MXPA06014941A - Time synchronization using spectral estimation in a communication system - Google Patents

Abstract

To perform time synchronization using spectral estimation, a receiver obtains a frequency response estimate for pilot symbols received on each set of frequency subbands used for pilot transmission. The receiver performs spectral estimation on frequency response estimates for different sets of sub bands to obtain a measured arrival time for a transmission from a transmitter. The spectral estimation determines a dominant frequency component in the frequency response estimates and derives the measured arrival time based on this dominant frequency component. A time error between the measured arrival time and a desired arrival time is computed and possibly filtered. The filtered or unfiltered time error is scaled with a fixed or adjustable gain. A time adjustment is then generated based on the scaled time error and using linear and/or non-linear functions. The time adjustment is sent to the transmitter and used to adjust the transmit timing at the transmitter.

Description

SYNCHRONIZATION OF TIME USING THE CALCULATION OF THE SPECTRUM IN A COMMUNICATION SYSTEM Field of the Invention The present invention refers generally to communications, and specifically to time synchronization in a communication system.

BACKGROUND OF THE INVENTION A multiple access communication system can support communication for multiple user terminals by sharing the available resources of the system (eg, time, frequency, and / or transmission power). Each user's terminal communicates with one or more base stations through transmissions on the direct and inverse links. The forward link (or downlink) refers to the communication link from the base stations to the user terminals, and the reverse link (or uplink) refers to the communication link from the user terminals to the base stations. In the reverse link, the base station can receive transmissions from multiple user terminals. The transmission of each user terminal can travel through a different group of signal paths. The signal paths of different user terminals usually have different channel gains and propagation delays. Accordingly, transmissions from these user terminals may arrive at different times in the base station during the same transmission start time. Transmissions can interfere with each other, if they are not properly aligned for the time to the base station. This interference may adversely impact the ability of the base station to recover the transmission from each user's terminal, and may degrade the performance of all affected user's terminals. A time control circuit can be used to adjust the synchronization of each user terminal, so that its transmission arrives at the appropriate time at the base station. The timing control circuit design can be challenging due to various factors such as, for example, difficulty in obtaining an accurate measurement of the arrival time of a transmission. This difficulty may be due to the way in which the transmission is sent, the adverse channel conditions, etc. Accordingly, there is a need in the art for techniques to accurately measure the arrival time in a receiver and to properly adjust the timing in a transmitter in a communication system.

SUMMARY OF THE INVENTION In the present invention, techniques are described for carrying out time synchronization using the spectrum calculation. Time synchronization comprises determining the arrival time of a transmission (or a signal) in a receiver and adjusting the transmission time in a transmitter, so that the transmission arrives at a desired time to the receiver. The receiver normally does not know a priori the arrival time of the transmission, because the transmitter can transmit starting at an arbitrary time, and the wireless channel can introduce an unknown delay. The receiver can obtain a relatively accurate arrival time measurement for the transmitter from the transmitter, using spectrum calculation. In one embodiment for performing time synchronization using spectrum calculation, the receiver obtains a frequency response calculation of the pilot symbols received in each group of frequency sub-bands used for the pilot transmission. The receiver can obtain different frequency response calculations for different groups of subbands.

Subsequently, the receiver can carry out spectrum calculation in the frequency response calculations to obtain a measured arrival time for transmission from the transmitter. The calculation of the spectrum determines a dominant frequency component in the frequency response calculations and derives the time of arrival measured based on this dominant frequency component, as described later. A time error between the measured arrival time and a desired arrival time is computed and can be filtered as a low pass filter. The filtered or unfiltered time error is scaled with a gain, which can be adjusted based on one or more criteria. Subsequently, a time adjustment is generated based on the scaled time error and using linear and / or non-linear functions. For example, the time setting may be limited (or saturated) to be within a predetermined range of values to take into account a possible impression as the time of arrival. The time setting is sent to the transmitter and use the transmit timing on the transmitter.

Next, various aspects and embodiments of the present invention are described in more detail.

