US20160254933A1

Movatterモバイル変換

Info

Publication number: US20160254933A1
Application number: US15/067,924
Authority: US
Inventors: Amir Eliaz
Original assignee: Magnacom Ltd
Current assignee: Magnacom Ltd
Priority date: 2012-06-20
Filing date: 2016-03-11
Publication date: 2016-09-01
Also published as: WO2013190395A2; US8548072B1; US8599914B1; US8548097B1; EP2865151A4; US20130343487A1; US8582637B1; US9252822B2; US20150071389A1; US8605832B1; US20140108892A1; CN104521203A; EP2865097A4; US20140133540A1; US20140036986A1; US20150312000A1; EP2865151A2; US8559498B1; US8553821B1; US8675769B1

Abstract

A method and system for configuring one or both of a transmitter pulse-shaping filter and a receiver pulse-shaping filter to generate a total partial response that incorporates a predetermined amount of inter-symbol interference (ISI), based on one or more defined performance-related variables and one or more set constraints that are applicable to one or both of the transmitter pulse-shaping filter and the receiver pulse-shaping filters. The predetermined amount of ISI is determined based on an estimation process during extraction of data from an output of the receiver pulse-shaping filter, such that performance of total partial response based communication matches or surpasses performance of communication incorporating filtering based on no or near-zero ISI. The configuring may comprise determining optimized filtering configuration, by applying an optimization process which is based on, at least in part, the one or more constraints and the one or more performance-related variables.

Description

CLAIM OF PRIORITY

This application is a continuation of U.S. patent application Ser. No. 14/551,466 filed Nov. 24, 2014, which is a continuation of U.S. patent application Ser. No. 13/754,998, filed Jan. 31, 2013 (now patented as U.S. Pat. No. 8,897,387), which claims the benefit of priority to U.S. Provisional Patent Application Ser. No. 61/662,085 filed on Jun. 20, 2012, U.S. Provisional Patent Application Ser. No. 61/726,099 filed on Nov. 14, 2012, U.S. Provisional Patent Application Ser. No. 61/729,774 filed on Nov. 26, 2012, U.S. Provisional Patent Application Ser. No. 61/747,132 filed on Dec. 28, 2012.

Each of the above identified applications is hereby incorporated herein by reference in its entirety.

INCORPORATION BY REFERENCE

This application makes reference to:

Each of the above stated applications is hereby incorporated herein by reference in its entirety.

FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[Not Applicable].

MICROFICHE/COPYRIGHT REFERENCE

[Not Applicable].

TECHNICAL FIELD

Aspects of the present application relate to electronic communications. More specifically, certain implementations of the present disclosure relate to design and optimization of partial response pulse shape filter.

BACKGROUND

Existing communications methods and systems are overly power hungry and/or spectrally inefficient. Further limitations and disadvantages of conventional and traditional approaches will become apparent to one of skill in the art, through comparison of such approaches with some aspects of the present method and apparatus set forth in the remainder of this disclosure with reference to the drawings.

BRIEF SUMMARY

A system and/or method is provided for design and optimization of partial response pulse shape filter, substantially as shown in and/or described in connection with at least one of the figures, as set forth more completely in the claims.

These and other advantages, aspects and novel features of the present disclosure, as well as details of illustrated implementation(s) thereof, will be more fully understood from the following description and drawings.

BRIEF DESCRIPTION OF SEVERAL VIEWS OF THE DRAWINGS

FIG. 1 is a block diagram depicting an example system configured for low-complexity, highly-spectrally-efficient communications.

FIG. 2 is a block diagram depicting an example equalization and sequence estimation circuit for use in a system configured for low-complexity, highly-spectrally-efficient communications.

FIG. 3 is a block diagram depicting an example sequence estimation circuit for use in a system configured for low-complexity, highly-spectrally-efficient communications.

FIG. 4 is a block diagram depicting an example partial response pulse-shaping filtering setup for use in a system configured for low-complexity, highly-spectrally-efficient communications.

FIG. 5 is a block diagram depicting an example finite impulse response (FIR) implementation of partial response pulse-shaping filters, in accordance with an embodiment of the present invention.

FIG. 6 is a flow chart depicting an example of a process for joint optimization of transmitter and receiver pulse-shaping filters, in accordance with an embodiment of the present invention.

FIG. 7 is a chart diagram depicting a comparison between the frequency-domain responses of a legacy transmit filter and an optimized partial response pulse-shaping transmit filter.

DETAILED DESCRIPTION

The present disclosure relates to a method and system for design and optimization of partial response pulse shape filter.

As utilized herein the terms “circuits” and “circuitry” refer to physical electronic components (i.e. hardware) and any software and/or firmware (“code”) which may configure the hardware, be executed by the hardware, and or otherwise be associated with the hardware. As used herein, for example, a particular processor and memory may comprise a first “circuit” when executing a first plurality of lines of code and may comprise a second “circuit” when executing a second plurality of lines of code. As utilized herein, “and/or” means any one or more of the items in the list joined by “and/or”. As an example, “x and/or y” means any element of the three-element set {(x), (y), (x, y)}. As another example, “x, y, and/or z” means any element of the seven-element set {(x), (y), (z), (x, y), (x, z), (y, z), (x, y, z)}. As utilized herein, the terms “block” and “module” refer to functions than can be performed by one or more circuits. As utilized herein, the term “exemplary” means serving as a non-limiting example, instance, or illustration. As utilized herein, the terms “for example” and “e.g.,” introduce a list of one or more non-limiting examples, instances, or illustrations. As utilized herein, circuitry is “operable” to perform a function whenever the circuitry comprises the necessary hardware and code (if any is necessary) to perform the function, regardless of whether performance of the function is disabled, or not enabled, by some user-configurable setting.

FIG. 1 is a block diagram depicting an example system configured for low-complexity, highly-spectrally-efficient communications. Thesystem100 comprises amapper circuit102, a pulse shapingfilter circuit104, a timingpilot insertion circuit105, a transmitter front-end circuit106, achannel107, a receiver front-end108, afilter circuit109, a timingpilot removal circuit110, an equalization andsequence estimation circuit112, and ade-mapping circuit114. The

components

102,104,105, and106 may be part of a transmitter (e.g., a base station or access point, a router, a gateway, a mobile device, a server, a computer, a computer peripheral device, a table, a modem, a set-top box, etc.), the

components

108,109,110,112, and114 may be part of a receiver (e.g., a base station or access point, a router, a gateway, a mobile device, a server, a computer, a computer peripheral device, a table, a modem, a set-top box, etc.), and the transmitter and receiver may communicate via thechannel107.

Themapper102 may be operable to map bits of the Tx_bitstream to be transmitted to symbols according to a selected modulation scheme. The symbols may be output viasignal103. For example, for an quadrature amplitude modulation scheme having a symbol alphabet of N (N-QAM), the mapper may map each Log₂(N) bits of the Tx_bitstream to single symbol represented as a complex number and/or as in-phase (I) and quadrature-phase (Q) components. Although N-QAM is used for illustration in this disclosure, aspects of this disclosure are applicable to any modulation scheme (e.g., amplitude shift keying (ASK), phase shift keying (PSK), frequency shift keying (FSK), etc.). Additionally, points of the N-QAM constellation may be regularly spaced (“on-grid”) or irregularly spaced (“off-grid”). Furthermore, the symbol constellation used by the mapper may be optimized for best bit-error rate performance that is related to log-likelihood ratio (LLR) and to optimizing mean mutual information bit (MMIB). The Tx_bitstream may, for example, be the result of bits of data passing through a forward error correction (FEC) encoder and/or an interleaver. Additionally, or alternatively, the symbols out of themapper102 may pass through an interleaver.

Thepulse shaper104 may be operable to adjust the waveform of thesignal103 such that the waveform of the resultingsignal113 complies with the spectral requirements of the channel over which thesignal113 is to be transmitted. The spectral requirements may be referred to as the “spectral mask” and may be established by a regulatory body (e.g., the Federal Communications Commission in the United States or the European Telecommunications Standards Institute) and/or a standards body (e.g., Third Generation Partnership Project) that governs the communication channel(s) and/or standard(s) in use. Thepulse shaper104 may comprise, for example, an infinite impulse response (IIR) and/or a finite impulse response (FIR) filter. The number of taps, or “length,” of thepulse shaper104 is denoted herein as LTx, which is an integer. The impulse response of thepulse shaper104 is denoted herein as hTx. Thepulse shaper104 may be configured such that itsoutput signal113 intentionally has a substantial amount of inter-symbol interference (ISI). Accordingly, thepulse shaper104 may be referred to as a partial response pulse shaping filter, and thesignal113 may be referred to as a partial response signal or as residing in the partial response domain, whereas thesignal103 may be referred to as residing in the symbol domain. The number of taps and/or the values of the tap coefficients of thepulse shaper104 may be designed such that thepulse shaper104 is intentionally non-optimal for additive white Gaussian noise (AWGN) in order to improve tolerance of non-linearity in the signal path. In this regard, thepulse shaper104 may offer superior performance in the presence of non-linearity as compared to, for example, a conventional zero (or near zero) positive ISI pulse shaping filter (e.g., root raised cosine (RRC) pulse shaping filter). Thepulse shaper104 may be designed as described in at least some of the following figures (e.g.,FIGS. 4-7), and in one or more of: the United States patent application titled “Constellation Map Optimization For Highly Spectrally Efficient Communications,” and the United States patent application titled “Dynamic Filter Adjustment For Highly-Spectrally-Efficient Communications,” each of which is incorporated herein by reference, as set forth above.

