HK1117298A - Symbol time tracking for ofdm communication system - Google Patents

Description

Symbol time tracking for OFDM communication systems

Priority declaration in accordance with 35U.S.C. § 119

This application claims the benefit of U.S. provisional application serial No. 60/660,717 filed on 10.3.2005 and is a non-provisional application for this provisional application, which is assigned to the assignee herein and is hereby expressly incorporated herein by reference in its entirety for all purposes.

Background

The present invention relates generally to data or voice communications, and more particularly to synchronization in a communication system.

Orthogonal Frequency Division Multiplexing (OFDM) is a multi-carrier modulation technique that effectively partitions the overall system bandwidth into multiple (N) orthogonal frequency subbands. These subbands are also referred to as tones (tones), subcarriers, bins (bins), and channels. With OFDM, each subband is associated with a respective subcarrier that may be modulated with data, pilot, or overhead information.

In an OFDM system, a transmitter processes data to obtain modulation symbols and further performs processing on the modulation symbols to generate OFDM symbols. The transmitter then conditions and transmits the OFDM symbols over the communication channel. The OFDM system may utilize a transmission structure whereby data is transmitted in frames, each frame having a particular duration. Different types of data (e.g., traffic/packet data, overhead/control data, pilot, etc.) may be transmitted in different portions of each frame. The term "pilot" generally refers to data and/or transmissions that are known a priori by both the transmitter and the receiver.

The receiver typically needs to obtain accurate frame and OFDM symbol timing in order to properly recover the data transmitted by the transmitter. For example, in order to correctly recover the different types of data transmitted in each frame, the receiver may need to know the beginning of each frame. The receiver often does not know the time at which each OFDM symbol was transmitted by the transmitter, nor the propagation delay introduced by the communication channel. In order to correctly perform complementary OFDM demodulation on the received OFDM symbols, the receiver should need to determine the timing of each OFDM symbol received over the communication channel.

Synchronization refers to the process performed by the receiver to obtain the frame and OFDM symbol timing. The receiver may also perform other tasks such as frequency error estimation and channel estimation. Synchronization may occur at different times to improve timing and correct for variations in the channel. Because sudden changes in the channel are unlikely, wireless systems can change their timing coherently.

Often the channel experiences varying delays and multipaths. Different reflections or paths of the signal may arrive at the receiver at different times and have different magnitudes. Fading affects the size of the received signal. The delay spread is the difference between the First Arriving Path (FAP) and the Last Arriving Path (LAP). The LAP may not be the last actual reflection received, but the last reflection that meets some time delay constraint and/or size criteria. If both FAP and LAP can be estimated correctly and the OFDM symbol timing adjusted accordingly, then most of the received signal reflections can be used constructively for data demodulation.

SUMMARY

In one aspect, the present disclosure provides a method for characterizing a communication channel. The detection window is moved across a channel profile (profile) to aggregate tap energies in the channel profile within the detection window into an accumulated energy curve. The peak of the maximum in the accumulated energy curve is determined. A frequency band is defined relative to the accumulated energy curve. A First Arriving Path (FAP) is detected using a trailing edge found near a second end of the region in the accumulated energy curve. A leading edge found near a first end of a region in the accumulated energy curve is detected. The leading edge is used to determine the Last Arrival Path (LAP). The band defines a region of the accumulated energy curve at or near a maximum within the band.

In one aspect, the present disclosure provides a receiver for characterizing a communication channel. The receiver includes: means for moving a detection window across a channel profile (profile) to accumulate tap energies in the channel profile within the detection window into an accumulated energy curve; means for determining a peak value of a maximum value in the accumulated energy curve; means for defining a frequency band relative to the accumulated energy curve; means for detecting the FAP using a trailing edge found near a second end of the region in the accumulated energy curve; means for detecting a leading edge found near a first end of a region of the accumulated energy curve; and means for determining the LAP using the leading edge. The band defines a region of the accumulated energy curve at or near a maximum within the band.

In one aspect, the present disclosure provides a communication device for characterizing a communication channel. The communication device includes a processor and a memory coupled together. The processor is configured to: moving a detection window across a channel profile (profile) to accumulate tap energies in the channel profile within the detection window into an accumulated energy curve; determining a peak value of the maximum value on the accumulated energy curve; defining a frequency band relative to the accumulated energy curve; detecting a First Arriving Path (FAP) using a trailing edge found near a second end of a region in the accumulated energy curve; detecting a leading edge found near a first end of a region in the accumulated energy curve; and determining a Last Arrival Path (LAP) using the leading edge. The band defines a region of the accumulated energy curve at or near a maximum within the band.

Further areas of applicability of the present invention will become apparent from the detailed description provided hereinafter. It should be understood that the detailed description and specific examples, while indicating various embodiments, are intended for purposes of illustration only and are not intended to necessarily limit the scope of the disclosure.

Brief Description of Drawings

The present disclosure is described in connection with the accompanying drawings:

FIG. 1 is a block diagram of an embodiment of a base station and a wireless receiver in an Orthogonal Frequency Division Multiplexing (OFDM) system;

fig. 2A, 2B, and 2C are block diagrams of embodiments of superframe structures with increasing levels of detail;

FIG. 3 is a block diagram of an embodiment of an OFDM modulator;

FIG. 4 is a block diagram of an embodiment of a synchronization and channel estimation unit;

FIG. 5 is a block diagram of an embodiment of an OFDM symbol timing detector and channel estimator;

FIG. 6 is a block diagram of an embodiment of a symbol timing detector;

FIG. 7 is a block diagram of an embodiment of a time filter unit;

FIG. 8 is a diagram of an embodiment of a pilot transmission scheme with both TDM and FDM pilots;

FIGS. 9A, 9B, and 9C are block diagrams illustrating embodiments of three receive signal paths for OFDM symbols having different delay spreads;

FIGS. 10A and 10B are graphs of processing a channel profile to determine an accumulated energy curve;

FIG. 11 is a graph illustrating an example of the effect of timing drift on the resulting channel tap energies;

FIG. 12 is a chart representing an embodiment of a search window for a programmable channel arrangement;

FIG. 13 is a block diagram of an embodiment of a portion of a receiver; and

fig. 14 is a flow diagram of a method for characterizing a communication channel.

In the drawings, similar components and/or features may have the same reference numerals.

Detailed Description

The ensuing description provides preferred exemplary embodiments only, and is not intended to limit the scope, applicability, or configuration of the invention. Rather, the description of the preferred exemplary embodiments will provide those skilled in the art with a description of the preferred exemplary embodiments for practicing the present invention. It being understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope of the invention as set forth in the appended claims.

In the following description, specific details are given to provide a thorough understanding of the embodiments. However, it will be understood by those of ordinary skill in the art that the embodiments may be practiced without these specific details. For example, circuits may be shown in block diagrams in order not to obscure the embodiments in unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail in order to avoid obscuring the embodiments.

Also, it is noted that the embodiments may be described as a process which is depicted as a flowchart, a flow diagram, a data flow diagram, a structure diagram, or a block diagram. Although a flowchart may describe the operations as a sequential process, many of the operations can be performed in parallel or concurrently. In addition, the order of the operations may be rearranged. A process is terminated when its operations are completed, but the process may have additional steps not included in the figure. A process may correspond to a method, a function, a procedure, a subroutine, a subprogram, etc. When a procedure corresponds to a function, its termination corresponds to a return of the function to the calling function or the main function.

Moreover, as described herein, the term "storage medium" may represent one or more devices for storing data, including Read Only Memory (ROM), Random Access Memory (RAM), magnetic RAM, core memory, magnetic disk storage media, optical storage media, flash memory devices, and/or other machine-readable media for storing information. The term "machine-readable medium" includes, but is not limited to portable or fixed storage devices, optical storage devices, wireless channels and various other mediums capable of storing, containing or carrying instruction(s) and/or data.

Furthermore, embodiments may be implemented by hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware or microcode, the program code or code segments to perform the necessary tasks may be stored in a machine-readable medium such as a storage medium. The processor may perform the necessary tasks. A code segment or machine-executable instruction may represent a procedure, a function, a subprogram, a program, a routine, a subroutine, a module, a software package, a class, or any combination of instructions, data structures, or program statements. A code segment may be coupled to another code segment or a hardware circuit by passing and/or receiving information, data, arguments, parameters, or memory contents. Information, arguments, parameters, data, etc. may be passed, forwarded, or transmitted via any suitable means including memory sharing, message passing, token passing, network transmission, etc.

The synchronization techniques described herein may be used for various multicarrier systems, for the downlink as well as the uplink, and for broadcast systems. The downlink (or forward link) refers to the communication link from the base stations to the radio receivers, and the uplink (or reverse link) refers to the communication link from the radio receivers to the base stations. For clarity, these techniques are described below for the downlink in an Orthogonal Frequency Division Multiplexing (OFDM) or Orthogonal Frequency Division Multiple Access (OFDMA) system. The pilot detection structure is well suited for broadcast systems, but may also be used for non-broadcast systems. In a broadcast topology, forward-linlcing will be transmitted by a single base station and received by several wireless receivers. In one embodiment, the forward link may dedicate some channels to a single wireless receiver, a subset of multiple wireless receivers, or all wireless receivers.