BRIEF DESCRIPTION OF THE DRAWINGS The characteristics and nature of the present invention will be better appreciated from the detailed description set forth below, when taken in conjunction with the drawings in which the similar reference characters are identified in a manner corresponding throughout the description. Figure 1 shows a wireless communication system, multiple access. Figure 2 shows a process for adjusting the timing of a user's terminal for transmitting data to a base station. Figure 3 shows a model in a time control circuit (TCL). Figures 4A and 4B show two exemplary transmission schemes. Figure 5 shows a channel profile and a response response of time of arrival. Figure 6 shows responses of the passage of the time control circuit for different TCL gains. Figure 7 shows a process for carrying out time synchronization using spectrum calculation. Figure 8 shows a block diagram of a user terminal and a base station. Figure 9 shows an OFD demodulator and a time control unit.

Detailed Description of the Invention The word "example" is used in the present invention to mean "that it serves as an example, case or illustration". Any design modality described herein as "exemplary" will not necessarily be constructed as preferred or advantageous over other modalities or designs. Figure 1 shows a multiple access wireless communication system 100. The system 100 includes a number of base stations 110 that support the communication of a number of terminals of the user 120. A base station is usually a fixed station used to communicate with the user terminals and can be called an access point, a Node B, or any other terminology. The user terminals 120 are normally dispersed throughout the entire system, and each user terminal can be fixed or mobile. A user terminal may also be referred to as a base station, a user equipment (UE), a wireless communication device, or some other terminology. Each user terminal can communicate with one or more base stations in the direct and inverse links at any given time. For simplicity, Figure 1 shows only transmissions on the reverse link. For a centralized system, a system controller 130 is coupled to the base stations and provides coordination and control for these base stations. The time synchronization techniques described herein can be used by several wireless and wired communication systems. For example, these techniques can be used for an orthogonal frequency division multiple access (OFDMA) system, a Time Division Multiple Access (TDMA) system, a Code Division Multiple Access (CDMA) system, a Frequency Division Multiple Access System (FDMA), etc. Other examples of multiple access systems include Multiple-Carrier CDMA (MC-CDMA), and Broadband CDMA (W-CDMA), High-Speed Downlink Packet Access (HSDPA). The techniques can also be used by the direct link, to adjust the timing of the base stations and by the reverse link to adjust the timing of the user terminals. For clarity, these techniques are described below for the reverse link in a wireless OFDMA system. An OFDMA system uses orthogonal frequency division multiplexing (OFDM). OFDM is a multiple carrier modulation technique that effectively divides the bandwidth in a general manner into multiple sub-bands of orthogonal frequency (N). These subbands are commonly called shades, sub-transporters, deposits and frequency channels. Each subband is associated with a respective sub-carrier that can be modulated with data. Up to N modulation symbols can be sent in the N total subbands in each period of OFDM symbols. These modulation symbols are converted to the time-domain with an N-point inverse fast Fourier transporter (IFFT) to generate a transformed symbol containing N chips or time-domain samples. To combat inter-symbol interference (ISI), which is caused by selective frequency fading in a multiple channel, the C chips of the transformed symbol are repeated to form an OFDM symbol containing N + C chips, where C is normally a fraction of N. The chips repeated by C are often referred to as a cyclic prefix, and C is the length of the cyclic prefix. An OFDM symbol period is the duration of an OFDM symbol, it is equal to the N + C chip periods.