It should be noted that a partial response signal (or signals in the “partial response domain”) is just one example of a type of signal for which there is correlation among symbols of the signal (referred to herein as “inter-symbol-correlated (ISC) signals”). Such ISC signals are in contrast to zero (or near zero) ISI signals generated by, for example, raised-cosine (RC) or root-raised-cosine (RRC) filtering. For simplicity of illustration, this disclosure focuses on partial response signals generated via partial response filtering. Nevertheless, aspects of this disclosure are applicable to other ISC signals such as, for example, signals generated via matrix multiplication (e.g., lattice coding), and signals generated via decimation below the Nyquist frequency such that aliasing creates correlation between symbols.

The timingpilot insertion circuit105 may insert a pilot signal which may be utilized by the receiver for timing synchronization. Theoutput signal115 of the timingpilot insertion circuit105 may thus comprise thesignal113 plus an inserted pilot signal (e.g., a sine wave at ¼×fbaud, where fbaud is the symbol rate). An example implementation of thepilot insertion circuit105 is described in the United States patent application titled “Timing Synchronization for Reception of Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.

The transmitter front-end106 may be operable to amplify and/or upconvert thesignal115 to generate thesignal116. Thus, the transmitter front-end106 may comprise, for example, a power amplifier and/or a mixer. The front-end may introduce non-linear distortion and/or phase noise (and/or other non-idealities) to thesignal116. The non-linearity of thecircuit106 may be represented as FnITx which may be, for example, a polynomial, or an exponential (e.g., Rapp model). The non-linearity may incorporate memory (e.g., Voltera series).

Thechannel107 may comprise a wired, wireless, and/or optical communication medium. Thesignal116 may propagate through thechannel107 and arrive at the receive front-end108 assignal118.Signal118 may be noisier than signal116 (e.g., as a result of thermal noise in the channel) and may have higher or different ISI than signal116 (e.g., as a result of multi-path).

The receiver front-end108 may be operable to amplify and/or downconvert thesignal118 to generate thesignal119. Thus, the receiver front-end may comprise, for example, a low-noise amplifier and/or a mixer. The receiver front-end may introduce non-linear distortion and/or phase noise to thesignal119. The non-linearity of thecircuit108 may be represented as FnIRx which may be, for example, a polynomial, or an exponential (e.g., Rapp model). The non-linearity may incorporate memory (e.g., Voltera series).

The timing pilot recovery andremoval circuit110 may be operable to lock to the timing pilot signal inserted by thepilot insertion circuit105 in order to recover the symbol timing of the received signal. Theoutput122 may thus comprise thesignal120 minus (i.e. without) the timing pilot signal. An example implementation of the timing pilot recovery andremoval circuit110 is described in the United States patent application titled “Timing Synchronization for Reception of Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.