An improved method and system for timing synchronization after initial acquisition in an OFDM system is disclosed. Fine Timing Acquisition (FTA) may be superior to the performance of Data Mode Time Tracking (DMTT) described in this disclosure. The result of initial time acquisition based on Time Division Multiplexed (TDM) pilot-1 processing is a coarse timing estimate. This coarse timing estimate provides information about the start of the superframe and gives a coarse estimate of the start of TDM pilot 2. By means of a further timing estimation with the TDM pilot-2 structure, the receiver estimates a more accurate starting position of the following OFDM symbol. This step is called FTA.

Once the FTA is made, ongoing timing corrections in DMTT mode keep the receiver synchronized even though the channel may be temporarily fading, experiencing a wide range of delay spreads, experiencing newly emerging energy concentrations, or other problems. The DMTT may use TDM or Frequency Division Multiplexed (FDM) pilots, but the following description refers primarily to FDM pilots, although applicable to TDM pilots. Synchronization involves not only detecting the First Arriving Path (FAP) for the channel, but also finding the best position for the FFT collection window to capture the most useful energy from the channel. In one embodiment, the process is designed to successfully process 1024 sample channel estimates with a delay spread of up to 768 chips (chip).

In one embodiment, the DMTT correction depends on the FAP location and the estimated channel delay spread D. The time tracker unit first finds the position T where the maximum accumulated energy occurs_MAnd storing the maximum value E_M. Next, T is checked_MTo locate the drop of the accumulated energy below the value (1-b) E_MFor some predetermined value b less than 1. In other words, the leading and trailing edges of the flat region in the accumulated energy curve are defined as the positions on the detection window where the accumulated energy is a few percent (e.g., 5% or 3%) away from its maximum value. The percentage defines a band around the maximum of the cumulative energy curve. Entering the band defines the leading edge T of the flat portion in the band_LAnd a trailing edge T defining a flat portion in the strip away from the strip_T. The trailing edge coincides with the location of the FAP, and the leading edge equals the Last Arriving Path (LAP) minus N_W. The difference between the leading edge and the trailing edge is equal to N_WThe delay spread D is subtracted. Thus, the delay spread may be calculated as D ═ N_W-T_T-T_L. Once at least two of FAP, LAP or D have been calculated, DMTT is performed along with the layout of the FFT collection window.

Referring initially to fig. 1, a block diagram of an embodiment of a base station 110 and a wireless receiver 150 in an OFDM system 100 is shown. The base stations 110 are typically fixed stations and may also be referred to as Base Transceiver Systems (BTSs), access points, or some other terminology. The wireless receiver 150 may be fixed or mobile and may also be referred to as a user terminal, a mobile station, or some other terminology. The wireless receiver 150 may also be a portable unit such as a cellular phone, a handheld device, a wireless module, a Personal Digital Assistant (PDA), a television receiver, and so forth.

At base station 110, a TX data and pilot processor 120 receives different types of data (e.g., traffic/packet data and overhead/control data) and processes (e.g., encodes, interleaves, and maps) the received data to generate data symbols. As used herein, a "data symbol" is a modulation symbol for data, a "pilot symbol" is a modulation symbol for pilot, and a modulation symbol is a complex value for a point in a signal constellation for a modulation scheme (e.g., M-PSK, M-QAM, etc.). Pilot processor 120 also processes pilot data to generate pilot symbols and provides the data and pilot symbols to an OFDM modulator 130.

As described below, the OFDM modulator 130 multiplexes the data and pilot symbols into appropriate subbands and symbol periods, and further performs OFDM modulation on the multiplexed modulation symbols to generate OFDM symbols. In this embodiment, one OFDM symbol consists of 4096 modulation symbols, one subcarrier per modulation symbol, in the frequency domain. A transmitter (TMTR) unit 132 converts the OFDM symbols into one or more analog signals and further conditions (e.g., amplifies, filters, frequency upconverts, etc.) the analog signals to generate a modulated signal. Base station 110 then transmits the modulated signal from antenna 134 to a wireless receiver in OFDM system 100. In the time domain for this embodiment, each OFDM symbol period is 4096+512+ 17-4625 samples long.

At wireless receiver 150, the transmitted signal from base station 110 is received by antenna 152 and provided to receiver unit 154. Receiver unit 154 conditions (e.g., filters, amplifies, frequency downconverts, etc.) the received signal and digitizes the conditioned signal to obtain a stream of input samples. An OFDM demodulator 160 performs OFDM demodulation on the input samples to obtain received data and pilot symbols. OFDM demodulator 160 also performs detection (e.g., matched filtering) on the received data symbols with the aid of channel estimates (e.g., frequency response estimates) to obtain detected data symbols, which are estimates of the data symbols transmitted by base station 110. OFDM demodulator 160 provides detected data symbols to a Receive (RX) data processor 170.

As described below, a synchronization/channel estimation unit (SCEU)180 receives input samples from receiver unit 154 and performs synchronization to determine frame and OFDM symbol timing. SCEU180 also derives channel estimates using pilot symbols received from OFDM demodulator 160. SCEU180 provides OFDM symbol timing and channel estimates to OFDM demodulator 160 and may provide frame timing to RX data processor 170 and/or controller 190. OFDM demodulator 160 utilizes the OFDM symbol timing to perform OFDM demodulation and channel estimation to perform detection on the received data symbols.

An RX data processor 170 processes (e.g., symbol demaps, deinterleaves, decodes, etc.) the detected data symbols from OFDM demodulator 160 and provides decoded data. RX data processor 170 and/or controller 190 may utilize the frame timing to recover different types of data transmitted by base station 110. In general, the processing by OFDM demodulator 160 and RX data processor 170 is complementary to the processing by OFDM modulator 130 and TX data and pilot processor 120, respectively, at base station 110.

Controllers 140, 190 control the operation at base station 110 and radio receiver 150, respectively. The controller may be a processor and/or a state machine. Memory units 142, 192 provide storage for program codes and data used by controllers 140 and 190, respectively. The memory units 142, 192 may use various types of storage media to store information.

The base station 110 may send point-to-point transmissions to a single wireless receiver, multicast transmissions to a group of wireless receivers, broadcast transmissions to all wireless receivers under its coverage area, or any combination thereof. For example, base station 110 may broadcast pilot and overhead/control data to all wireless receivers under its coverage area. In various different situations and embodiments, the base station 110 may further unicast transmission of user-specific data to a particular wireless receiver, multicast data to a group of wireless receivers, and/or broadcast data to all wireless receivers.

With respect to fig. 2A, 2B, and 2C, an embodiment of a superframe structure 200 that may be used for the OFDM system 100 is shown. Data and pilot may be transmitted in superframes 204, with each superframe 204 having a predetermined duration. The superframe 204 may also be referred to as a frame, a slot, or some other terminology. In this embodiment, each super-frame 204 includes a TDM pilot 1 field 212 for the first TDM pilot, a TDM pilot 2 field 214 for the second TDM pilot, an overhead field 216 for overhead/control data, and a data field 218 for traffic/packet data.

The four fields 212, 214, 216, 218 for each superframe 204 hold data. Various allocation schemes may be used, such as burst-TDM, periodic-TDM, and/or burst-TDM/FDM. In one embodiment, the four fields 212, 214, 216, 218 may also be arranged to facilitate synchronization and data recovery. The pilot TDM symbols 212, 214 transmitted first in each super-frame 204 may be used for detection of the open-ended OFDM symbols in the overhead field 216 transmitted next in that super-frame 204. In addition, the TDM pilot fields 212, 214 may be used for timing acquisition of the OFDM signal. The overhead information obtained from the overhead field 216 may then be used for recovery of traffic/packet data in the data field 218 that was last transmitted in the super-frame 204.

In an embodiment, TDM pilot 1 field 212 carries one OFDM symbol for TDM pilot 1 and TDM pilot 2 field 214 carries one OFDM symbol for TDM pilot 2. In general, each field may have any duration, and the fields may be arranged in any order. In some embodiments, TDM pilot 1212 and/or TDM pilot 2214 may be periodically broadcast in each superframe 204 to facilitate synchronization of the wireless receivers.

The OFDM system 100 has an overall system bandwidth of BW MHz, which is divided into N orthogonal subbands using OFDM. The spacing between adjacent subbands is BW/N MHz. Of the N total subbands, M subbands may be used for pilot and data transmission, where M < N, and the remaining N-M subbands may be unused and serve as guard subbands. In one embodiment, OFDM system 100 utilizes an OFDM structure with a total of N4096 subbands, M4000 usable subbands, and N-M96 guard subbands. In addition, FDM pilot symbols are interleaved in each OFDM symbol 226, e.g., an FDM pilot symbol is inserted into the data symbols every seven subbands, so that 500 FDM pilot symbols and 3500 data symbols are available outside of the guard subbands. In general, any OFDM structure with any total number of subbands, any number of usable subbands, and any number of guard subbands may be used for OFDM system 100.

The data field 218 includes several frames 222, which are described in detail in FIG. 2B. In this embodiment, there are four frames 222 per data field 218, but different embodiments may use more or fewer frames. Each frame 222 includes a number of OFDM data symbols 226 as shown in fig. 2C. In one embodiment, each OFDM data symbol 226 includes 3500 data symbols after accounting for the unused guard subbands and the removed FDM pilot symbols.