As shown in Figure 1, the multiple terminals of the user can transmit to each base station. The user terminals can be located throughout the entire system and may have different propagation delays for the base station. The propagation delay of each user terminal can also change over time due to movement through the user's terminal, change in the wireless channel, etc. The timing of the transmission from each user terminal can be adjusted to take the delay for the propagation observed by said user terminal. This ensures that transmissions from different user terminals arrive aligned in time at each base station, and do not interfere with each other. Figure 2 shows a process for adjusting the timing of the user terminal for a data transmission on the reverse link from the user terminal to a base station. Initially, the user's terminal transmits a request message to the base station. This message may require access to the system, permission to transmit on the reverse link, etc. This message may contain an indication of the start time of that moment in which the message was sent, based on the transmission timing in the user's terminal. The base station receives the request message and measures the arrival time of the transmission from the user's terminal. The base station subsequently determines the initial time compensation, for example, based on the transmission start time indicated by the request message and the arrival time measured by the base station. This initial time compensation is the adjustment amount for the transmission timing at the 1-user terminal, with the object that the base station receives the transmission from the user's terminal at just the right time. The base station sends the initial time compensation through a signaling channel (SCH). The user terminal receives the compensation of the initial time, adjusts its transmission timing accordingly, and transmits data using the initial time compensation. The base station receives the data transmission from the user's terminal and measures the arrival time of this transmission. The propagation delay of the user's terminal may have changed since the last transmission. In this case, the arrival time of the data transmission may be different from the desired arrival time for data transmission, which may be the beginning of a time interval that the base station has assigned to the user's terminal. The base station computes the difference or error between the measured arrival time and the desired arrival time and determines the time setting for the user's terminal. Subsequently, the base station sends the time adjustment through the SCH. The user terminal receives the time adjustment, updates its transmission timing accordingly, and transmits the data using the updated time offset. The timing adjustment process normally continues throughout the entire data transmission from the user's terminal. The base station measures the arrival time of each transmission received from the user's terminal and determines the time adjustment. The user terminal updates its transmission timing from each timing setting received from the base station. The current transmission timing in the user's terminal is determined through the initial time compensation in addition to all the received time settings of the base station. The user terminal uses the current transmission timing for transmission to the base station. Figure 3 shows an exemplary model 300 of a time control circuit (TCL) between a user terminal 120x and a base station 110x. The user's terminal transmits the reverse link using its current transmission timing. The transmission observes a propagation delay through the wireless channel and can also be distorted by multiple paths in the wireless channel. The propagation delay may change arbitrarily over time.