The input filter109 may be operable to adjust the waveform of the partial response signal119 to generate partial response signal120. The input filter109 may comprise, for example, an infinite impulse response (IIR) and/or a finite impulse response (FIR) filter. The number of taps, or “length,” of the input filter109 is denoted herein as LRx, an integer. The impulse response of the input filter109 is denoted herein as hRx. The number of taps, and/or tap coefficients of the input filter109 may be configured based on: a non-linearity model,
, signal-to-noise ratio (SNR) ofsignal120, the number of taps and/or tap coefficients of the Txpartial response filter104, and/or other parameters. The number of taps and/or the values of the tap coefficients of theinput filter109 may be configured such that noise rejection is intentionally compromised (relative to a perfect match filter) in order to improve performance in the presence of non-linearity. As a result, theinput filter109 may offer superior performance in the presence of non-linearity as compared to, for example, a conventional zero (or near zero) positive ISI matching filter (e.g., root raised cosine (RRC) matched filter). Theinput filter109 may be designed as described in at least some of the following figures (e.g.,FIGS. 4-7), and in one or more of: the United States patent application titled “Constellation Map Optimization For Highly Spectrally Efficient Communications,” and the United States patent application titled “Dynamic Filter Adjustment For Highly-Spectrally-Efficient Communications,” each of which is incorporated herein by reference, as set forth above.
As utilized herein, the “total partial response (h)” may be equal to the convolution of hTx and hRx, and, thus, the “total partial response length (L)” may be equal to LTx+LRx−1. L may, however, be chosen to be less than LTx+LRx−1 where, for example, one or more taps of theTx pulse shaper104 and/or theRx input filter109 are below a determined level. Reducing L may reduce decoding complexity of the sequence estimation. This tradeoff may be optimized during the design of thesystem100. In various embodiments, the design and/or configuration of filters and/or filtering operations of thepulse shaper104 and the input filter109 (particularly, to realize the desired total partial response) may be optimized. The filter(s) used and/or the filtering performed (e.g., at transmit-side and/or receive-side) during partial response pulse shaping may be optimized symmetrically (i.e. the linear phase) and/or asymmetrically (i.e. nonlinear phase). In this regard, in symmetric implementations the tap coefficients may be set symmetrically around zero, assuming the filter is centered around zero (i.e. tap coefficient function, h(k), is an even function that provides linear phase). Optimizing filters/filtering design and/or configuration may comprise, for example, optimizing Symbol Error Rate function of transmitter and receiver partial response taps; optimizing (weighted) minimal distance reduction (e.g., over transmit-side filter only or both transmit-side and receive-side filters); optimizing distance between a spectrum mask imposed on transmission and the response of the transmit-side filter incorporating nonlinear distortion model; optimizing non-compensated (residual) preceding taps of the total response (i.e. both the transmit-side and receive-side filters); optimizing noise enhancement of receiver side; reducing noise enhancement of receiver side in case of frequency selective fading (multipath) using transmit-side and/or receive-side filters; optimizing adjacent and/or interferes signals and noise folding caused by decimation (anti-aliasing) at the receiver; optimizing non-linear tolerance at the receiver originated at Tx side by the overall time domain response; and/or optimizing ‘early’ taps. In some instances, the optimization process may be performed per modulation—i.e. based on the particular modulation scheme (e.g., PSK, QAM, etc.) that is to be used. In some instances, real time optimization of the transmit-side and/or receive-side filters may be performed. In some instances, filtering optimization may comprise optimizing a constellation symbol mapping used for modulation, such as for the smallest minimum distance reduction possible or Symbol Error Rate (SER). In some instances, filtering operations may comprise use of adaptive transmit-side and/or receive-side filters—e.g., for adaptive baud rate vs. minimal distance reduction or Symbol Error Rate (SER), based on particular parameters or criteria, such as link condition parameters, SER, metric function of sequence estimation detection and/or other performance indication measurements (e.g., any parameters that may pertain to receiver ability to recover timing needed for information detection). The adaptive transmit-side and/or receive-side filters may also be configured to account for dynamic channel (fading) compensation. Filter/filtering optimization is described in more details with respect toFIG. 4.
The equalizer andsequence estimator112 may be operable to perform an equalization process and a sequence estimation process. Details of an example implementation of the equalizer andsequence estimator112 are described below with respect toFIG. 2. Theoutput signal132 of the equalizer andsequence estimator112 may be in the symbol domain and may carry estimated values of corresponding transmitted symbols (and/or estimated values of the corresponding transmitted information bits of the Tx_bitstream) ofsignal103. Although not depicted, thesignal132 may pass through an interleaver en route to the de-mapper114. The estimated values may comprise soft-decision estimates, hard-decision estimates, or both.
The de-mapper114 may be operable to map symbols to bit sequences according to a selected modulation scheme. For example, for an N-QAM modulation scheme, the mapper may map each symbol to Log₂(N) bits of the Rx_bitstream. The Rx_bitstream may, for example, be output to a de-interleaver and/or an FEC decoder. Alternatively, or additionally, the de-mapper114 may generate a soft output for each bit, referred as LLR (Log-Likelihood Ratio). The soft output bits may be used by a soft-decoding forward error corrector (e.g. a low-density parity check (LDPC) dedecoder). The soft output bits may be generated using, for example, a Soft Output Viterbi Algorithm (SOVA) or similar. Such algorithms may use additional information of the sequence decoding process including metrics levels of dropped paths and/or estimated bit probabilities for generating the LLR, where
$LLR (b) = \log (\frac{P_{b}}{1 - P_{b}}),$
where P_bis the probability that bit b=1.
In an example implementation, components of the system upstream of thepulse shaper104 in the transmitter and downstream of the equalizer andsequence estimator112 in the receiver may be as found in a conventional N-QAM system. Thus, through modification of the transmit side physical layer and the receive side physical layer, aspects of the invention may be implemented in an otherwise conventional N-QAM system in order to improve performance of the system in the presence of non-linearity as compared, for example, to use of RRC filters and an N-QAM slicer.
FIG. 2 is a block diagram depicting an example equalization and sequence estimation circuit for use in a system configured for low-complexity, highly-spectrally-efficient communications. Shown are anequalizer circuit202, asignal combiner circuit204, a phase adjustcircuit206, asequence estimation circuit210, andnon-linearity modeling circuits236aand236b.
Theequalizer202 may be operable to process thesignal122 to reduce ISI caused by thechannel107. Theoutput222 of theequalizer202 is a partial response domain signal. The ISI of thesignal222 is primarily the result of thepulse shaper104 and the input filter109 (there may be some residual ISI from multipath, for example, due to use of the least means square (LMS) approach in the equalizer202). The error signal,201, fed back to theequalizer202 is also in the partial response domain. Thesignal201 is the difference, calculated bycombiner204, between222 and apartial response signal203 that is output bynon-linearity modeling circuit236a. An example implementation of the equalizer is described in the United States patent application titled “Feed Forward Equalization for Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.
Thecarrier recovery circuit208 may be operable to generate asignal228 based on a phase difference between thesignal222 and apartial response signal207 output by thenon-linearity modeling circuit236b. Thecarrier recovery circuit208 may be as described in the United States patent application titled “Coarse Phase Estimation for Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.
The phase adjustcircuit206 may be operable to adjust the phase of thesignal222 to generate the signal226. The amount and direction of the phase adjustment may be determined by thesignal228 output by thecarrier recovery circuit208. The signal226 is a partial response signal that approximates (up to an equalization error caused by finite length of theequalizer202, a residual phase error not corrected by the phase adjustcircuit206, non-linearities, and/or other non-idealities) the total partial response signal resulting from corresponding symbols ofsignal103 passing throughpulse shaper104 andinput filter109.
Thebuffer212 buffers samples of the signal226 and outputs a plurality of samples of the signal226 viasignal232. Thesignal232 is denoted PR1, where the underlining indicates that it is a vector (in this case each element of the vector corresponds to a sample of a partial response signal). In an example implementation, the length of the vector PR1 may be Q samples.
Input to thesequence estimation circuit210 are thesignal232, thesignal228, and a response ĥ. Response ĥ is based on h (the total partial response, discussed above). For example, response ĥ may represent a compromise between h (described above) and a filter response that compensates for channel non-idealities such as multi-path. The response ĥ may be conveyed and/or stored in the form of LTx+LRx−1 tap coefficients resulting from convolution of the LTx tap coefficients of thepulse shaper104 and the LRx tap coefficients of theinput filter109. Alternatively, response ĥ may be conveyed and/or stored in the form of fewer than LTx+LRx−1 tap coefficients—for example, where one or more taps of the LTx and LRx is ignored due to being below a determined threshold. Thesequence estimation circuit210 may output partial response feedback signals205 and209, asignal234 that corresponds to the finely determined phase error of thesignal120, and signal132 (which carries hard and/or soft estimates of transmitted symbols and/or transmitted bits). An example implementation of thesequence estimation circuit210 is described below with reference toFIG. 3.
Thenon-linear modeling circuit236amay apply a non-linearity function
(a model of the non-linearity seen by the received signal en route to the circuit210) to thesignal205 resulting in thesignal203. Similarly, thenon-linear modeling circuit236bmay apply the non-linearity function
to thesignal209 resulting in thesignal207.
may be, for example, a third-order or fifth-order polynomial. Increased accuracy resulting from the use of a higher-order polynomial for
may tradeoff with increased complexity of implementing a higher-order polynomial. Where FnITx is the dominant non-linearity of the communication system100,
modeling only FnITx may be sufficient. Where degradation in receiver performance is above a threshold due to other non-linearities in the system (e.g., non-linearity of the receiver front-end108) the model
may take into account such other non-linearities
FIG. 3 is a block diagram depicting an example sequence estimation circuit for use in a system configured for low-complexity, highly-spectrally-efficient communications. Shown are acandidate generation circuit302, ametrics calculation circuit304, acandidate selection circuit306, acombiner circuit308, abuffer circuit310, abuffer circuit312, a phase adjustcircuit314, andconvolution circuits316aand316b. The sequence estimation process described with respect toFIG. 3 is an example only. Many variations of the sequence estimation process are also possible. For example, although the implementation described here uses one phase survivor per symbol survivor, another implementation may have PSu (e.g., PSu<Su) phase survivors that will be used commonly for each symbol survivor.
For each symbol candidate at time n, themetrics calculation circuit304 may be operable to generate a metric vector D_n¹. . . D_n^M×Su×Pbased on the partial response signal PR1, thesignal303aconveying the phase candidate vectors PC_n¹. . . PC_n^M×Su×Pand thesignal303bconveying the symbol candidate vectors SC_n¹. . . SC_n^Mn×Su×P, where underlining indicates a vector, subscript n indicates that it is the candidate vectors for time n, M is an integer equal to the size of the symbol alphabet (e.g., for N-QAM, M is equal to N), Su is an integer equal to the number of symbol survivor vectors retained for each iteration of the sequence estimation process, and P is an integer equal to the size of the phase alphabet. In an example implementation, the size of phase alphabet is three, with each of the three symbols corresponding to one of: a positive shift, a negative phase shift, or zero phase shift, as further described in the United States patent application titled “Low-Complexity, Highly-Spectrally-Efficient Communications,” and in the United States patent application titled “Fine Phase Estimation for Highly Spectrally Efficient Communications,” each of which is incorporated herein by reference, as set forth above. In an example implementation, each phase candidate vector may comprise Q phase values and each symbol candidate vector may comprise Q symbols. An example implementation of the metrics calculation block is described in the United States patent application titled “Low-Complexity, Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.