TDM pilot 1212 and TDM pilot 2214 are designed to facilitate, among other things, synchronization of wireless receivers in OFDM system 100. The wireless receiver can detect the beginning of each frame using the TDM pilots 1212, obtain a coarse estimate of the OFDM symbol timing, and estimate the frequency error. The TDM pilot 2214 may be utilized by a wireless receiver to obtain more accurate or fine timing acquisition. FDM pilot symbols interleaved within the data symbols may further allow synchronization timing to optimize capture of signal energy. In particular, the FDM pilots may be used for channel estimation, which may be used to optimize the capture of signal energy and ultimately readjust OFDM symbol timing.

Referring next to fig. 3, a block diagram of an embodiment of OFDM modulator 130 of base station 110 is shown. A symbol-to-Subband mapping or multiplexer unit 510 receives data and pilot symbols from TX data and pilot processor 120 and maps the symbols to the appropriate subbands based on a Subband multiplexing control (Subband _ Mux _ Ctrl) signal from controller 140. Symbol-to-subband mapping unit 510 provides one data or pilot symbol on each subband used for data or pilot transmission and a "zero symbol" (having a signal value of zero) for each unused subband in each OFDM symbol period. For each OFDM symbol period, symbol-to-subband mapping unit 510 provides N modulation symbols for a total of N subbands, where each modulation symbol may be a data symbol, a pilot symbol, or a zero symbol.

An N-point Inverse Discrete Fourier Transform (IDFT) unit 520 receives N modulation symbols for each OFDM symbol period, transforms the N modulation symbols to the time domain using an N-point IDFT, and provides a "transformed" symbol that contains N time domain samples. Each sample is a complex value to be transmitted in one sample period. Instead of an N-point IDFT, an N-point Inverse Fast Fourier Transform (IFFT) may also be performed if N is a power of 2, which is typically the case.

A parallel-to-serial (P/S) converter 530 serializes the N samples for each transformed symbol. Cyclic prefix generator 540 then repeats a portion (or C samples) of each transformed symbol to form an OFDM symbol containing N + C samples. For example, in one embodiment, the cyclic prefix 1004 is the last 512 samples of an OFDM symbol. The cyclic prefix is used to combat inter-symbol interference (ISI) and inter-carrier interference (ICI), e.g., caused by long delay spread in a communication channel. Typically, the delay spread is the time difference between the FAP and LAP of the signal at the receiver 150. An OFDM symbol period (or simply "symbol period") is the duration of one OFDM symbol and is equal to N + C sample periods. In one embodiment, N4096 and C512, so the symbol period is 4608. In some embodiments there may be an inter-symbol guard band of 17 sample periods between OFDM symbols, such that the OFDM symbol period is 4625.

Referring next to fig. 4, an embodiment of a block diagram of the SCEU180 for the wireless receiver 150 is shown. In the depicted embodiment, the SCEU180 operates in a time tracking (or data mode) DMTT state. Within the SCEU180, a window placement unit 725 aligns the aligned samples according to OFDM symbol timing corrections and excludes redundant CPs 1004 from OFDM symbols using symbol timing information from the symbol timing detector 720 or DMTT unit. In this embodiment, the OFDM symbol is represented by 4096 samples after the window placement unit 725. The relevant 4096 samples from the OFDM symbol are in the FFT window placement unit 725 and sent to the N-point DFT unit to create a transformed OFDM symbol at the receiver with the relevant 4096 samples.

A frequency error estimator 712 receives the filtered input samples and determines a frequency error estimate in the received signal. The frequency error estimate is provided to a frequency correction unit 715 to perform frequency correction. The frequency error may be due to various sources such as differences in the frequencies of the oscillators of the base station and the wireless receiver, doppler shift, and the like. The filtered and frequency corrected input samples are generated by the frequency correction unit 715 using the frequency estimate. Channel estimation unit 730 receives the FDM pilots in the transformed symbols and derives channel estimates therefrom.

The channel estimates are primarily used to assist in data demodulation, but are also used to determine symbol timing for future OFDM symbols. The symbol timing detector 720 determines the symbol timing from the channel estimate and provides the timing information to the window placement unit 725. In an iterative manner, the window layout is affected by the previous channel estimates.

With respect to fig. 5, a block diagram of an embodiment of a symbol timing detector 720 and a channel estimator 730 are shown coupled together for performing embodiments of timing synchronization and channel estimation based on FDM pilots. Channel estimator 730 generates channel estimates in both the time and frequency domains based on the FDM pilots. The time domain channel estimate is used by symbol timing detector 720 in generating a new timing offset that is fed back to the channel estimation unit to affect the acquisition of the next time domain channel estimate. This timing offset is also used by the FFT window placement unit 725 and other circuitry in the receiving unit 150. This loop allows for iterative determination of timing offsets for use throughout receive unit 150.

The channel estimator 730 generates a channel impulse response in the time domain; that is, the channel estimator 730 is responsible for estimating the channel from the FDM pilot in both the time and frequency domains using the time filter unit 528. In this embodiment, the channel estimator 730 includes an N-point DFT514, a pilot demodulator 516, a zero extrapolation unit 517, an M-point IDFT518, a time filter 528, and an M-point DFT 532. The N-point DFT514 performs a 4096-point fourier transform on the OFDM symbol after, for example, the redundant information in the cyclic prefix is removed by the FFT window placement unit 725. Although data symbols are used elsewhere after the N-point DFT514, our discussion focuses on the 500 FDM pilots output from the DFT. The FDM pilots are demodulated in pilot demodulation unit 516 to produce 500 demodulated FDM pilots. A zero extrapolation unit 517 converts the 500 actual pilots into 512 extrapolated FDM pilots. The M-point IDFT518 generates time-domain channel observations using a 512-point inverse fourier transform based on the 512 extrapolated FDM pilots. Time domain channel observations may have aliasing.

The time filter 528 removes any possible aliasing by collecting channel observations over several consecutive OFDM symbols. This embodiment of the time filter 528 filters channel observations over three consecutive OFDM symbols, but other embodiments may perform averaging over more or fewer OFDM symbols. Through this process, three consecutive 512-sample long channel observations are combined in this embodiment into a 1024-sample long time-domain channel estimate. The timing offset is used to align the three consecutive channel observations.

Within the symbol timing detector 720, a channel averaging unit 508 and a time tracker block 520 are used to determine the symbol timing. Symbol timing detector 720 receives successive time domain channel estimates, which are a by-product of channel estimation unit 730, and processes the time domain channel estimates to track the signal and control the generation of future channel estimates by channel estimator 730. The location of the channel energy is determined by the time tracker 520 based on an analysis of the channel estimates generated with the FDM pilots.

Referring next to fig. 6, a block diagram of an embodiment of a symbol timing detector 720, the symbol timing detector 720 is used to help determine the location of the channel energy. This embodiment uses two stages of filtering, but other embodiments may have only one filter or even no filtering. The channel impulse response or time domain channel estimate is received sequentially one tap at a time and filtered by the short term averaging block 908. Short-term averaging utilizes the last few channel estimates to maintain a short-term average of the channel estimates. Typically, the averaged channel impulse response is within one frame period. After providing the short-term average to the long-term averaging block 912, the short-term averaging process is periodically cleared. In this embodiment, short-term averaging block 912 helps to discern useful channel information from background noise to more accurately identify the channel taps and smooth the average channel impulse response for further processing.

The interval timer 928 clears the short-term averaging block 608 after a delay 632 that allows the delay 632 to provide the results to the long-term averaging block 912 before clearing. In one embodiment, each frame 222 is followed by an interval timer trigger, such that channel estimates over one frame period are used in the short-term average block. During the clear operation, the output from the short term averaging block is disconnected from the long term averaging 912 by a switch. In some embodiments, the period of interval timer 928 is adjustable and may depend on the expected coherence time.

In this embodiment, the channel impulse response is 1024 taps long, but may be other sizes in other embodiments. The false channel estimate is filtered by a digital filter in the short-term averaging block 908, for example, an Infinite Impulse Response (IIR) is shown, but a Finite Impulse Response (FIR) filter may alternatively be used in other embodiments. Filtering the channel tap energies one after the other over time as in the short term averaging block 908 allows averaging the channel energy profile (profile) over time and also helps to emphasize the active portion of the channel with respect to background noise. The short-term averaged channel estimates are passed through an instantaneous detector to find instantaneous FAP and LAP, which correspond to the channel profile over the past few OFDM symbols.

In another filtering step, the long-term averaging block 912 obtains and filters the short-term averaged channel estimates against historical channel estimates. These historical channel estimates typically come from the channel profile of several previous frames 222 (also spanning one or more superframes 204). In any event, the long-term averaging block 912 uses a wider range of channel estimates than the short-term averaging block 908. An FIR or IIR filter is used to combine the historical channel estimates with the current short-term averaged channel estimates. In one embodiment, long term averaging blocks are used to remember less common channel characteristics that have occurred in the past and are likely to occur in the future. The long-term averaged channel estimates are passed to a trend detector 920 to find trends in FAP and LAP behavior.

The channel locator block 924 obtains the instantaneous FAP and LAP and the trend FAP and LAP to determine the offset Off, which is used to arrange the FFT collection window in the window placement unit 725 and to align the aligned channel observations in the time filter 528. The operation of the algorithm used in the offset determination and application is explained further below.