The base station, a arrival time measurement unit 310 - receives the transmission from the user terminal, and the arrival time of the received transmission, and provides the measured arrival time. An adder 312 subtracts the measured arrival time from the desired arrival time and provides the time error for the received transmission. A delay unit 314 provides a delay of an update period for the time control circuit. The delay unit 314 is included in the model 300 to take into account the delay in updating the time control circuit. This update delay is due to the fact that the measurement of the arrival time processed in the normal TCL update time period is not reflected in the transmission timing in the user terminal until the next TCL update period. The TCL update range, and hence the TCL update period, is determined through the range in which time adjustments are sent by the base station of the user's terminal. The TCL update range may be fixed or may be variable depending on the design of the system. A multiplier 316, multiplies the delayed time error of the delay unit 314, with a TCL gain, and provides an error in scaled time. The TCL gain determines the characteristics of the time control circuit, as described below. A subsequent processor 318 quantizes the error in scaled time, performs post processing (if it exists) on the quantized value, and provides an adjustment in time for the received transmission. For example, the back processor 318 may perform saturation and limit the adjustment in time to be within a range of previously determined values. The time adjustment indicates whether the user's terminal should advance or delay its transmission timing, so that the transmission from the user's terminal is received at the desired arrival time. The time adjustment can be provided in previously determined units (for example, eight of chips). The base station sends the adjustment in time through the direct link to the user's terminal. At the user's terminal, an adder 320 receives the time adjustment of the base station, adds this time adjustment to the pre-transmission timing stored in a register 322, and provides the normal transmission timing. The normal transmission timing is stored in the register 322 and is also used for the next transmission to the base station. The adder 320 and the register 322 form an accumulator which updates the transmission timing in the user's terminal, based on the time settings received from the base station. The processing for the initial time compensation may be different from the processing of the time adjustment. For example, the error in time for the first transmission received from the user's terminal can be scaled by one unit (instead of the TCL gain) to obtain the initial time compensation, and subsequent processing can be supported. The adder 320 can add the compensation in the initial time or the pre-transmission timing in the same way as for the time adjustment. Figure 3 shows a first order time control circuit with an accumulator within the circuit. Other time control circuits may also be used to adjust the transmission timing at the user terminal, to take into account the propagation delay introduced by the wireless channel. For example, a second order time control circuit may also be used. The arrival time of a transmission can be measured using several techniques. You can select an appropriate measurement technique based on several factors, such as, for example, the way in which the data and pilot are transmitted in the system. A pilot is known data that is processed and transmitted in a known way. A transmitter can transmit a pilot to help a receiver perform various versions, such as time synchronization, channel calculation, frequency correction, automatic gain control, etc. Following are several example transmission schemes for the transmission of data and pilots. Figure 4A shows a frequency hopping transmission scheme (FH) 410 that can be used for the OFDMA system. Frequency hopping can randomize the interference and provide frequency diversity against detrimental effects of the trajectory. With the frequency hop, each user terminal is assigned a different frequency FH indicating the particular subband (s) to be used in each hop period. Each hop period can span one or multiple periods of OFDM symbols. Each FH sequence can select pseudo-random sub-bands that are used for transmission. The FH sequences of different user terminals in communication with the same base station are orthogonal to each other, so that two user terminals do not use the same subband at any given hop period. This avoids the "intra-cell" interference of the user terminals communicating with the same base station. The FH sequences of each base station are pseudo-random with respect to the FH sequences for the nearby base stations. This randomizes the "inter-cell" interference between the user terminals in communication with different base stations. For the example shown in Figure 4A, groups of sub-bands S with N sub-bands are formed, and each group contains M sub-bands, where N = M x S, M = 1, and S >; 1. The