Thecandidate selection circuit306 may be operable to select Su of the symbol candidates SC_n¹. . . SC_n^Mn×Su×Pand Su of the phase candidates PC_n¹. . . PC_n^Mn×Su×Pbased on the metrics D_n¹. . . D_n^Mn×Su×P. The selected phase candidates are referred to as the phase survivors PS_n¹. . . PS_n^Su. Each element of each phase survivors PS_n¹. . . PS_n^Sumay correspond to an estimate of residual phase error in thesignal232. That is, the phase error remaining in the signal after coarse phase error correction via the phase adjustcircuit206. The best phase survivor PS_n¹is conveyed via signal307a. The Su phase survivors are retained for the next iteration of the sequence estimation process (at which time they are conveyed viasignal301b). The selected symbol candidates are referred to as the symbol survivors SS_n¹. . . SS_n^Su. Each element of each symbol survivors SS_n¹. . . SS_n^Sumay comprise a soft-decision estimate and/or a hard-decision estimate of a symbol of thesignal232. The best symbol survivor SS_n¹is conveyed to symbol buffer310 via thesignal307b. The Su symbol survivors are retained for the next iteration of the sequence estimation process (at which time they are conveyed via signal301a). Although, the example implementation described selects the same number, Su, of phase survivors and symbol survivors, such is not necessarily the case. Operation of examplecandidate selection circuits306 are described in the United States patent application titled “Low-Complexity, Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.
Thecandidate generation circuit302 may be operable to generate phase candidates PC_n¹. . . PC_n^M×Su×Pand symbol candidates SC_n¹. . . SC_n^Mn×Su×Pfrom phase survivors PS_n-1¹. . . PS_n-1^Suand symbol survivors SS_n-1¹. . . SS_n-1^Su, wherein the index n−1 indicates that they are survivors from time n−1 are used for generating the candidates for time n. In an example implementation, generation of the phase and/or symbol candidates may be as, for example, described in one or more of: the United States patent application titled “Low-Complexity, Highly-Spectrally-Efficient Communications,” and the United States patent application titled “Joint Sequence Estimation of Symbol and Phase with High Tolerance of Nonlinearity,” each of which is incorporated herein by reference, as set forth above.
Thesymbol buffer circuit310 may comprise a plurality of memory elements operable to store one or more symbol survivor elements of one or more symbol survivor vectors. Thephase buffer circuit312 may comprise a plurality of memory elements operable to store one or more phase survivor vectors. Example implementations of thebuffers310 and312 are described in the United States patent application titled “Low-Complexity, Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.
Thecombiner circuit308 may be operable to combine the best phase survivor, PS_n¹, conveyed via signal307a, with thesignal228 generated by the carrier recovery circuit208 (FIG. 2) to generate fine phase error vector FPE_n¹, conveyed viasignal309, which corresponds to the finely estimated phase error of the signal222 (FIG. 2). At each time n, fine phase error vector FPE_n-1¹stored inphase buffer312 may be overwritten by FPE_n¹.
The phase adjustcircuit314 may be operable to adjust the phase of thesignal315aby an amount determined by thesignal234 output byphase buffer312, to generate thesignal205.
Thecircuit316a, which performs a convolution, may comprise a FIR filter or IIR filter, for example. Thecircuit316amay be operable to convolve thesignal132 with response ĥ, resulting in the partial response signal315a. Similarly, theconvolution circuit316bmay be operable to convolve the signal317 with response ĥ, resulting in thepartial response signal209. As noted above, response ĥ may be stored by, and/or conveyed to, thesequence estimation circuit210 in the form of one or more tap coefficients, which may be determined based on the tap coefficients of thepulse shaper104 and/orinput filter109 and/or based on an adaptation algorithm of a decision feedback equalizer (DFE). Response ĥ may thus represent a compromise between attempting to perfectly reconstruct the total partial response signal (103 as modified bypulse shaper104 and input filter109) on the one hand, and compensating for multipath and/or other non-idealities of thechannel107 on the other hand. In this regard, thesystem100 may comprise one or more DFEs as described in one or more of: the United States patent application titled “Decision Feedback Equalizer for Highly-Spectrally-Efficient Communications,” the United States patent application titled “Decision Feedback Equalizer with Multiple Cores for Highly-Spectrally-Efficient Communications,” and the United States patent application titled “Decision Feedback Equalizer Utilizing Symbol Error Rate Biased Adaptation Function for Highly-Spectrally-Efficient Communications,” each of which is incorporated herein by reference, as set forth above.
Thus, signal203 is generated by taking a first estimate of transmitted symbols, (an element of symbol survivor SS_n¹SS_n¹), converting the first estimate of transmitted symbols to the partial response domain viacircuit316a, and then compensating for non-linearity in thecommunication system100 viacircuit236a(FIG. 2). Similarly, signal207 is generated from a second estimate of transmitted symbols (an element of symbol survivor SS_n¹) that is converted to the partial response domain bycircuit316bto generatesignal209, and then applying a non-linear model to the signal209bto compensate for non-linearity in the signal path.
FIG. 4 is a block diagram depicting an example partial response pulse-shaping filtering setup for use in a system configured for low-complexity, highly-spectrally-efficient communications. Referring toFIG. 4, there is shown afiltering system400, which may comprise a transmit (Tx) pulse-shaper410, a receive (Rx) pulse-shaper420, and afiltering manager430.
Each of the Tx pulse-shaper410 and the Rx pulse-shaper420 may comprise suitable circuitry, interfaces, logic, and/or code for performing pulse shaping, such as during communication of data between a transmitter and a receiver which may comprise the Tx pulse-shaper410 and the Rx pulse-shaper420, respectively. The Tx pulse-shaper410 and the Rx pulse-shaper420 may correspond to thepulse shaper104 and theinput filter109, respectively, ofFIG. 1. In this regard, the Tx pulse-shaper410 and the Rx pulse-shaper420 may enable adjusting signal waveforms (e.g., of communicated signals) such as to comply with particular spectral requirements (i.e. the “spectral masks”) of communication channels that may be used. The waveform adjustments provided by the Tx pulse-shaper410 and the Rx pulse-shaper420 may entail various processing operations and/or functions, particularly filtering. In this regard, processing signals by the Tx pulse-shaper410 and the Rx pulse-shaper420 may comprise filtering. For example, filtering may be performed (at the transmit-side and the receive-side) to ensure that communicated signals conform to the applicable spectral mask(s) and/or to ensure successful and/or proper recovery of signals at the receiver side.
The Tx pulse-shaper410 and the Rx pulse-shaper420 may be implemented using various types of filters, including, for example, infinite impulse response (IIR) filters and/or finite impulse response (FIR) filters. The invention is not necessarily limited, however, to any particular type of filter. In various embodiments of the invention, the pulse-shaper410 and the Rx pulse-shaper420 may be configured to provide partial response pulse-shaping, substantially as described with respect to FIG.1, for example. In this regard, partial response pulse shaping may comprise generating (and/or handling) waveforms that intentionally incorporate a substantial amount of inter-symbol interference (ISI). Allowing such a substantial amount of ISI may provide enhanced communication performance. For example, partial response based modulation, as compared to legacy and/or existing modulation schemes, may allow for use of less bandwidth for communicating the same amount of data (and with the similar error rates), for communication of same amount of data over same bandwidth but with lower power consumption, and/or for communication more data with similar bandwidth. Accordingly, while legacy/existing pulse shaping (and particularly filtering related thereto) may be configured to have no-ISI (or minimal amount of ISI—e.g., corresponding to applicable error rates), the PR based modulation applied using the Tx pulse-shaper410 and the Rx pulse-shaper420 may be implemented to operate with a substantial amount of ISI.
Thefiltering manager430 may comprise suitable circuitry, interfaces, logic, and/or code that may be operable to manage filtering related processing, such as filtering processing performed via one or both of the Tx pulse-shaper410, the Rx pulse-shaper420. In this regard, filtering management may comprise configuring filters and/or filtering operations to meet spectrum mask limitations according to an underlying communication/modulation standard without degrading demodulation performance. Thus, thefiltering manager430 may be operable to determine and/or provide filtering configuration related parameters to one or both of the Tx pulse-shaper410 and the Rx pulse-shaper420 (Tx_side configuration data452 andRx_side configuration data454, respectively). In this regard, the configuration data may comprise, for example, filtering configuration parameters such as number of filter taps and/or filter coefficients values.
In various embodiments of the invention, thefiltering manager430 may be configured to implement and/or execute filtering optimization processes, which may be used to determine filtering configurations that optimize filtering operations based on particular criteria or conditions (e.g., for partial response based filtering). For example, thefiltering manager430 may be used to perform filtering optimization processes, which may be based on, for example,control data442 that may comprise preconfigured criteria or user inputs (e.g., definitions, preferences, etc.), to produce optimized filter coefficient values (e.g., for the Tx pulse-shaper410 and the Rx pulse-shaper420). In some instances, the optimization processes may also be based on information pertaining to and/or provided by the transmitter, the receiver, and/or the communication channels/links. Theresults444 of the optimization process (e.g., optimized filter configuration parameters) may be used in generating theTx_side configuration data452 andRx_side configuration data454 which may be sent directly to the filters and/or may be outputted to the users (e.g., to system designers or operators), such as using log files or screens.
Thefiltering manager430 may comprise various components and/or subsystems which may be utilized during filtering management. For example, in some instances, thefiltering manager430 may comprise an input/output (I/O)component432, which may incorporate various I/O mechanisms or devices (e.g., mice, keyboards, keypads, user interface, displays, touch screens or touchpads, and the like) to enable interacting with users (e.g., system designers and/or operators). In this regard, the I/O component432 may be used to allow inputting control data or commands, such as implementation-specific parameters, variables, constraints and cost functions that may define the desired optimization. The I/O component432 may also be used to allow outputting data to the system users. For example, the I/O component432 may comprise suitable graphic user interfaces (GUIs) and/or interface (e.g., a file transfer interface) for acknowledging (or repeating—e.g., for confirmation) user inputs, and/or to provide corresponding output (e.g., optimization results), such as suitable filter tap coefficients or other configuration parameters, for example.
Thefiltering manager430 may also comprise aprocessing component434, which may be configured to perform various processing functions or tasks in support of operations performed by thefiltering manager430. For example, theprocessing component434 may be used to implement and/or execute any filtering optimization processes that may be supported by the filtering manager430 (e.g., PR filtering optimization process). In this regard, designing and/or configuring filtering operations may entail analysis of the spectrum response of any filters used (at the transmit-side and/or the receive-side). The spectrum response may be calculated using, for example, Fourier transform. In this regard, the spectrum response of the PR signal (i.e. the total partial response of the system) may be calculated or estimated using, for example, a model for nonlinear distortion (e.g., 3rd order polynomial, or Voltera series).