Referring back to fig. 5, the time tracker 520 may determine the FAP by searching for a drop in the peak of the accumulated energy curve 1050. The integration and peak detection may be performed by a sliding length of N_WIs performed across a channel impulse response profile or channel profile. At each detection window start position, the energy of all taps falling within the detection window 1016 is calculated to find the accumulated energy curve 1050. The accumulated energy curve 1050 is analyzed to determine the FAP and LAP by finding the leading and trailing edges of the plateau around the maximum of the accumulated energy curve 1050.

A plot of the accumulated energy in the channel taps at the beginning of the different detection windows 1016 is shown in fig. 10B below for the accumulated energy curve 1050 of one embodiment. The detection window 1016 is cyclically shifted to the right so that when the right edge of the detection window 1016 reaches the last tap, the window 1016 wraps around to the first tap. As such, the accumulated energy is collected for the same number of channel taps for each detection window 1016 start position throughout the channel impulse response taps of the channel profile 1030.

With respect to fig. 7, a block diagram of an embodiment of the time filter unit 528 and the M-point DFT532 is shown detailing these blocks shown at a high level in fig. 5. Fig. 7 illustrates operations performed on channel observations to obtain 512-point channel estimates in the frequency domain for the data interleave of interest.

In one embodiment, channel estimation is performed based on the 500-FDM pilot subcarriers present in each OFDM symbol. The collected FDM pilots are processed in N-point DFT unit 514, pilot demodulation unit 516, zero extrapolation and M-point IDFT units 517 and 518, respectively. Thereby obtaining time-domain channel observations corresponding to pilot interlaces. Fig. 7 illustrates operations performed by blocks 528 and 532 for interleaving data of interest for an FDM carrier on time domain channel observations that result in 512-point channel estimates in the frequency domain.

After the 512-point IFFT518, a phase ramp (phase ramp)604 is performed to account for the offset of the pilot interlace from the zero interlace. The 512 time-domain channel observations obtained at the end of the phase ramp 604 are then filtered with two different time filters to produce 1024-point channel estimates in the time domain. This improved resolution time domain channel estimate is a byproduct of the channel estimation unit 730 and is at the same time an input to the symbol timing detector block 720.

Two different 3-tap non-correlated time filters 612, 616 are used for the filtering operation. The filtering operation uses three additional 512 long buffers 608-two for storing pilot observations corresponding to previous and future OFDM symbols, and a third buffer is used to store the additional 512 channel estimates obtained from the second time filter 616 employed. The result of the first time filter 612 operation is written back to the long 512 buffer containing pilot observations corresponding to the earliest OFDM symbol, while the result of the second time filter 616 operation is written to the extra long 512 buffer employed for this purpose. This additional filtering operation depends on the number of symbols, which determines the position of the pilot interlaces.

In combining time domain channel observations from three consecutive OFDM symbols, time filtering unit 528 takes into account any timing offset (or correction) detected by time tracking block 520. This is because once the timing correction is applied to the FFT window placement unit 725, the corresponding time domain channel observations are no longer aligned at the input of the time filter unit 528, and alignment occurs within the time filter unit 528. The timing offset is applied to buffers corresponding to future and past OFDM symbol observations while the current OFDM symbol observation is assumed to have the correct timing. The offset is applied with temporal filtering unit 528 prior to combining the time-domain channel observations.

We get input from the time tracking block 520, which we call the new timing offset (newTimingOffset). We also maintain two registers, which we call Offset 1(Offset1) and Offset 2(Offset 2). Offset1 corresponds to the effective Offset that is applied to the buffer corresponding to the future OFDM symbol (h (n +1)), while Offset2 corresponds to the Offset that has to be applied to the buffer corresponding to the past OFDM symbol (h (n-1)).

Combining 638 is performed to form a vector of length 512. This vector represents equivalent time-domain channel observations of length 512, which correspond to a different (non-pilot) interlace. Next, the vector is transformed to the frequency domain using a 512-point FFT unit 650, and this equivalent frequency domain channel estimate is used to demodulate the data on the interlace under consideration.

When performing timing correction, the time filter unit 528 transitions from one time base to a new time base as channel conditions change. The time base for the FDM pilot of an OFDM symbol corresponds to the time base for the data in that OFDM symbol. Time filter 528 typically combines time-domain channel observations from three consecutive OFDM symbols (i.e., past, current, and future), but when transitioning, may only consider channel observations from those OFDM symbols on the same time base. Additionally, channel observations on another time base may be corrected to the current time base before use. In any event, when the time filter considers consecutive ODFM symbols, only channel observations that are, or are corrected to, the same time base are used.

With respect to fig. 8, an embodiment of a pilot transmission scheme with a combination of TDM and FDM pilots is shown. Base station 110 may transmit TDM pilot 1212 and TDM pilot 2214 in each super-frame 204 to facilitate initial and fine timing acquisition by wireless receiver 150. In this embodiment, the overhead for the TDM pilots 212, 214 is two OFDM symbol periods, which may be small compared to the size of the super-frame 204. The base station 110 may also send FDM pilots in most or some of the reserved subbands according to various different schemes.

Each group for an OFDM symbol period contains a sufficient number (L) in the group_fdm) To support channel estimation and frequency and time tracking for wireless receivers. The set of subbands for the FDM pilot symbol is a subset of all the subbands for the OFDM symbol. The subbands in each group may be uniformly distributed across a total of N subbands and uniformly spatially separated by S_fdm＝N/L_fdmThe sub-bands are separated. Different sets of subbands may be used for different OFDM symbol periods, such that adjacent OFDM symbols have different sets of subbands. Also, the subbands in one group may be staggered or offset relative to the subbands in the other group such that the subbands in the two groups are interleaved with each other without overlap. Thus, each set of mutually disjoint and non-overlapping subbands described above is collectively referred to as an "interlace". As an example, N4096, L_fdm＝512，S_fdmThus, there are 8 interlaces per OFDM symbol, and each interlace consists of 512 subbands. In general, any number of interlaces (groups of subbands) may be used for the FDM pilot and each group may contain any number of subbands in the total of N subbands. In one embodiment, a single interlace (consisting of 512 subbands including guard subbands) is used for the FDM pilot.

Wireless receiver 150 utilizes the FDM pilot for channel estimation, time tracking, and/or possibly frequency tracking. The wireless receiver may obtain an initial channel estimate based on the pilot-2 OFDM symbol 214. The wireless receiver may utilize FDM pilots to increase the accuracy of channel estimation within the superframe 204. The wireless receiver 150 may also utilize the FDM pilot to update a frequency tracking loop that can correct for frequency errors in the received signal. Wireless receiver 150 may further update the time tracking loop with the FDM pilot (after conversion to time domain channel estimation by channel estimation unit 730) and place FFT collection window 1012 at an offset based on the observed channel position and delay spread (e.g., due to changes in the channel impulse response of the communication channel).

For the embodiment shown in FIG. 8, the FDM pilot is sent at 8 interleaving intervals such that each eight subband includes seven data symbols and one FDM pilot symbol. In this embodiment, the position of the interleaved FDM pilots is staggered from one OFDM symbol to the next OFDM symbol. If the pilots for OFDM symbol period m are placed on interlace 2, they will be placed on interlace 6 in OFDM symbol m + 1.

Interleaving allows channel estimation to utilize twice the actual FDM pilot subbands for converting them to the time-domain channel impulse response. The channel estimation block assumes that the conditions (channel, etc.) on successive OFDM symbols are stationary. Channel observations from OFDM symbol m-1 with FDM pilots on interlace 6 are combined with those from OFDM symbol m with pilots on interlace 2 and OFDM symbol m +1 with pilots back on interlace 6. Through this process, an effect is created that is similar to having FDM pilots on interlace 2 and interlace 6 simultaneously for a total of twice the actual number of FDM pilots. For example, for a given OFDM symbol period with 512 FDM pilots, the channel estimation block 730 doubles those FDM pilots with an adjacent OFDM symbol period to have 512 actual FDM pilots and 512 imaginary FDM pilots.

Referring next to fig. 9A, 9B, and 9C, an embodiment of three receive signal paths shown for one OFDM symbol at different delay spreads is shown. Each OFDM symbol includes a cyclic prefix 1004 of C samples and a transformed symbol 1008 of N samples. In this embodiment, the OFDM symbol is shown received on three paths, where each path has a different size and time shift. In some embodiments, paths having OFDM symbols below a predetermined size may be ignored. For example, in fig. 9A it is shown that there are many more paths than the three paths, but the smaller size paths are ignored when characterizing the channel position.

The difference between FAP and LAP is the delay spread D. In one embodiment, for example, cyclic prefix 1004 is 512 samples long and delay spread is 490 chips (c)hip). Placing D relative to the FFT Collection Window 1012_midAnd determining D by analyzing FAP, LAP and/or delay spread_mid。D_midIs the distance between the start of the FFT collection window 1012 for the current OFDM symbol and the desired channel center for the next OFDM symbol. The offset is used to adjust the position of a collection window 1012 of N samples (e.g., 4096 samples) between the current OFDM symbol and the next OFDM symbol. For example, the collection window 1012 defines the relevant portion of the incoming signal transformed to the frequency domain by the FFT 514.