sub-bands M in each group may be contiguous (as shown in Figure 4A) or non-contiguous. Each user terminal can be assigned a group of subbands in each period. Data symbols (which are modulation symbols for data) can be multiplexed by time division (TDM) with pilot symbols (which are modulation symbols for pilot), as shown in Figure 4A. The transmission scheme 410 can be used, for example, for the reverse link in the OFDMA system. Figure 4B shows an interlaced transmission scheme 420 that can also be used by the OFDMA system. For the example shown in Figure 4B, the sub-band groups S are formed, each containing sub-bands M. To provide frequency diversity, the sub-bands M in each group, can be separated uniformly between the sub-bands S. The groups of sub-bands subsequently interlaced as shown in Figure 4B. In each OFDM symbol period, a group of sub-bands can be used for the pilot transmission and the remaining S-1 sub-band groups can be used for data transmission. Groups of different sub-bands can be used for pilot transmission in different periods of OFDM symbols. These groups of sub-bands may be selected in a pseudo-random manner (for example, with a PN sequence) or in a determinant manner (for example, with a predetermined sequence of length S, as shown in Figure 4B) . The transmission scheme 420 can be used for example, for the direct link in the OFDMA system. A transmission scheme that is a combination of two schemes shown in Figures 4A and 4B, can be used for the reverse link. For this transmission scheme, interlaced sub-band groups S are formed, as shown in Figure 4B. However, each assault period encompasses multiple periods of symbols, and the data and pilots are multiplexed by time division in each sub-band in each jumping period, for example, as shown in Figure 4A.

In general, any number of sub-band groups can be formed, and each group can contain any number and can contain any combination of sub-bands. The measurement of arrival time and time synchronization can be carried out in different ways depending on the particular scheme used for data transmission and pilot transmission. In one embodiment, time synchronization is carried out using spectrum calculation. Spectrum calculation can provide relatively accurate time-of-arrival measurements based on a narrow-band pilot that is transmitted only through a portion of the entire system bandwidth at a time, for example, in the sub-bands M between the total subbands N as shown in FIGS 4A or 4B. The spectrum calculation can also be used for a broadband pilot that is transmitted in all or a large percentage of the total subbands N, for example, in a TDM form with data. A wireless channel in the OFDMA system can be characterized either by a channel impulse response or a corresponding channel response frequency. As used in the present invention, and which is consistent with conventional terminology, a channel impulse response is a wireless channel-time domain response, and a channel frequency response is a frequency-domain response of the Chanel. In a sampled data system, the channel frequency response is an independent Fourier transformation (DFT) of the channel impulse response. The impulse response of the channel is composed of a sequence of "channel derivations", each branch of the channel being defined by a channel derivation gain (or simply, "derivation gain") and a channel derivation delay (or simply "derivation delay"). The channel frequency response is composed of a group of "channel gains", with each channel gain for a specific subband. The channel impulse response has L channel branches of interest, eg, L channel leads of sufficient strength, where L = N. Each branch of the channel has a complex gain of hi and is located in a delay of d. In general, each branch of the channel can be located anywhere between 1 and N (or l = dj. = N), where N is also the duration or length of time of the wireless channel. The derivation gains L are denoted as. { hi} , or hi for i = 1,2, ... L. The delays of the derivation L are denoted as. { gave} , or say for i = 1,2, ... L. The gains of derivation. { hi} they are correlated random variables that vary in a range determined by the Doppler spread of the wireless channel. The derivation gains L. { hj ..}. , as well as the derivation delays L. { gave} , are not known and can be calculated as will be described below. The channel impulse response can be represented in the z-domain through a finite impulse response (FIR) derivative-L, H (z), as indicated below: í = 1 Eq. (1) where z "1 denotes a delay of a chip period and z" di denotes the delay of the derivation of the i-th channel. The impulse response of the channel can also be represented by a vector L x 1 and h, without explicit reference to the derivation delays. { gave} , as indicated below: h = [hi h2 ... hL] T, Eq. (2) where "T" denotes a transpose. A channel profile can be defined as follows: P = diag < h-hT), Eq. (3) where () denotes an average-time operation; diag. { M.}. is a diagonal matrix just with the diagonal elements of a matrix M; and P is a diagonal matrix for the channel profile.