Thefiltering manager430 may be implemented as a dedicated, stand-alone system or device (e.g., general-purpose computer, dedicated controller, or the like), which may be connected (e.g., using wireless and/or wired connections, based on any appropriate interface/standard) to one or both of the transmitter and receiver comprising the Tx pulse-shaper410, the Rx pulse-shaper420, respectively, to enable interactions therewith during filtering management. Alternatively, in some instances, at least a portion of the filtering manager430 (and/or functions or operations performed thereby) may be incorporated into the transmitter and/or the receiver, or component(s) thereof. In some instances, thefiltering manager430 may be implemented in distributed manner, with components and/or functions thereof being distributed among the transmitters, the receiver, and/or a stand-alone device/system. In some instances, thefilter manager430 may be configured to optimize filtering processing in the Tx pulse-shaper410 and/or the Rx pulse-shaper420.
In operation, theTx pulse shaper410 and theRx pulse shaper420 may be used to provide pulse shaping—e.g., waveform adjustments based on particular spectral masks, nonlinearity, symbol constellation, etc.—which would typically entail performing filtering. In this regard, thefilter manager430 may be used to configure, control, and/or manage filtering performed by theTx pulse shaper410 and/or theRx pulse shaper420. For example, theTx pulse shaper410 and theRx pulse shaper420 may be implemented using infinite impulse response (IIR) filters and/or finite impulse response (FIR) filters, and as such configuring, controlling, and/or managing filtering may comprise, inter alia, determining or setting the number of taps (or “length”) and/or the values of the tap coefficients of theTx pulse shaper410 and theRx pulse shaper420. In this regard, the number of taps may comprise the number of filter taps for each of theTx pulse shaper410 and theRx pulse shaper420, and/or the total number of taps (i.e. combined number of taps of both of theTx pulse shaper410 and the Rx pulse shaper420).
In some instances, theTx pulse shaper410 and/or theRx pulse shaper420 may be configured for use in conjunction with partial response based communications. In this regard, one distinction between a partial response (PR) based scheme and existing/legacy schemes pertains to inter-symbol interference (ISI). Inter-symbol interference (ISI) is typically undesired, and existing/legacy schemes are typically tailored to eliminate or significantly minimize ISI (e.g., either being configured to achieve zero ISI or near-zero ISI, or when allowing for ISI it would be minimal amount, typically based on specified acceptable error rates). Achieving such zero (or near-zero) ISI usually comes at the expense of the baud rate (or the bandwidth efficiency)—e.g., being significantly worse than the channel's Shannon capacity bound.
For example, most existing filtering implementations, particularly Root Raised Cosine (RRC) and/or Raised Cosine (RC) based implementations, are typically configured to have near-zero ISI, to assure demodulating performance. For example, with RRC-based implementations, RRC-based filters would typically be used in both of the transmitter and receiver, serving as pulse shaping filter and match filter, respectively. The challenge in RRC-based filter design is to adjust the truncated and quantized RRC taps to meet spectrum mask limitations without generating inherent ISI. In this regard, the conventional root raised cosine (RRC) pulse shaping filter is an even function to assure linear phase response (i.e. the time domain filter response is symmetric) and the length of the casual part is equal or similar to the non-casual. The modulated signal, which is the convolution between the filter taps and the symbols, may incorporate peaks that are associated with nonlinear distortion that typically cannot be compensated in the receiver because the non-casual part may not be known to the receiver.
In partial response (PR) based communications as implemented in accordance with aspects of the present invention, however, end-to-end system/path (particularly theTx pulse shaper410 and the Rx pulse shaper420) may be configured such that signals communicated between the transmitter and the receiver would intentionally have a substantial amount of inter-symbol interference (ISI). In this regard, allowing for substantial amount of ISI may allow for increasing the symbol rate (and/or reducing required bandwidth), which may allow, when drawbacks of having substantial ISI are mitigated (e.g., through signal processing techniques in the receiver) for enhanced performance. In other words, use of partial response pulse shaping may allow for generating signals with particular inherent and substantial amount of ISI such that to enable compressing the signal spectrum to allow improved spectral efficiency. In this regard, ISI level allowed in partial response based implementations may be worse than the threshold signal-to-noise ratio (SNR) whereas in legacy systems/schemes, which typically have near zero ISI (having similar spectral efficiency), the ISI level would typically be much better than threshold SNR. For example, in a legacy system that is configured (e.g., based on particular standard) to have particular SNR threshold (e.g., 30 dB), the ISI, which would be treated as an interference and thus factor into the overall noise, would have to be negligible (e.g., be 20 to 30 dB down from the SNR threshold). On the other hand, in a corresponding partial response based implementation that specifically allows for the presence of ISI (and account for it in the PR based estimation process), the ISI may be substantial relative to the same SNR threshold (30 dB) since it does not be counted as noise (e.g., ISI can be equal to or greater than the SNR threshold of 30 dB).
Therefore, to achieve the desired performance enhancements when using partial response pulse shaping, various aspects of operations and/or functions that may be required for providing end-to-end communicability may be configured and/or adjusted to account for the partial response (and particularly, for the presence of substantial ISI). With respect to the filtering operations, in PR-based implementations, the number of taps and/or the values of the tap coefficients of theTx pulse shaper410 and the Rx pulse shaper420 (e.g., as determined by the filtering manager430) may be designed such that the pulse shaper intentionally may be non-optimal in terms of noise tolerance in order to improve tolerance of nonlinearity. In this regard, the PR pulse shaping providing by theTx pulse shaper410 and theRx pulse shaper420 may offer superior performance in the presence of nonlinearity as compared to, for example, the conventional root raised cosine (RRC) pulse shaping filters. In this regard, the PR communication path may be designed and/or configured such that effects of the substantial ISI may be mitigated and/or accounted for based on use of a specifically configured estimation process, as described, for example, in the United States patent application titled “Low-Complexity, Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.
In various embodiments, the design and configuration of the pulse shape filtering (i.e. transmit-side and/or receive-side filters) may be based on optimization processes, which may be configured to achieve and/or balance between several objectives. Furthermore, the filtering optimization processes may be subject to implementation-specific performance-related parameters, variables, constraints, and/or cost functions, which may be defined for particular desired optimization. In this regard, the one or more performance-related parameters may pertain to the filtering itself and/or other aspects of the communication that may be affected by it (e.g., error measurements). The constraints may comprise particular limits or requirements, which may apply to, for example, the performance-related variables and/or the filters (or their configuration). For example, spectral mask limits and/or compliance therewith may be constraint in any filtering configuration or optimization thereof. The cost functions may be defined over one or more performance-related parameters or variables—e.g., specifying the filter configuration (for example, filter tap coefficient values) is to be optimized based on a function of one or more variables. The cost functions may, in some instances, specify different weights to pertinent parameters or variables. The implementation-specific parameters, variables, constraints and/or cost functions may be defined by system designers or operators, such as via user input (which may be provided as part of the control data442). Thus, the filter optimization process may provide optimal filtering configuration that may achieve the primary objective(s), as applied to the cost functions to meet (as much as possible) the specified performance-related parameters or variables, while complying with the specified constraints.
For example, the primary design objective for filtering optimization may be spectrum mask compliance, including impact of nonlinear distortion. In particular, it may be desirable (and thus be considered in designing and/or optimizing filtering operation) to comply with applicable spectrum masks as closely as possible to achieve as close to the Shannon capacity limit as possible. Another (or alternative) primary design objective for filtering optimization may be symbol error rate minimization for equalization processing. In this regard, typical adaptive equalization method may be based on least means square (LMS) algorithms, which may not be optimal in certain conditions (e.g., in case of severe multipath) and may not assure lowest SER and/or BER. Accordingly, the equalization process used in conjunction with partial response based communications may be configured as an adaptive process which may minimize SER (thus possibly providing optimal performance subject to the finite equalizer length). In some instances, the adaptive equalization used may incorporate parameters that may be based on the SER expression of a PR system/path. An example implementation of such equalization method is described in the United States patent application titled “Decision Feedback Equalizer Utilizing Symbol Error Rate Biased Adaptation Function for Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.
The filtering design and/or configuration may account for nonlinearity, such as by considering nonlinear distortion when configuring the filters (e.g., filter taps). In this regard, tolerating nonlinearity (particularly at the receive-side) may allow for use of an optimal observation model—i.e. not needing to account for nonlinearity at the transmit-side. Other design objectives may comprise partial response minimum distance reduction (e.g., for worst case error patterns), minimization of symbol error rate (SER) and/or bit error rate (BER), early taps amplitude maximization, residual non-compensated early response, minimization of noise enhancement (e.g., at the receive-side), minimization of amplitude of tail response, nonlinear tolerance, adjacent channel rejection, and/or decimation filtering to avoid noise folding (i.e. anti-aliasing).
The partial response (PR) filtering, and/or the configuring thereof, may be associated with minimum distance reduction. In this regard, minimum distance reduction may be caused by the partial response's inherent ISI and the fact that every sample (at the transmit-side and/or the receive-side) may correspond to a linear combination of multiple symbols. In other words, rather than slice or sample for a single symbol as many legacy/existing systems do, in PR based implementations, the receive side may be configured to estimate a sequence of symbols. The degradation associated with PR pulse shaping may be compared to flat pulse shape (or RRC/RC shaping) which may typically introduce no (or very little) ISI. The determined degradation may represent a theoretical bound for the more optimal (e.g., Maximum Likelihood or ML) estimation, and it may be perceived as the penalty for spectrum compression—i.e. a tradeoff for using less bandwidth or being able to send more symbols. In various instances, partial response detection may be associated with error bursts that may have typical patterns, which may be determined based on analysis, simulations and/or based on tracking/monitoring. In this regard, every symbol mapping may have a worst case error pattern that may yield the minimum distance reduction. Error patterns and/or minimum distance are described in more detail in the United States patent application titled “Decision Feedback Equalizer for Highly-Spectrally-Efficient Communications,” which is incorporated herein by reference, as set forth above.
Typically, high order mapping (e.g., 128-QAM) may associate with higher reduction of minimum distance than low order mapping (e.g. QPSK) because high order mapping may have much more combinations of error patterns than low order mapping would have. For example worst case pattern error of QPSK mapping may be [1 −1] and for 16-QAM is [1 −2 2 −1], where a ‘1’ in the error pattern vector may reflect the distance to an adjacent symbol. The following table shows examples of partial response pulse shape performance:
TABLE 1
Examples of particular partial response based schemes
Error Pattern
Gross Minimum Related to
PR Base Capacity Spectral Distance Minimum
Scheme Modulation Gain Efficiency Reduction Distance
PR5 QPSK 2.5 5 b/s/Hz −2.5 dB [1 −1 1 −1]
PR10 32-QAM 2 10 b/s/Hz −8.0 dB [1 −2 2 1]