The collection window 1012 is positioned to capture the signal portion of the most useful energy of the package. As described below, at least two of the FAP, LAP, and delay spread are determined in order to characterize the channel location. The FAP, LAP, and delay spread may be current metrics, time-averaged metrics, and/or worst-case metrics. To place the collection window 1012, the beginning of the collection window 1012 may be arranged such that subsequent channel estimates surround the value D_midAround the value D of_midIs programmable. In one embodiment, D_midIs set to a value of about half the length of the cyclic prefix 1004 (i.e., 256 samples in a cyclic prefix that is 512 samples long) and is measured from the beginning of the collection window 1012.

In the embodiment of FIGS. 9A, 9B and 9C, D_midPlaced in the middle of the delay spread for the next OFDM symbol, the collection window is referenced to D_midAnd (4) placing. As long as the delay spread D is less than the length of the cyclic prefix 1004, as is the case in fig. 9A and 9B, all of the signal energy collected within the FFT window corresponds to the desired OFDM symbol and can be constructively combined for data demodulation. In contrast, the delay spread in fig. 9C does not allow all of the energy collected within the FFT window to come from the desired OFDM symbol due to the large delay spread. In the embodiment of fig. 9A and 9B, the collection window 1012 is positioned at relative D_midBut in the embodiment of fig. 9C, the collection window 1012 is placed at the FAP_min。

The FAP distance is the beginning of the FFT collection window 1012 and the end of the cyclic prefix for the first pathA measure of time. The LAP distance is a measure between the beginning of the FFT collection window 1012 and the end of the cyclic prefix for the last path. D_mid' is D for the current OFDM symbol_midThe desired position of the device. D_midIs the next OFDM symbol period D_midThe desired position of the' device. In one embodiment, D_midIs placed somewhere between the FAP and LAP or at an intermediate point between the FAP and LAP. In other words, in the next OFDM symbol period, D_mid' become D_mid. Since the channel conditions in FIG. 9A have not changed, D_midAnd D_mid' generally in agreement with each other.

Fig. 9A and 9B illustrate examples of where the collection window 1012 may be placed to capture useful signal energy for an OFDM symbol. In both cases, the delay spread is smaller than the size of the cyclic prefix 1004. In these cases, the beginning of the collection window 1012 is placed within the intersection of the cyclic prefixes of all the arrival paths corresponding to the OFDM symbol of interest. The intersection is defined as the time period during which each signal path is receiving the cyclic prefix of the same OFDM symbol, which may be filtered to exclude paths that are weak in signal. In other words, the intersection begins at the beginning of the cyclic prefix for LAP and ends at the end of the cyclic prefix 1004 for FAP. In one embodiment, the collection window 1012 is placed in the middle of the intersection of the cyclic prefixes corresponding to the first and last arrival paths, as long as the intersection is a non-empty set. Typically the FFT collection window 1012 is placed in such a way that subsequent (future) channel estimates occur at D_midWhen D is in_mid' different from D_midThen the change is possible. In an iterative manner, at D_mid' different from D_midWhere D is corrected by an offset_mid’。

In FIG. 9A, no change in channel position results in D_mid' and D_midAre generally identical to each other. However, this is not the case in FIG. 9B, since at D_midAnd D_midThere is an Offset between'. The difference between fig. 9A and 9B is the collection window 1012 have shifted to the right with respect to the channel position so that D_midNo longer with D_midAnd (5) the consistency is achieved. The Offset is passed from the time tracker 520 to the time filter 528 so that the position of the collection window 1012 for the next OFDM symbol can be adjusted. The collection window 1012 for the next symbol is moved to the left and away from D_midAnd orientation D_mid', as they exist in the current OFDM symbol. In this manner, the beginning of the receive window 1012 may be maintained at the intersection of the cyclic prefixes 1004 for all paths of interest.

When the delay spread exceeds the length of the cyclic prefix 1004, as is the case in fig. 9C, it is no longer possible to eliminate paths from other OFDM symbols that are part of the collection window 1012. In these cases, the collection window 1012 is placed the smallest FAP before the end of the estimate of the cyclic prefix 1004 for the FAP corresponding to the current OFDM symbol is the FAP_minTo (3). One or more past OFDM symbols are used to predict where the cyclic prefix 1004 will end for the current OFDM symbol. In one embodiment, the FAP is for a cyclic prefix 1004 of 512 samples_minIs 24 samples. In other embodiments, FAP_minMay be about 0%, 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, or 10% of the length of the cyclic prefix 1004.

With respect to fig. 10A and 10B, these figures represent the processing of channel tap energies to determine the accumulated energy. In fig. 10A, the detection window 1016 moves through the impulse response tap energy to accumulate energy within the detection window 1016. In this embodiment, the tap energies are limited to the short-term averaging block 908 and/or the long-term averaging block 912 such that the trend FAP and LAP and/or the instantaneous FAP and LAP may be determined using various techniques including a sliding detection window 1016 and other steps described below.

As shown in fig. 10B, the accumulated energy curve 1050 is formed with the accumulated energy within the detection window 1016 as the detection window 1016 in fig. 10A traverses the channel profile 1030. From the accumulated energy curve 1050, the delay spread, FAP and LAP can be determined. By knowing the delay spreadAny two of the extension, FAP, or LAP may determine the missing one. In this embodiment, a vector of channel tap energies is used as an input to the DMTT algorithm, with N_CThe estimation is done 1024 points. However, this need not be the case in other embodiments. All of the lengths and sizes described herein may be scaled down appropriately if the resolution of the DMTT algorithm is reduced. This is achieved by combining N_CThe energy of several adjacent taps (1024-long channel estimation) to get a lower resolution (shorter) channel estimation. In another embodiment, for example, eight adjacent taps may be combined and the lower resolution N_C1＝128。

The tracking power or resolution of time tracking algorithms often depends on the length of the channel estimate, i.e., N_C. If all N of the channels are cyclically convolved_C4096 time domain taps are available, then the time tracking resolution is typically at its maximum. In this case, it is possible to uniquely determine what the amount of offset introduced in the FFT collection window 1012 position relative to the OFDM symbol boundary is. However, in most practical cases, the length of the channel estimation is limited by the number of FDM subcarriers used for channel estimation. For example, there are a (2, 6) pilot interlace pattern at the position shown in FIG. 8, a number N of time domain channel taps available after zero extrapolation and interpolation of 500 useful FDM subcarriers in an OFDM symbol, and averaging channel observations over the OFDM symbols_C1024. Staggered pilots increase the resolution of channel estimation, e.g., in one embodiment, as described above, with L per OFDM symbol_fdm512 pilots, and N_C1024 channel taps (also per OFDM symbol).

The time tracking capability in this embodiment depends on the actual time tracking algorithm implemented. In one embodiment, to improve the ability to detect channel changes, the algorithm uses information about past channel placements. Assuming that the maximum non-zero delay spread of the channel is D_MAXWherein no further knowledge about the channel is available, as long as D_MAX＞N_C2, cannot solveAnd to resolve ambiguities in channel placement. However, when assuming that the FAP and LAP information was estimated correctly in the past, the absolute tracking capability is extended to a total of N_C-D_MAXAnd (4) a position. In other words, it is assumed that the channel may change its position equally in both directions (i.e., the channel content may appear equally before and after the current timing reference). Then, the channel position in the future may be as much as + - (N) away from the current timing reference_C-D_MAX) A/2 chip position. This is illustrated in fig. 11, which shows the effect of timing drift on the resulting channel tap energy.

One factor in tracking capability is the estimated channel delay spread D instead of D_MAX. In one embodiment, when information about the channel delay spread is available, the total number of feasible channel positions is increased to N_C-D. Delay spread estimate D and algorithm pairs longer than N_CResistance to delay spread of/2 taps may lead to a modification of the FAP detection method. For ease of labeling, we introduce the terms "positive search region" or "negative search region", both of which are shown in FIG. 11. The positive search area is a portion of the area outside of the non-zero channel content (i.e., 0 and D in fig. 11)_MAXThe area in between) where it is assumed that recent channel content may be present. Similarly, paths detected in the negative search area are assumed to have moved a shorter distance than the previously observed channel content and, therefore, have occurred "earlier" in time. The introduction of search areas relative to previous channel content introduces memory (or causality) that makes it possible to enhance the tracking capabilities of the DMTT. The decision to produce the layout of the detected channel content within the maximum detection region and the layout of the boundary 1104 between the two search regions is explained next.

Timing synchronization is based on the channel estimate and the values in the accumulated energy curve 1050. Consider the channel profile 1030 shown in fig. 10A for timing synchronization. It is possible that the channel energy will be grouped into bins (bins), in which case the channel profile 1030 is coarser, and in FIG. 10AN_CIs lowered. For reasons of clarity below, we always assume that N_C1024 and explicitly introduce a length scaling factor of 2 when required^m. The timing search algorithm is performed on the long-term and/or short-term averaged channel estimate energy as shown with respect to fig. 6. During the time averaging process, one container at a time, the container with the largest channel energy is identified, i.e., n in FIG. 10A_MAX. Moreover, the maximum energy is storedUsing the E_MAXValue to determine a threshold T for noise thresholding_DMTTThe goal is to eliminate spurious taps in the channel estimate that do not correspond to true channel content.