A diagonal matrix contains possible non-zero values along the diagonal and zeroes anywhere. The diagonal elements P represent a channel profile defined by h. The channel profile indicates the averaged long-term energy of the channel branches in the impulse response of the channel. The channel profile does not include short-term effects, such as fading, Doppler, etc. The profile of the channel is therefore indicative of the capacity of reflection / transmission capacity of a medium through which a signal can travel. A gain of the frequency-domain channel can be calculated for each sub-band used for pilot transmission, as indicated below: * Pk 'for k = 1, 2, ... M, Eq. (4) where y¾ is a pilot symbol received for the subband k; pk is a pilot symbol transmitted in the subband k; and Hk is a channel gain calculation for the subband k. For simplicity, equation (4) assumes that the consecutive subbands M are used for pilot transmission, for example, as shown in Figure 4A, so that the sub-band index k runs from 1 to M. In general, any subbands can be used for pilot transmission, and these subbands are known to the receiver.

The earnings of channel M. { Hk} for the M-subbands used in the pilot transmission, they can be estimated based on pilot symbols received in these subbands, as shown in equation (4). Channel gains are frequency-domain values. Each gain of the channel can be expressed as a Fourier transformation of the time-domain channel derivations L (not known), as indicated below: ".|¿ (^ '-' ^^ - a ^" - »-,) for k .1 (2> ... M (Eq. (5) where GH = 2n di / N is an angular frequency (in radians) for the derivation of the i-th channel, and ni is the noise for the derivation of the i-th channel The factor "k-1" (instead of just "k") in the exponents in the equation ( 5), is due to a subband index that starts with 1 instead of 0. The angular frequencies, GH for i = 1,2, ... L, are frequency components of the channel gains { Hk.}, And are directly related to the unknown derivation delays for the channel impulse response, the derivation delays can be estimated thereby carrying out the spectrum calculation in the channel gains { Hk}, as described below, Equation (5) can be expressed in the matrix, as indicated below: or H = Qh + n, Eq. (7) where H is an M x 1 vector containing the channel gain calculations of the M subbands; Q is a "Fourier type" matrix M x L containing the elements shown in equation (6); and n is a noise vector L x 1. The pilot can be transmitted in different groups of sub-bands M at different time intervals, for example, as shown in Figure 4A or 4B. For the transmission scheme shown in Figure 4A, the pilot can be sent in the sub-bands k = 1, 2, ... M in a time interval, subsequently in the sub-bands k = 1 + b, 2 + b, ... + b in a next time interval, and so on, where b can be any arbitrary compensation value that is determined by the sequence FH. The channel gains for the pilot sent in the sub-bands k = 1 + b, 2 + b, ... M + b can be expressed as: ¾ = Q ·? -h + n, Eq. (8) where B is a diagonal matrix L x L determined by B = diag (ejbui, ejbu2, ... ejbuL). A correlation matrix M x M (or an external product) of H can be defined as H * HH, where "H" denotes a conjugate transpose. An average long-term time of the matrix of the correlation H, denoted as R, can be expressed as: R = < H · ??) = Q -P-QH + o2 -I, Eq. (9) where I_ is an identity matrix with ones along diagonals and zeros in any place; and s2 is the variance of the noise. Equation (9) is obtained based on equations (3), (7), and (8). The different compensation values b can be selected (for example, in a pseudo-random form as shown in Figure 4A or in a determinant form as shown in Figure 4B), so that the matrices B for an assembly of different compensation values is averaged to zero. In this case, if a sufficient amount of average is made in the correlation matrices obtained for different time intervals, then the matrices B cancel out and do not appear in R. Equation (9) also assumes that the channel noise is noise Gaussian white additive (AWGN) with an average of zero, a variance of o2, and a matrix of autocovariance of (£ nm = O2 'I. The decomposition of Eigen value can be carried out in the matrix R as follows: R = V -D-VH, Eq. (10) where V is a unitary matrix M x M of values Eigen of R, and D is a diagonal matrix of Eigen values of R. An unitary matrix M is characterized by the property MHM = _ I. The columns of a unitary matrix are orthogonal to each other, and each column has the power of The decomposition of Eigen value is described by Gilbert Strang in a book entitled "Linear Algebra and Its Applications", Second Edition, Academic Press, 1980. The diagonal elements M of D are referred to as Eigen values of R. The M columns of V are referred to as Eigen vectors of R. Each column of V, corresponds to an Eigen D value. Therefore, the first column or the far left column V corresponds to the diagonal element in the first column of D, the second column of V corresponds to the diagonal element in the second column of D, and so on. The Eigen M values in D can be ordered from the smallest to the largest and denoted as. { ??, ?2, ... ??} after the ordination, where ?? Is the Eigen value smaller and ?? It is the largest Eigen value. When ordering the Eigen values in D, the Eigen V vectors are ordered accordingly. Yes > L, then the smaller Eigen values M - L in D (for example, ?? a ?? - L) are equal to the noise variance s2 and are referred to as "noise" Eigen values. The Eigen M - L vectors in V that correspond to the Eigen M - L values (for example, the left - most columns M - L of V after the ordering), are referred to as Eigen "noise" vectors of R and are denoted as. { y_i, y_2, ... vM_L} . The Eigen noise vectors are orthogonal to the columns of Q. The gains / power of derivation L are contained in the matrix P and the derivation delays L are contained in the matrix Q. Each of the columns L of Q have the following form: ^ = [1, e 2nU-l) Ní ej2n (í-l) Ní ... gj2n (-l) (í-l) / N) j ^? c _ (n) where i is an index representing an unknown derivation delay and is within a range of 1 to N, or t C. { 1, 2, ... N.}. .

