of the patent application scope further includes a lookup table that is consumed by each filter feedback tap set, the table stores the weighting coefficient. 6. —A signal processing architecture for a direct sequence spread spectrum receiver in a multi-path environment, including: a first codeword corrector that uses each of a plurality of predetermined codewords to correct a received signal and outputs a corresponding A plurality of correction signals; a plurality of finite impulse response filters, each corresponding to one of a plurality of predetermined codewords, and each operable to synthesize a multipath channel impulse response of a chip code set corresponding to the codeword to provide Representative of multi-path echo of the back cursor corresponding to the codeword; -25-This paper size applies to China National Standard (CNS) A4 (210 X 297 mm) Hold 561711 A8 B8 C8 Patent application range A plurality of second codeword modifiers, each coupled to the output of one of a plurality of individual finite impulse response filters, and each providing a corresponding one of a plurality of evaluation error signals to the corresponding One of the plurality of codewords; the plurality of differences are combined into one each having a first input for receiving one of the corresponding plurality of correction signals and one for receiving one of the corresponding plurality of estimation error signals A second input; and a peak detector coupled to the output of each of the plurality of difference combiners. 7 The signal processing architecture of item 6 of the patent application, wherein each finite impulse response filter includes a separate filter feedback tap set. 8. The signal processing architecture of item 7 of the patent application, wherein each of the plurality of filter feedback tap sets includes a weighting coefficient established during the training interval. 9. The signal processing architecture of item 8 of the patent application scope further includes a lookup table coupled to the plurality of filter feedback tap sets, the table storing the weighting coefficient. 10. The signal processing architecture according to item 6 of the patent application scope, further comprising: an expansion unit for receiving a plurality of correction signals and generating a set of quadrature phase correction signals for each correction signal; a plurality of mother code first codes The modifiers each provide a corresponding set of orthogonal evaluation error signals to one of the corresponding plurality of codewords; and a plurality of difference combiners, each including a set of orthogonal combiners for receiving a corresponding orthogonal phase correction A signal group and a corresponding orthogonal evaluation error signal group. -26-The paper size is in accordance with Chinese National Standard (CNS) A4 (210 X 297 mm) Installation 561711 8 8 8 8 ABC D 6. Application scope of patent 11. The signal processing architecture of item 6 of the scope of patent application, wherein the first codeword corrector is implemented with a fast Walsh conversion structure. -27-This paper size applies to China National Standard (CNS) A4 (210 X 297 mm)