Referring to fig. 10B, a cumulative energy curve 1050 resulting from sliding the energy detection window 1016 across the channel profile 1030 is shown. The examples of fig. 10A and 10B are intended to address the unique features of this approach. Selecting a detection window length N_WSuch that the complete channel profile 1030 can fit within the detection window 1016. Thus, N is selected in this embodiment_W≥D_MAXAnd N is_W768 samples, where the channel impulse response length (N)_C) Is 1024 samples, but may have other sizes in other embodiments. The location of the beginning of the detection window 1016, which includes the complete channel energy (or a substantial portion thereof), forms a relatively flat region 1040 of the accumulated energy curve 1050. The length of the flat zone 1040 is N_WD and assumes positive during which D is the actual channel delay spread. By estimating the boundaries of the flat zone 1040, it is possible to determine that (N) is_CLong channel estimate) of the channel energy of interest and the channel delay spread D. The flat zone 1040 is defined as a continuous portion of the accumulated energy curve 1050, where the curve 1050 is within a predetermined range from the maximum of the curve 1050.

In fig. 10B, the channel location is identified by the presence of FAP and LAP. Once the FAP and D of the current channel estimate are known, the timing correction or offset to be introduced into the future FFT collection window 1012 is such that the channel in the future OFDM symbol is in the maximum detection region D_MAXAt a predetermined position D_midAs the center. If the goal is to minimize the probability of an early path occurring before the maximum detection region while keeping the effective channel delay spread low, D is chosen_midIs D_MAX/2. Generally, selected for D_midThe value of (d) depends on the deployment area and is kept programmable. The resulting timing offset (offset) can be calculated as:

note that the second term in equation (1) corresponds to a variable compensation that depends on the estimated "headroom" between the maximum allowed channel length and the current estimated channel length, and on the optimal placement of the channel that takes into account robustness issues in future OFDM symbols. In other words, equation (1) for calculating the timing offset to be applied to the FFT collection window 1012 results in shifting the time domain channel content to point D in future OFDM symbols_midAs the center. In this embodiment, once it reaches the tip overThe value (roll-over value)4625, the calculated arbitrary offset moves the collection window 1012. The above procedure applies to a single instance of timing decision.

In other embodiments, timing decisions may be done independently in Hardware (HW) and/or Software (SW), where HW decisions are based on short-term average or instantaneous channel estimates and SW decisions are based on long-term average or trend channel estimates. Other embodiments may utilize SW or HW interchangeably to perform short or long term averaging. Each decision instance (HW and SW) then makes a decision on the channel location, i.e., FAP and LAP. These decisions may then be combined in one embodiment of the channel locator block 924, and as such

FAP＝min(FAP_HW，FAP_SW) And LAP ═ max (LAP)_HW，LAP_SW)(2)。

As long as LAP-FAP is less than or equal to D_MAXThe value of equation (2) can be used. If a condition is violated, the fast filtered instantaneous value has priority over the slow filtered trend value. In other words, if LAP-FAP > D_MAXThen:

if FAP_HW＜FAP_SW*FAP_SW：＝FAP＝FAP_HW，LAP_SW：＝LAP＝FAP+D_MAX；(3)。

If LAP_HW＞LAP_SW*LAP_SW：＝LAP＝LAP_HW，FAP_SW：＝FAP＝LAP-D_MAXThe only remaining cases: LAP_HW-FAP_HW＞D_MAXOr LAP_SW-FAP_SW＞D_MAXIs processed as follows. The parameter D used in equation (1) can be found to be D: LAP-FAP.

Next, the actual algorithm for computing the FAP and LAP is described for one embodiment. The input to the algorithm is at N_C/2^mAverage channel energy | h (n) in a container²Where m may take 0 and m_max(in some embodiments m_max2 or 3). Energy quiltAveraged, but if below a value T that is programmable, before being used in the timing synchronisation block_DMTTThe channel tap is selected. The output of the algorithm is two integers, FAP and LAP. Note that in some embodiments the following algorithms in the steps listed below may be applied independently in the form of HW and/SW, and the results may be combined. Two variables describing the FAP/LAP detection algorithm: single-pass (single-pass) and two-pass (two-pass) algorithms. The single pass algorithm requires slightly less time to compute, but in one embodiment is slightly more complex in implementing logic. The two-pass algorithm is more straightforward to implement and can be used whenever computing time is not the most stressful resource.

Single-pass enhanced DMTT algorithm

"spread out" the channel profile 1030 to distinguish two regions-positive and negative search regions as shown in fig. 11 and 12. In some embodiments, the boundary point 1104 between the negative and positive search regions is another programmable parameter. Assuming that the new signal path may equally occur before or after the currently detected channel content, the boundary points 1104 are chosen (from the end of the channel estimation) such that the regions are of the same length; in other words, Δ_N＝(N_C/2-D_mid)/2^m. Thus, with T_DMTTAfter thresholding we have:

e(n)＝|h(N_C/2^m-Δ_N+n)|²for n < Δ > 0 ≦ n_N(ii) a And

e(n)＝|h(n-Δ_N)|²for a_N≤n＜N_C/2^mConceptually, to ensure that the slide length N is achieved by the implementation described below_WThe detection window 1016 to detect the leading and trailing edges of the flat region exemplified in fig. 10B, so that the spread channel estimate is zero-padded on both sides.

2. Setting an initial value: n is a radical of_W＝D_MAX/2^m，n＝0，E_uTwo steps δ with different resolutions, 0_E＝E_MAX/γ，Δ_E＝N_γ·δ_E(programmable parameters γ and N)_γ) (ii) a Setting three forward and one backward thresholds E_TF0＝E_MAX-2Δ_E，E_TF1＝E_MAX-Δ_E，E_TF，end＝E_TF1-δ_EAnd E_TB＝E_TF，end(ii) a Setting binary flag found_beg＝false，found_endFalse; initializing a buffer BEG of leading edge positions_buffIs long by N_γAll zeros of (c).

3.For0≤n＜N_WPerforming (do) the following in a single pass algorithm:

a)E_n＝E_n+ e (n); imagine thatAnd e (n) adding zero before.

b)If(E_n＞E_TF0And E_n＞E_TFI) Value exceeds threshold by a large number: in this embodiment the BEG_buffFill value N (N)_γSecond).

c)Else if(E_n＞E_TF0But E_n≤E_TF1) When the value just exceeds the threshold value, the current position n is once moved to the buffer BEG_buffIn (1). In both cases (b) and (c), both do:

E_TF0＝E_n+δ_E，E_TF1＝E_TF0+Δ_E，E_TB＝E_n-Δ_E，found_beg＝true。

d)Else if(found_beg＝true，found_endfalse and E_n＜E_TB)，do：END＝n，found_end＝true，E_TF，end＝E_TB+δ_E。

e)Else if(found_endTrue and E_n＞E_TF，end)，do：found_end＝false。

4.For N_W≤n＜N_C/2^m，do：[E_n＝E_n-e(n-N_W) + e (n); then repeating the above steps (b) - (e)]。

5.While found_end＝false，do：[n＝n+1，E_n＝E_n-e(n-N_W) Then step (d) is also performed]。

Finally, we obtain the output parameters that depend on the channel position and delay spread as:

LAP＝2^m·(BEG_buff[0]-Δ_N)，FAP＝2^m·(END-Δ_N-N_W)(4)

the algorithm described in steps 1-5 above has some or all of the following characteristics:

to construct the effective step size for the algorithm, a small fine step size δ is used_EAnd corresponding value N_γ(chosen as a power of 2) because of the small value delta_ETo help determine the absolute maximum, delta, of the cumulative energy profile with greater accuracy_EIs advantageous. Buffer BEG for leading edge position_buffEffectively acting as a coarse delta_EReturn from the maximum position and declare that this point is the contribution of the leading edge. The actual amount returned may depend on several factors.

The estimated leading edge BEG is placed by the algorithm ═ BEG_buff[0]Not later than the true beginning of the flat region 1040, toAnd the estimated trailing edge END is not earlier than the true END of the flat region 1040, as long as Δ_E1060 is greater than E in flat region 1040_nMaximum peak-to-peak fluctuation. The resulting flat area 1040 may be wider than the true flat area 1040. However, in most practical cases, at Δ_EAnd 2. delta_EAny position between the actual maximum points of the distance declares the leading edge, while the trailing edge is about below the maximum value Δ_E1060.

For values γ and N_γSome of the compromises of (a) include: from the viewpoint of accuracy, a large value is preferable. When gamma and the corresponding N_γTowards infinity, the maximum value maxE of the accumulated energy is determined more accurately_n. With N_γThe ambiguity of returning from the maximum (to determine the leading edge position) also increases. In one embodiment, the selection values γ 256 and N_γ8. These values are programmable in some embodiments.

And γ and N_γOf interest is the threshold T_DMTTThe value of (c). Thresholding is introduced in the symbol timing detection or DMTT block 720 to exploit the coherent combining gain of the time channel estimate averaging. The thresholding eliminates the positive bias (positive tilt in the accumulated energy curve 1050) seen in step 3 above due to zero padding. The threshold is not greater than the obtained equivalent step size Δ_EAnd according to a fine step size delta_EMultiple N of_thRemain programmable. In one embodiment, the selected value is N_thIs 4, therefore T_DMTT＝E_MAX/64。

In general, for gamma, N_γAnd N_thAnd 2^mSuitable values of (a) may be empirically or algorithmically derived, but stored in programmable registers. Also, the AGC setpoint (setpoint) is kept as a reference value instead of E_MAXIs possible.