A cost function can be defined as follows: for t = 1, 2, ... N. Eq. (12) The unknown derivation delays L can be obtained based on the cost function C. { ¿), As indicated below. The cost function is evaluated for each of the possible values N of t, that is, for t -1, 2, ... N. Each value of representa represents a hypothesized delay value for a channel derivation. For each value of t, the vector q_t, is first determined as shown in equation (11) and multiplied with each of the Eigen noise vectors M - L to obtain internal products M - L, gk = qH t -vk for k = 1, 2, ... M - L. The power of each internal product is subsequently computed as | gk | 2 = gk -g * ^ where "*" denotes a complex conjugate. The powers of the internal products ML are added later, and the inverse of the summed power is provided as a cost value C (for this value of T. The cost values N, C, for t = 1,2,. .. N, are obtained for possible values N of t Since the columns of Q are orthogonal for noise vectors Eigen, the internal product of any column of Q with any Eigen vector of noise is small or is zero. Consequently, the summed power of the internal products M - L for each column of Q is small, and the inverse of this summed power is large.The larger values L between the cost values N, are subsequently identified. These correspond to these larger cost values L, represent the non-known derivation delays L for the channel impulse response.These values identified by L of t, are used to form the Q matrix and are also used together with the compensation value bc The derivative gains L can be derived later as follows: where Hb is an M x 1 vector for the frequency response calculation for a group of pilot sub-bands M; and hb is a vector L x 1 for a channel impulse response calculation with derivations L. An impulse response estimate may be computed hb for the frequency response estimate Hb, which is obtained from a pilot transmission received in one of the sub-band groups. The impulse response calculations for different groups of subbands can be averaged to obtain an impulse response estimate averaged with L derivations. In the previous description, L denotes the number of channel derivations that will be calculated, where L = M. In general, L may or may not be the number of channel derivations (Lact) in the actual impulse response of the wireless channel. If L = Lact < , then the Lact channel derivations can be estimated as described above. Yes L? Lact and L < M, then the L channel leads representative of the channel profile for the wireless channel can be obtained as described above. In general, as M increases, more channel derivations can be calculated with good accuracy and high resolution. For a broadband pilot with M being equal or reaching N, a full-channel impulse response can be calculated with up to N leads, based on the broadband pilot. To measure the arrival time of a transmission received through a wireless channel, the spectrum calculation technique shown in equations (1) to (2) can be carried out with an adjusted L equal to one. If L = 1, then the spectrum calculation technique provides a single channel derivation located in the center of the channel profile, assuming that M is greater than one. The derivation delay corresponds to this single channel derivation, it can be used as the arrival time measured for the transmission. A computer simulation was carried out for the measurement of arrival time using spectrum calculation for an exemplary wireless channel in an example OFDMA system. This system has OFDMA has a general system bandwidth of 5 MHz and uses an OFDM structure with 1024 total subbands (N = 1024), 16 subbands in each group (M = 16), and a prefix length cyclic 48 chips (C = 48). Each OFDM symbol period is 214.4 ysec, or (1024 + 48) / (5xl06). Each jump period covers seven OFDM symbol periods or 1.5 msec. A pilot symbol is sent in the middle of each jump period as shown in Figure 4A. The wireless channel has a channel profile that is 38 chips in length, or. { gave} =. { 1, 2, ... 38.}. , which is approximately 80% of the cyclic prefix length. The channel profile is modeled with a different sequence of random values for each TCL update period. Changing the channel profile from one TCL update period to the next so that the sequence of random values for the channel profile of the normal TCL update period is not correlated with the sequence of random values for an update period TCL previous. The channel noise is an AWGN noise, and the signal to noise ratio (SNR) for the received pilot is 0 decibels (dB). A measurement of arrival time is made in each TCL update period, so that the measurement range is the same as the TCL update range. The TCL update period is 90 msec, and each measure of arrival time is based on pilot symbols received in periods of 60 hops. A channel frequency response vector, Hb, is obtained for each jump period based on the pilot symbols received in the jump period. The matrix R is updated in each jump period based on the Hb vector. The arrival time is computed after 60 periods of jumps carrying out the equations from (10) to (12) with L = 1. Figure 5 shows a trace 510 of the channel derivations for the wireless channel in a TCL update period. The 38 channel branches are selected randomly for the TCL update period. Figure 5 also shows a trace 520 of a time-of-arrival measurement response provided by the spectrum calculation technique. The peak of trace 520 is the center of the channel profile, as estimated through the spectrum calculation technique. Figure 5 shows that the spectrum calculation technique can determine the center of the channel profile, even if the number of derivations of the channel exceeds the number of subbands used for the pilot transmission in each hopping period. The time corresponding to the center of the channel profile is given as the time of arrival measured. Referring again to Figure 3, the time error from adder 312 is scaled at the TLC gain through multiplier 316. The TLC gain determines the circuit