Accordingly, one of the filtering design optimization goals may be minimizing minimum distance reduction. In some instances, there may be several error patterns which have similar reduction of minimum distance, and thus more than one error pattern should be considered in the optimization. Another optimization goal may be based on symbol error ration. In this regard, the SER function applicable in the system may be configured such that it may consider the contribution of worse error patterns, each associated with its weight.
The design method can be based on an analytic or a numerical approach (or may use both). The optimization may be driven by one or more design considerations or goals, such as minimum distance reduction for selected error patterns for a given spectrum compression gain, or for spectrum compression gain for a given minimum distance level. In this regard, the selected error patterns may be the ‘n’ most frequently occurring error patterns (‘n’ being a non-zero natural number), such as seen during simulation and/or operation (i.e. use) of the system. The optimization may be limited by one or more constraints, such as spectrum mask limitations of the pulse shape taps (with or without the impact of a nonlinear model).
Due to spectrum mask limitations the pulse shape impulse response is typically smooth. Accordingly, it may be desirable to select filter tap coefficients that grow gradually from zero. Thus the first (“early”) taps of the pulse shape filter may be chosen (e.g., by filtering manager430) to have low values. The partial response signal (at the transmit-side) may be a convolution between the partial response filter taps and symbol stream. The new (current) symbol contributes only to the most new sample and the contribution level in the newest sample is proportional to the value of the first (earliest) tap coefficient of the partial response filter. On the receive side, data (e.g., symbol) extraction may be based on use of suboptimal search mechanisms (e.g., based on M-ary algorithm, an implementation of which is referred to herein as High Probability Sequence Estimation or ‘HPSE’). Use of such mechanism (i.e. sub-optimal estimation) may result in estimates of a symbol with low reliability. In case of low signal-to-noise ratio (SNR), new symbol estimation might be faulty, and may cause error bursts. Therefore, to improve error performance, a consideration (goal) of the filter optimization may be maximizing “early” taps amplitude while complying with other constraints. Alternatively, constraints can be applied for some (i.e. a subset) of the “early” taps amplitude. By optimizing the level of early taps, the demodulator resources and complexity may be significantly relaxed. In this regard, demodulating/decoding (at the receive-side) based on sequence estimation may grow significantly (e.g., exponentially) with the search dimension. The HPSE process may use the early taps for exercising the search. The number of survivors and taps needed for effective search, which provide high performance in low SNR conditions, may highly depend on the amplitude of the early taps. Therefore, optimizing early taps level may be important for reducing the number of survivors and taps participating in the search process, thus saving dramatically resources, power, and cost of the demodulator (in particular resources used for the sequence estimation).
Minimum distance of the partial response signal may affect sequence estimation performance, thus it can be considered when optimizing the minimum distance reduction of the entire path (i.e. including both of the transmit-side and the receive-side filters). In this regard, the receive-side filter may be optimized to minimize minimum distance reduction while also providing a determined (e.g., based on the control data442) amount of noise rejection and out-of-band rejection to reject interferences such as adjacent channels.
As discussed above, the total partial response consists of the convolution of pulse shaping transmit-side and receive-side filters. The convolution may yield a preceding tail, which may not be compensated by the demodulator (thus resulting in introduction of floor to the demodulator). Therefore, an optimization design goal may be to minimize the preceding tail, such as below the operating point. In this regard, because the tail of the partial response may cause error duplication in the sequence estimation, the optimization process may be tailored to control the value of tail tap coefficients to improve stability, such as at low SNR conditions.
The receive-side filter may be configured, for example, to reject interference and provide optimal SNR performance in presence of additive white Gaussian noise (AWGN). Design considerations (goals) particularly for the receive-side filter may include noise enhancement (e.g., in comparison to an ideal match filter), partial response performance, out-of-band rejection and impacts on tap coefficients of early taps as well as tap coefficients of residual taps.
The observation model (referred to as the “WAM” observation) used in designing/optimizing PR pulse shaping filters as described herein may not necessarily provides “sufficient statistics” and consequently may be sub-optimal in comparison to estimation approaches currently in use (which are typically based on maximum likelihood or ‘ML’). The WAM observation model, however, may perform better in case of reduced complexity sequence estimation with less degradation around threshold SNR. Additionally the observation model may incorporate, as an optimization consideration or goal (e.g., as part of the transmit-side and/or receive-side pulsing shaping design), nonlinear distortion compensation. Therefore, the nonlinear model may be applied on the reconstructed partial response survivors/successors to ensure that the partial response being used by the time domain sequence estimation would be close to the transmit-side pulse shape taps (i.e. assure that nonlinearities generated on the transmit-side, such as by the power amplifier, may be observed by the sequence estimation in the receive-side). Accordingly, the observation model used herein may optimize the distance between transmit-side pulse shape and the observation response (overall PR) in presence of the nonlinear distortion.
The sequence estimation used in the receive-side may be configured such that its time domain response should be close to the transmit-side pulse shape response (e.g., to assure nonlinear tolerance). The transmit-side partial response signal may be affected by the transmit-side nonlinearities. The sequence estimation, however, is typically configured to use the total partial response, which may be different from transmit-side pulse shape due to the use of filter on the receive side (i.e. receive-side filter). In this regard, the nonlinear model is applied over the reconstructed signal in the receiver to satisfy the receive-side samples. Therefore it may be important that the time domain response of the total partial response be similar to (or very close to) the transmit-side partial response, so the impact of applying the nonlinear model on the receive-side may provide similar output as the actual nonlinearities (i.e. including the transmit-side nonlinearities). This may be achieved by using, as part of the filtering optimization processing, a cost function or a constraint that may be applied to assure the above requirement is a least means square (LMS) function between the transmit-side pulse shape taps and the total PR response taps that is used for the sequence estimation.
Another type of factors that may affect partial response (and thus filtering operations and/or filter configuration pertaining thereto) is channel conditions. In other words, channel conditions (and any measures or functions adapted to handle or compensate for them in one or both of the transmitter and the receiver) may constitute performance-variables and/or constraints which may pertinent to (and thus must be accounted for in) the optimization process. In this regard, channel variations, such as channel attenuation (fading), which has impact on signal-to-noise ratio (SNR), multipath (selective fading), and received interference signals, may be accounted and/or compensated for, preferably concurrently by both the transmit-side and the receive-side for best performance. In some implementations, certain channel conditions (e.g., flat channel attenuation) may be accounted and/or compensated for, including at the transmit side. For example, flat channel attenuation may be compensated for, such as by adapting transmit (Tx) power (in the Tx pulse shaper410), as well as supporting receiver (Rx) dynamic range (in the Rx pulse shaper420). Multipath can be compensated for by the receive-side equalization, which may incorporate, for example, a noise enhancement penalty. In addition, in some instances multipath may partially be addressed by use of transmit-side equalization, which advantageously may not suffer from the noise enhancement as in the case of the receive-side equalization. In this regard, use of partial response based communications may allow for use of equalization in the transmitter since introducing (controlled) ISI is acceptable (and mainly part of the partial response scheme). Thus, as result of the use of transmit side equalization, noise may not be enhanced as much in the receive-side equalization. Use of transmit-side equalization may have its limitations, however. For example, use of full equalization at the transmit-side may affect minimum distance, and/or transmit-side equalization may be limited by the transmit spectrum mask. Accordingly, in one embodiment, desired total equalization may be achieved based on combining of transmit-side and receive-side equalization. In this regard, in some instances, only “partial” equalization may be performed in the transmit-side (i.e. in the Tx pulse shaper410), with the remaining equalization being performed in the receive-side (i.e. in the Rx pulse shaper420).
In case of interference signals, the Rx pulse shaper420 (particularly filters therein and/or filtering operations performed thereby) may be adaptively configured to nullify the interference. Such adaptation may, however, cause degradations such as noise enhancement. Therefore an optimization may be used to achieve better performance. In this regard, transmit-side and receive-side filters adaptation may be associated with impact for partial response modulation attributes, such as minimum distance reduction, early taps amplitude, and/or uncompensated preceding response, which may be considered in the adaptation optimization. For example, the transmit-side and receive-side filters may adapt dynamically to optimize BER, SER and/or throughput performance, and to improve stability under channel variations and interferences mentioned above. The optimization objectives could be subject to the constrains that are applied in the filters design/optimization process—e.g., spectrum mask, minimum distance reduction, receiver noise enhancement, receiver adjacent rejection, receiver anti-aliasing, matching the transmitter time domain response with the total response for non-linear toleration, early taps amplitude and the uncompensated preceding response (floor). In this regard, although the SER and BER may be highly coupled, and BER may typically improve when SER improves, it is possible that BER may not provide the expected improvement for a given SER improvement due to possible dependency within symbol errors and/or the characteristics of the forward error correction (FEC) or channel coding being used. In this regard, in some instances the best BER performance for a given SER may result from less dependency within the symbol errors—i.e. when symbol errors have less bursty nature. The statistical length of a symbol error burst may depend, for example, on the time domain profile of the partial response (e.g., taps amplitude). In instances where the partial response decays fast (e.g., within a small number of taps), it may assure shorter bursts of symbol error, but it may impact other important attributes of the partial response, such as early taps amplitude and uncompensated preceding response, which are related to the overall performance. Therefore the BER may be considered for the optimization along with the SER.
In various embodiments, different configurations may be utilized for achieving a given (i.e. the same) gross spectral efficiency. The following table describes examples of different configuration modes for achieving a particular spectral efficiency—e.g., 10 bit/sec/Hz, which may correspond to PR10:
TABLE 2
Examples of different configuration modes for PR10
Bits Spectrum Spectrum
Symbol per Compression Compression
Constellation Symbol Rate Efficiency
16-QAM 4 2.50 10 b/s/Hz
32-QAM 5 2.00 10 b/s/Hz
64-QAM 6 1.67 10 b/s/Hz