As described above and in equation (3), when a particular channel estimate is detected in the timing decisions using the slow and fast averages 908, 912Where there is a difference, an alternative treatment may be applied. Similar counter-measures apply in the unlikely case that the above-mentioned independent (HW or SW) timing search algorithm returns unexpected results, e.g. LAP-FAP > D_MAX. In an embodiment, the values for FAP and LAP are independently limited to below D, regardless of the estimated leading and trailing edge positions_MAX. However, to avoid performance degradation over erroneous timing, it is proposed that LAP-FAP > D be detected_MAXSet up HW, and set up FAP ═ D in this case_MAX-N_C，LAP＝2D_MAX-N_C。

Using the above algorithm, channel parameters for channel estimation for short term average (HW) and long term average (SW) can be obtained. Final results of the respective algorithms: namely FAP_HW、LAP_HW、FAP_SW、LAP_SWAnd then combined and used for OFDM sample counter modification as explained earlier in equations (2) and (3) to locate the collection window 1012.

DMTT algorithm enhanced by two times

Assuming fast filter averaging channel estimate A_n(k) Information about FAP and LAP can be extracted using either a single-pass or a dual-pass DMTT algorithm. While single pass algorithms offer the benefit of faster processing times, the associated logic and HW resources can be more demanding.

According to a previously known timing (including a variable offset, depending on the length of the estimated useful channel content), included in A_n(k) The average channel estimates in (1) are aligned. In this embodiment, the averaged channel estimates are also not thresholded. One goal of the DMTT algorithm is to re-estimate the estimate a_n(k) And utilizes this information to update the FFT collection window 1012 placement in future OFDM symbols in order to establish desired performance and robustness to changing channel conditions. This re-estimation (timing estimation) is every N_udOne OFDM symbol is performed and includes the followingThe operation is as follows:

1. a break point is defined that separates the "future" from the "past" in the averaged channel estimates. Due to the limited number of staggered pilots (in this embodiment the channel estimate spans only 1024 time-domain chips), a hard decision is made as to the position of the starting position (the position of the minimum time) in the channel estimate. The estimates are then reordered in increasing chronological order.

2. According to the tap E selected as the largest time domain_MAXThe threshold value of a fraction of (2) is used to threshold the channel estimate. The uncancelled channel estimates are averaged in HW so that non-coherent combining of the noise taps can result in SNR gain of the channel estimates. To take advantage of this gain, thresholding is applied in DMTT block 720.

3. By sliding the size N on the thresholded channel estimate_WThe accumulated energy curve 1050 is calculated using the rectangular detection window 1016. The maximum of the cumulative energy profile is found. Based on preset forward and backward threshold coefficients epsilon_fAnd ε_bAnd calculating forward and backward thresholds for determining flat zone edges:

E_T，F＝E_n，MAX(1-ε_f)，E_T，B＝E_n，MAX(1-ε_b)

4. from corresponding to E_n，MAXPoint n of_MAXStarting and moving towards the edges of the cumulative energy profile, the farthest point at which the energy crosses the forward and backward thresholds is determined. In one embodiment, the forward and backward thresholds are E_n，MAX5%, 10%, 15% or 20%. These points determine the end and the beginning (respectively) of the flat region 1040. With these points, the start and end of the estimate of the non-zero channel content (FAP and LAP) are computed directly.

The input to the enhanced DMTT cell is N values A_n(k)、A_n(k) Maximum value of (E) namely_MAX(which may be determined by the block used for averaging), and every N in the normal operating mode by the FFT block_udOne OFDM code element generation"DMTT update request signal" once. The other "inputs" are the SW programmable parameters of the two-pass algorithm, given in Table 1 below.

Table 1: programmable parameters with default values for use by a two-pass DMTT algorithm

Programmable parameter	Description of the invention	Acceptable value
			NW	Length of window	At 512/NbAnd 768/NbBetween
Dmid	Intermediate point for centering channel estimate after timing correction	At 256/NbAnd 384/NbBetween
			ΔN	Cut-off position for unwrapping channel estimates	The nominal value is 512/Nb-DmidBut + -128/NbDeviation of (1) is possible (according to SW)
εf，εb	Forward and backward threshold coefficients	Division of 7 bitsA value between 0.02 and 0.1; nominal value: about 0.05, both are
			βT	Determining a noise threshold TDMTT＝EMAX/βT	The range of values is between 32 and 128; a good candidate is 64

These parameters may be written by software and while some of them will remain constant throughout modem operation, some of them may be modified on a frame basis, depending on a more sophisticated SW algorithm. Whenever the "DMTT update request signal" is set high, and based on other inputs, the enhanced DMTT unit produces two outputs, namely the estimated start and end of the non-zero channel content; these are two integers: FAP and LAP. Additional programmable parameters include minimum offset B_offAnd hard bound offset_max。

During the initialization phase, the relevant variables and buffers are initialized. They are used in the following stages. The first step is defined to include the length N1024/N_bBuffer A of_n(k) (also referred to as A)_n) The concept of the beginning and end of the channel estimation energy profile in (1). For energy accumulation purposes, the starting point in the buffer is defined as N- Δ_NAnd the last point is N-Delta_N-1; note that the subscripts increment modulo N. The address of the memory location is defined as: startIndex N-Delta_NAnd breakpoint position: breakpt ═ N- Δ_N+N_W]_modN. Accumulated energy E_nIs a 12-bit unsigned value (scale 2)⁵) Which in this embodiment is initialized to zero. The position counter n is initialized to zero (10 bits) and the values BEG and END contained at the leading and trailing edge positions, respectively, are the same (10 bits each). Channel based on averagingEstimated maximum tap E_MAXDetermining an unsigned 8 bit threshold T_DMTT＝E_MAX/β_T. Maximum cumulative energy value E_n，MAXAnd corresponding position n_MAXAre initialized to zero. Allocating storage for accumulated energy E_nIs (N + N)_WA long buffer containing 12 bit unsigned values). Finally, the binary flag is set: found_beg＝false，foud_end＝false。

After initialization is complete, the first pass of the DMTT algorithm is ready to begin. The result of this phase is to calculate the accumulated energies and locate their maximum.

For0≤n＜N，do：

f)e＝A_endIndex；if e＞T_DMTT，do：E_n＝E_n+ e (saturated back to l2 bits); endIndex [ endIndex +1 ]]_{mod N}；

g)if endIndex＞breakPt，do：e＝A_startIndex；if e＞T_DMTT，do：E_n＝E_nE (hold at 12 bits); startIndex [ startIndex +1 ]]_{mod N}(ii) a Save E at the corresponding location_n；

h)ifE_n＞E_n，MAXSetting up E_n，MAx＝E_nAnd n_MAX＝n。

For N≤n＜N+N_W，do：e＝A_startIndex；if e＞T_DMTT，do：E_n＝E_nE (hold at 12 bits); startIndex [ startIndex +1 ]]_modN(ii) a Save E at the corresponding location_n(ii) a After completion of the first pass, using E_n，MAXTo set forward and backward thresholds: e_T，F＝E_n，MAX·(1-ε_f)，E_T，B＝E_n，MAX·(1-ε_b). The threshold is maintained with a 12-bit unsigned value.

The second pass includes two parts: the backward search is used to find the leading edge BEG and the forward search is used to locate the trailing edge END.

Forn_MAx-1. gtoreq.n.gtoreq.0 (decreasing subscript), do: found_end＝false

a)if(found_beg═ false, and E_n＜E_T，B)，do：BEG＝n+1，found_beg＝true；

b)else if(found_begTrue, and E_n≥E_T，B)，do：found_beg＝false。

For n_MAx+1≤n＜N+N_W(increasing subscript), do:

a)if(found_end═ false, and E_h＜E_T，f)，do：END＝n，found_end＝true；

b)else if(found_endTrue, and E_n≥E_T，f)，do：found_end＝false。

Here, both the BEG and END should contain non-zero values and both binary flags should be set to true (true). If this is not the case, the timing offset value should not be modified. The channel positioning and delay spread dependent output parameters FAP and LAP are found to be:

LAP_HW＝BEG-Δ_N，FAP_HW＝END-Δ_N-N_W(5)。

alternative communication channel location algorithm

In addition to the above-described methods of determining the FAP, other methods may be used in some embodiments. In one embodiment, the leading and trailing edges of the flat region 1040 are found by scoring a weighted sum of the accumulated energy and its positive finite difference. The delay spread D may be determined after finding both the leading and trailing edges of the accumulated energy curve 1050. US patent application/, (attorney docket No. 040588), entitled "fine timing ACQUISITION" (FINE TIMING accuracy ACQUISITION), filed on even date herewith by reference, describes the use of detecting spikes (spikes) in the negative difference (negative differential) of the accumulated energy curve to determine the trailing edge of a flat region, which can also be used to find the leading edge of the flat region. Once these edges are found, the delay spread can be determined according to the above description. Having found these communication channel positioning parameters, they may be used in the layout of the channel estimation unit 720 and/or the collection window 1012 as described above.