As shown in table 2, it is possible to achieve the same spectrum compression efficiency by use of different configuration modes, each of which resulting in different spectrum compression rate, which may be measured in relation to the same reference (legacy) configuration—e.g., a spectrum compression rate value of 1.00 may correspond to, for all three modes, a legacy RRC based 1024-QAM. Having higher spectrum compression rate may provide higher dimensions in the Euclidian domain (which may generate a steeper SER curve). However, due to the limited complexity of practical sequence estimation performance at low SNR are degraded (cutoff). The opposite result may be obtained by using low spectrum compression rate with high order modulation. The amplitude distribution characteristics (CCDF) of the partial response signals which may have been deeply compressed may be degraded—i.e. such PR signals may be more sensitive to nonlinear distortion than PR signals configured with lower spectrum compression rate and higher constellation order. In some instances, spectrum compression rate and constellation order may be optimized according to system requirements (e.g., performance under nonlinearities, SNR dynamic range, and complexity). In some instances, when using high spectrum compression rate it may be possible to exclude FEC in the system as the SER and BER curve may be steep enough.
For example, filtering operations performed via theTx pulse shaper410 and theRx pulse shaper420, corresponding to the total partial response, may be configured using, e.g., simulation (which may be performed during system design). For example, the filtering of theTx pulse shaper410 and theRx pulse shaper420 may be designed in a MATLAB environment, such as using the function ‘fmincon( )’. In this regard, the function ‘fmincon( )’ may minimize a particular object under predefined constraints. In the present partial response (PR) scheme described herein the minimization objective may be minimal distance reduction, or a weighted combination of minimal distance reductions for the worst case error patterns, or SER.
The constraints may comprise, for example, spectrum mask limits, tap coefficients of early taps, preceding power response(s), receive-side out-of-band rejection, matching the transmitter time domain response with the total response for non-linear toleration and/or noise mismatch. In this regard, a significant constraint in designing and/or configuring filtering operations may the transmit-side (i.e. at the Tx filters) spectrum mask limits, with or without the impact of nonlinear distortion. In other words, the pulse shaping resulting from the filtering operations (particularly at the transmit-side) must typically comply with specific spectral masks, such as in accordance with the applicable governmental regulation. The tap coefficients of the filters' ‘early’ taps may also be treated as a constraint. In this regard, the tap coefficients of the filters' ‘early’ taps may be constrained to assure the performance of the sequence estimation, particularly the High Probability Sequence Estimation (HPSE) used herein or any other sub-optimal sequence estimation. The preceding (i.e. pre cursor) power responses may be included as part of the optimization constraints. In this regard, the preceding (pre-cursor) power response of the total partial response may be limited to a certain level, such as to minimize the impact of sensitivity degradation and flooring. The total pre-cursor power may be a summation of all individual pre-cursor taps, but the sequence estimation (e.g., HPSE) may consist of convolution with the overall partial response taps that reject out-of-band signal components, thus the actual pre-cursor floor may be evaluated by the total taps power of the convolution of pre-cursor taps with partial response taps. Such approach may allow further flexibility and consequently better performance of the system. Noise mismatch may also be treated as an optimization constraint. In this regard, the receive-side filtering (e.g., filtering by the Rx pulse shaper420) may be incorporated into the minimization objective, but may also affect the SNR at its output. For example, the output SNR degradation (at the receive-side filter) comparing to a perfectly match filter may be set as a constraint.
FIG. 5 is a block diagram depicting an example finite impulse response (FIR) implementation of partial response pulse-shaping filters, in accordance with an embodiment of the present invention. Referring toFIG. 5, there is shown a transmit (Tx)filter510 and a receive (Rx)filter520. In this regard, theTx filter510 and theRx filter520 may correspond to theTx pulse shaper410 and the pulse shaper420 (or at least the filtering components/functions thereof), respectively.
In the example embodiment shown inFIG. 5, theTx filter510 and theRx filter520 may be implemented as finite impulse response (FIR) filters. It is understood, however, that the invention is not so limited, and that the FIR implementation is provided as non-limiting example embodiment.
TheTx filter510 may be configured as FIR of LTx-1 order—i.e. having LTx taps. In this regard, theTx filter510 comprise a plurality of delay (e.g., Z transform, applied as z⁻¹operator) elements512₁-512_LTx-1, a plurality of multipliers514₁-514_LTx(for applying a plurality of corresponding tap coefficients or weights: T₁-T_LTx), and an accumulator516 (which may be implemented using a plurality of combiners516₁-516_LTx1, which may connected in series—such that to allow accumulating the weighted sum of the current and a finite number of previous values of the Tx filter input).
Similarly, theRx filter520 may be configured as FIR of LRx-1 order—i.e. having LRx taps. In this regard, theRx filter520 comprise a plurality of delay (e.g., Z transform, applied as z⁻¹operator) elements522₁-522_LRx-1, a plurality of multipliers524₁-524_LRx(for applying a plurality of corresponding tap coefficients or weights: R₁-R_LRx), and an accumulator526 (which may be implemented using a plurality of combiners526₁-526_LRx-1, which may connected in series—such that to allow accumulating the weighted sum of the current and a finite number of previous values of the Rx filter input).
In operation, the Tx filter510 (operating in the transmitter) and the Rx filter520 (operating in the receiver) may be used to provide the required overall filtering during communications, particularly to provide pulse shaping and noise rejection. For example, in some instances theTx filter510 and theRx filter520 may be configured provide a total partial response, as described inFIGS. 1-4 for example. In this regard, the number and values of the filters taps (i.e. number of taps at the transmit-side and receive-side: LTx and LRx, and the values of the transmit-side and receive-side tap coefficients: T₁-T_LTxand R₁-R_LRx) may be determined based on application of filtering optimization that may be tailored for partial response based paths, substantially as described with respect toFIG. 4 for example.
FIG. 6 is a flow chart depicting an example of a process for joint optimization of transmitter and receiver pulse-shaping filters, in accordance with an embodiment of the present invention. Referring toFIG. 6, there is shown aflow chart600 comprising a plurality of exemplary steps for configuring and/or optimizing partial-response pulse-shaping based filtering.
Instep602, filtering related fixed constants and/or optimization-related parameters may be configured—e.g., determined and/or set. In this regard, filtering-related fixed constants may comprise, for example, Tx filter length and/or Rx filter length, sampling per symbol, nonlinear distortion backoff (which may be a proxy for Tx power), baud rate, and/or channel spacing. The optimization-related parameters may be selected based on the optimization process. For example, in instances where the optimization is performed using MATLAB function ‘fmincon( )’, the optimization parameters may comprise maximum number of iterations, tolerance, and/or search algorithm. Instep604, optimization constraints may be defined. In this regard, the optimization constraints may comprise, for example, spectral mask limit(s), Rx filter noise mismatch, Rx filter out-of-band rejection, early taps amplitude, preceding tail power, matching the transmitter time domain response with the total response for non-linear toleration and the like. Instep606, optimization variables may be structured, such as based on the applicable optimization process. For example, in instances where the optimization is performed by use of MATLAB function ‘fmincon( )’, the variables under optimization may comprise filters tap coefficients, which may be structured in vectors format according to the fmincon( ) definitions, for example.
Instep608, the filtering optimization processing may be performed, such as by initiating running of the MATLAB function fmincon( ) (or any other constrained nonlinear optimization technique). The function fmincon( ) may use one or more user defined scripts. For example, the function fmincon( ) may use two user defined scripts, including one for the cost function which holds the minimal distance reduction or SER as function of the filter tap coefficients, and a second function for holding the constraints expressions as function of the optimization variables (i.e. filters taps). Instep610, results of the optimization process may be monitored, and may be used in configuring the filtering operations (and/or the results may be stored for subsequent use).
FIG. 7 is a chart diagram depicting a comparison between the frequency-domain responses of a legacy transmit filter and an optimized partial response pulse-shaping transmit filter. Referring toFIG. 7, there is shown twocharts710 and720.
In this regard, chart710 shows a frequency-domain response of a reference transmit filter whereas thechart720 shows a frequency-domain response of a corresponding optimized partial response (PR) pulse-shaping transmit filter. As shown inFIG. 7, the reference transmit filter may comprise an RRC-based filter configured for 1024-QAM (with the transmitter RRC filter being configured to have a roll-off factor of 0.2), whereas the corresponding partial response (PR) pulse-shaping transmit filter may comprise an optimized PR transmit filter for processing 32-QAM symbols. In this regard, the PR pulse-shaping transmit filter represented inFIG. 7 may be optimized (along with the corresponding receive filter) using an optimization process as described inFIG. 6 for example.
Both ofcharts710 and720 show aspectral mask730 that may be defined for the transmitted signal. Spectral masks may be defined by a particular standards body, such as the European Telecommunications Standards Institute (ETSI). For example, thespectral mask730 may represent a spectral mask defined by theETSI EN 302 217-2-2 standard titled “Fixed Radio Systems; Characteristics and Requirements for Point-to-Point Equipment and Antennas.” This standard is available at www.etsi.org.
Inchart710, representing the frequency-domain response of the reference transmit filter (1024-QAM RRC transmit filter with roll-off factor of 0.2),curve712 represent the modulated signal of the reference (RRC-based) transmit filter whereas thecurve714 represent the distorted signal of the reference (RRC-based) transmit filter. Inchart720, representing the frequency-domain response of the optimized PR transmit filter (driven by 32-QAM),curve722 represent the modulated signal of the optimized PR transmit filter whereas thecurve724 represent the distorted signal of the optimized PR transmit filter. Comparing the corresponding signals (i.e.712 vs.722 and/or714 vs.724) demonstrates that the optimized PR transmit filter provides enhanced performance. For example, the response of the optimized PR transmit filter shows that it may be able to match the spectral mask limits as good or even better than reference (RRC-based) transmit filter (e.g., up to a particular target symbol rate—such as ˜15 MHz).
Other implementations may provide a non-transitory computer readable medium and/or storage medium, and/or a non-transitory machine readable medium and/or storage medium, having stored thereon, a machine code and/or a computer program having at least one code section executable by a machine and/or a computer, thereby causing the machine and/or computer to perform the steps as described herein for design and optimization of partial response pulse shape filter.
Accordingly, the present method and/or system may be realized in hardware, software, or a combination of hardware and software. The present method and/or system may be realized in a centralized fashion in at least one computer system, or in a distributed fashion where different elements are spread across several interconnected computer systems. Any kind of computer system or other system adapted for carrying out the methods described herein is suited. A typical combination of hardware and software may be a general-purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein.
The present method and/or system may also be embedded in a computer program product, which comprises all the features enabling the implementation of the methods described herein, and which when loaded in a computer system is able to carry out these methods. Computer program in the present context means any expression, in any language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after either or both of the following: a) conversion to another language, code or notation; b) reproduction in a different material form.
While the present methods and/or apparatus may have been described with reference to certain implementations, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the scope of the present method and/or apparatus. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the present disclosure without departing from its scope. Therefore, it is intended that the present method and/or apparatus not be limited to the particular implementations disclosed, but that the present method and/or apparatus will include all implementations falling within the scope of the appended claims.