Referring next to fig. 13, the present disclosure provides an embodiment of a receiver 1300 for characterizing a communication channel. The receiver includes: means 1304 for acquiring a plurality of pilot symbols; means 1308 for determining a channel profile from the plurality of pilot symbols; means 1316 for moving a detection window across the channel profile to accumulate tap energies in the channel profile within the detection window into an accumulated energy curve; and means for analyzing the accumulated energy curve to characterize the communication channel 1312. The means for analyzing 1312 includes means for determining 1320 a peak of a maximum in the accumulated energy curve; means 1324 for defining a frequency band relative to the accumulated energy curve; means 1328 for detecting the FAP using a trailing edge found near a second end of the region in the accumulated energy curve; means 1332 for detecting a leading edge found near a first end of a region of the accumulated energy curve; and means 1336 for determining LAP using the leading edge. The frequency band is arranged relative to the maximum and defines a region of the accumulated energy curve at or near the maximum within the frequency band.

With respect to fig. 14, an embodiment of a method 1400 for characterizing a communication channel is disclosed. At block 1404, the detection window 1016 is moved across the channel profile 1030 to accumulate tap energies in the channel profile within the detection window into an accumulated energy curve 1050. The peak of the maximum in the accumulated energy curve is determined at block 1408. A frequency band 1060 is defined relative to the accumulated energy curve at block 1412. A First Arriving Path (FAP) is detected at block 1416 using a trailing edge found near a second end of the region 1040 in the accumulated energy curve 1050. A leading edge is found near the first end of region 1040 of accumulated energy curve 1050. The leading edge is used to determine the Last Arrival Path (LAP) in step 1420. The frequency band 1060 defines a region 1040 of the accumulated energy curve at or near the maximum within the frequency band 1060.

The synchronization techniques described herein may be implemented in various ways. For example, these techniques may be implemented in hardware, software, or a combination thereof. For a hardware implementation, the processing units used at the base station to support synchronization (e.g., TX data and pilot processor 120) may be implemented within one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), processors, controllers, micro-controllers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof. The processing unit (e.g., SCEU 180) used at the wireless receiver to perform synchronization may also be implemented within one or more ASICs, DSPs, and the like.

Some of the above embodiments determine the leading and trailing edges of the flat region in a particular manner. Other embodiments may score a weighted sum of the tap energies and the finite difference of the maximum tap energy. The start and end of the flat region may be determined with this type of scoring algorithm.

For a software implementation, the synchronization techniques may be implemented with modules (e.g., procedures, functions, and so on) that perform the functions described herein. The software codes may be stored in a memory unit (e.g., memory unit 192 in fig. 1) and executed by a processor (e.g., controller 190). The memory unit may be implemented within the processor or external to the processor.

While the principles of the disclosure have been described above in connection with specific apparatuses and methods, it is to be clearly understood that this description is made only by way of example and not as a limitation on the scope of the invention.

Claims (32)

1. A method for characterizing a communication channel, the method comprising:

moving a detection window across the channel profile to accumulate tap energies in the channel profile within the detection window into an accumulated energy curve,

determining a peak value of a maximum value in the accumulated energy curve, defining a frequency band relative to the accumulated energy curve, wherein the frequency band defines a region of the accumulated energy curve at or near the maximum value within the frequency band,

detecting a First Arriving Path (FAP) using a trailing edge found near a second end of the region of accumulated energy curves,

detecting a leading edge found near a first end of the region of the accumulated energy curve, an

The leading edge is used to determine the Last Arrival Path (LAP).

2. A method for characterizing a communication channel as recited in claim 1, wherein the band is placed relative to the maximum.

3. The method for characterizing a communication channel as recited in claim 1, further comprising:

acquiring a plurality of pilot symbols, an

Determining the channel profile from the plurality of pilot symbols.

4. The method for characterizing a communication channel as claimed in claim 1, further comprising the step of determining a delay spread using the FAP and LAP.

5. The method for characterizing a communication channel as recited in claim 1, wherein the band of frequencies is different in size at a first end of the region than at a second end of the region, such that the band of frequencies is tapered.

6. A method for characterizing a communication channel as claimed in claim 1, wherein the step of defining a frequency band comprises one of the following sub-steps:

defining the frequency band as a percentage of the maximum value, and

the frequency band is defined at a predetermined decrement from the maximum value.

7. The method for characterizing the communication channel as recited in claim 1, further comprising a step of positioning an FFT collection window based at least in part on at least two of FAP, LAP, or delay spread.

8. A method for characterizing a communication channel as recited in claim 1, further comprising the step of selecting channel taps from the channel profile that are below a predetermined threshold.

9. A method for characterizing a communication channel as recited in claim 1, wherein the analyzing step uses a two-pass algorithm.

10. The method for characterizing a communication channel as recited in claim 1, wherein the step of determining a channel profile includes the step of determining a channel profile using a plurality of Orthogonal Frequency Division Multiplexing (OFDM) symbols.

11. The method for characterizing a communication channel as recited in claim 1, wherein:

collecting a plurality of pilot symbols from a plurality of OFDM symbols within a frame, an

The step of determining the channel profile comprises the step of determining the channel profile using the plurality of pilot symbols.

12. The method for characterizing a communication channel as recited in claim 1, wherein:

collecting a plurality of pilot symbols from a plurality of OFDM symbols in a plurality of frames, an

The step of determining the channel profile comprises the step of determining the channel profile using the plurality of pilot symbols.

13. A method for characterizing a communication channel as recited in claim 1, wherein the communication method uses OFDMA signals.

14. A receiver for characterizing a communication channel, the receiver comprising:

means for moving a detection window across the channel profile to accumulate tap energies in the channel profile within the detection window into an accumulated energy curve,

means for determining a peak value of a maximum value in the accumulated energy curve,

means for defining a frequency band relative to the accumulated energy curve, wherein the frequency band defines a region of the accumulated energy curve at or near the maximum within the frequency band,

means for detecting the FAP using a trailing edge found near a second end of the region of the accumulated energy curve,

means for detecting a leading edge found near a first end of said region of the accumulated energy curve, an

Means for determining the LAP using the leading edge.

15. The receiver for characterizing a communication channel as recited in claim 14, further comprising:

means for acquiring a plurality of pilot symbols, an

Means for determining a channel profile from the plurality of pilot symbols.

16. The receiver for characterizing the communication channel as recited in claim 14, wherein the frequency band is positioned relative to the maximum.

17. The receiver for characterizing a communication channel as claimed in claim 14, further comprising means for determining a delay spread using FAP and LAP.

18. The receiver for characterizing the communication channel as recited in claim 14, wherein the band is different in size at a first end of the region than at a second end of the region, such that the band is tapered.

19. The receiver for characterizing a communication channel as claimed in claim 14, wherein said defining means comprises one of:

means for defining the frequency band as a percentage of the maximum value, and

means for defining the frequency band with a predetermined decrement from the maximum value.

20. The receiver for characterizing the communication channel as recited in claim 14, further comprising means for locating an FFT collection window based in part on at least two of FAP, LAP, or delay spread.

21. The receiver for characterizing a communication channel as recited in claim 14, further comprising means for selecting channel taps from the channel profile that are below a predetermined threshold.

22. A receiver for characterizing a communication channel as recited in claim 14, wherein the means for analyzing uses a two-pass algorithm.

23. The receiver for characterizing a communication channel as recited in claim 14, wherein the means for determining the channel profile includes means for determining the channel profile using a plurality of Orthogonal Frequency Division Multiplexing (OFDM) symbols.

24. A communication device for characterizing a communication channel, the communication device comprising:

a processor configured to:

moving a detection window across the channel profile to accumulate tap energies in the channel profile within the detection window into an accumulated energy curve,

determining a peak value of a maximum value in the accumulated energy curve,

defining a frequency band relative to the accumulated energy curve, wherein the frequency band defines a region of the accumulated energy curve at or near the maximum within the frequency band,

detecting the FAP using a trailing edge found near a second end of the region in the accumulated energy curve,

detecting a leading edge found near a first end of the region of the accumulated energy curve, an

Determining the LAP using the leading edge;

a memory coupled to the processor.

25. The communication device for characterizing a communication channel as recited in claim 24, the processor further configured to:

acquiring a plurality of pilot symbols, an

A channel profile is determined from the plurality of pilot symbols.

26. A communications device for characterizing a communications channel as claimed in claim 24, wherein the frequency band is positioned relative to the maximum.

27. The communication device for characterizing the communication channel as recited in claim 24, wherein the band of frequencies is different in size at a first end of the region than at a second end of the region, such that the band of frequencies is tapered.

28. The communication device for characterizing the communication channel as recited in claim 24, wherein defining the frequency band comprises one of:

defining the frequency band as a percentage of the maximum value, and

the frequency band is defined at a predetermined decrement from the maximum value.

29. The communication device for characterizing the communication channel as recited in claim 24, wherein the processor further selects channel taps from the channel profile that are below a predetermined threshold.

30. The communication device for characterizing the communication channel as recited in claim 24, wherein the processor determines the channel profile using a plurality of OFDM symbols.

31. The communication device for characterizing the communication channel as recited in claim 24, wherein:

the plurality of pilot symbols are collected from a plurality of OFDM symbols within a frame, an

Determining the channel profile includes determining the channel profile using the plurality of pilot symbols.

32. The communication device for characterizing the communication channel as recited in claim 24, wherein:

the plurality of pilot symbols are collected from a plurality of OFDM symbols in a plurality of frames, an

Determining the channel profile includes determining the channel profile using the plurality of pilot symbols.