Claims

What is claimed is:

1. A method, comprising:

utilizing pulse-shaping filtering for communication of data, wherein:

the pulse-shaping filtering provides a total partial response that incorporates a predetermined amount of inter-symbol interference (ISI); and

the predetermined amount of inter-symbol interference (ISI) is determined based on an estimation process applied to an output of the pulse-shaping filtering.

2. The method ofclaim 1, wherein the pulse-shaping filtering comprises use of a transmit-side pulse-shaping filter and a receive-side pulse-shaping filter.

3. The method ofclaim 1, comprising configuring the pulse-shaping filtering to optimize an applicable weighted distances for one or more selected error patterns for a given spectral compression.

4. The method ofclaim 1, comprising configuring the pulse-shaping filtering to optimize a symbol error rate (SER) associated with the communication of data.

5. The method ofclaim 1, comprising configuring the pulse-shaping filtering, the configuring comprising:

defining a plurality of performance-related variables;

setting a plurality of constraints; and

determining optimized filtering configuration by applying an optimization process that is based on, at least in part, the plurality of constraints and the plurality of performance-related variables.

6. The method ofclaim 5, comprising configuring the optimization process to allow for setting different numbers of filter coefficients for each of a transmit-side pulse-shaping filter and a receive-side pulse-shaping filter utilized for the pulse-shaping filtering.

7. The method ofclaim 5, comprising configuring the optimization process to determine one or more early filter taps that are applicable to one or more filters utilized for the pulse-shaping filtering while ensuring that a resultant pulse shape remain within an applicable spectral mask.

8. The method ofclaim 5, comprising configuring the optimization process to determine late filter taps applicable to one or more filters utilized for the pulse-shaping filtering while ensuring that a resultant pulse shape remain within an applicable spectral mask.

9. The method ofclaim 5, wherein the optimization process comprises optimizing at least one of a tap configuration, a number of taps, and one or more of a plurality of tap coefficients.

10. The method ofclaim 5, comprising applying as part of the optimization process one or more cost functions while meeting the plurality of constraints.

11. The method ofclaim 10, wherein at least one of the one or more cost functions cost function is defined over one or more of the plurality of performance-related variables.

12. The method ofclaim 5, wherein the plurality of performance-related variables comprise minimum distance reduction, symbol error rate (SER) and/or bit error rate (BER), early taps amplitude, residual non-compensated early response, receive-side noise enhancement, receive-side out-of-band rejection, amplitude of tail response, channel response, and/or nonlinear distortion.

13. The method ofclaim 5, wherein the plurality of constraints comprise spectrum mask limits, amplitudes early filter taps, amplitude of tail response, residual non-compensated early response, receive-side out-of-band rejection, noise mismatch, and/or nonlinear distortion tolerance.

14. The method ofclaim 1, wherein the estimation process is configured such that communications based on the total partial response provide data rates and error measurements that are at least equal to the data rates and error measurements associated with no inter-symbol interference (ISI) or near-zero inter-symbol interference (ISI) based communications subject to same spectrum mask limits for said partial response and said no ISI or near-zero ISI communications.

15. The method ofclaim 1, wherein the predetermined amount of inter-symbol interference (ISI) is greater than the maximum inter-symbol interference (ISI) practically allowed in reference communication using zero or near-zero ISI filters.

16. A system, comprising:

one or more circuits for performing a pulse-shaping filtering, wherein:

17. The system ofclaim 16, wherein the pulse-shaping filtering comprises use of a transmit-side pulse-shaping filter and a receive-side pulse-shaping filter.

18. The system ofclaim 16, wherein the one or more circuits are operable to configure the pulse-shaping filtering to optimize an applicable weighted distances for one or more selected error patterns for a given spectral compression.

19. The system ofclaim 16, wherein the one or more circuits are operable to configure the pulse-shaping filtering to optimize a symbol error rate (SER) associated with the communication of data.

20. The system ofclaim 16, wherein the one or more circuits are operable to configure the pulse-shaping filtering, the configuring comprising:

defining a plurality of performance-related variables;

setting a plurality of constraints; and