bandwidth and the time control circuit damping factor, which in turn determines the temporal response of the time control circuit. In general, a larger TLC gain corresponds to a wider circuit bandwidth and a faster response time, but more timing distortion for a noisy wireless channel. Conversely, a smaller TLC gain corresponds to a narrower bandwidth and a slower response time, although with less timing distortion for a noisy wireless channel. The time control circuit performance is simulated by the example OFDMA system and the example wireless channel described for Figure 5 above. For the simulation, the initial arrival time for the received transmission is a zero chip compensation and the desired arrival time is a minus 15 chips compensation. The arrival example is measured each update period (which is each period represents jumps or 90) the time control circuit were also updated in each TLC update period. Figure 6 shows the steps of the response of the time control circuit of a step input with different TLC gains. The traces 610, 620 and 630 show the response responses of the time control circuit with TLC gains of 0.1, 1.0 and 1.5 respectively. The trace 610 shows a step response that has a settling time of about 3 seconds, which may be too slow. The trace 630 has a step response that has 8 over-firing and stamping chips, which may be undesirable. The trace 620 shows a step response that has a fast settling time for the desired arrival time and has no over-trip. Although not shown in Figure 6, a step response with a time constant gives approximately 200 msec and less noise can be achieved with a TLC gain of 0.5. The step responses shown in Figure 6 are obtained without saturation through the downstream processor 318 in Figure 3. As shown in Figure 6, the TLC gain affects the circuit characteristics, which determines the response time , as well as the noise performance of the time control circuit. The TLC gain can be a fixed value that is selected based on a negotiation between these two criteria. The TLC gain may also be a configurable value that is selected based on channel conditions. For example, a higher TLC gain can be used for a higher SNR, to improve the tracking performance, and a lower TLC gain can be used for a low SNR to improve the noise performance. The SNR can be calculated based on the received pilot. A higher TLC gain can also be used for a high range of displacement in the channel profile, and a low TLC gain can be used for a low range of displacement in the channel profile. The displacement range in the channel profile can be estimated based on the time of arrival measured. For example, a high displacement range can be deduced, if multiple time errors (or a high percentage of time errors) are for the same address. A higher TLC gain can also be used for a low variance in arrival time measures, which may indicate a stationary or relatively static channel for a user terminal with fixed mobility or low mobility. Conversely, a higher TLC gain can be used for a higher variance in carry-time measurements, which may indicate a rapidly changing channel for a user terminal with high mobility. The range of displacement of the channel profile can be confirmed and used to adjust the TLC gain, as well as to deduce certain characteristics (for example speed) of the wireless channel. The TLC gain can also be configurable based on operating modes. For example, a high TLC gain can be used during an acquisition phase and a low TLC gain during a tracking phase. The acquisition phase can span from the first FTA update periods, and the tracking phase can cover the remaining FTA update periods. The spectrum calculation technique generally provides accurate time-of-arrival measurements, although occasionally they can produce arrival time calculations that are inaccurate by a large amount. The maximum range that the channel profile can move is usually restricted by the speed of the user's terminal. For example, the spectrum calculation technique can produce arrival time estimates that are inaccurate for several chips (for example, by + 4 chips for the simulated scenario). The maximum range of displacement in the channel profile can be a fraction of a chip (for example, 0.1 chip) for each TLC update period. If spectrum calculation techniques produce a measure of arrival time that is different from the measure of the preceding arrival time by a large amount, then there is a high probability of inaccuracy in the measurement. In this case, the measurement of the normal arrival time can be provided with less weight during the updating of the time control circuit. You can achieve less weight in several ways. In one embodiment, the scaled time error from the multiplier 316 in FIG. 3 is saturated by the subsequent processor 318 that is within a predetermined range of values. This range can be selected based on the maximum displacement range in the channel profile. For example, if the maximum range of displacement is 0.1 chip per TLC update period, then the time adjustment can be limited to a range of +1 chip to -1 chip. This saturation reduces the amount of timing distortion due to inaccurate calculations of the arrival time. In another embodiment, the post processor 318 provides a large time adjustment if multiple large time errors are obtained. The after processor 318 can limit the time setting for the first large time error, as described above. The subsequent processor 318 can provide a large time adjustment if the subsequent time error is also large, which can confirm the accuracy of the previous arrival time measurement. This "wait and confirm" scheme can allow a faster circuit response, and at the same time eliminate many large arrival time measurement errors.

In another embodiment, the delay unit 314 is replaced with a low pass filter which filters the time errors from the adder 312. The bandwidth of this low pass filter is large enough so that the filter response does not alter adversely the closed circuit characteristics of the time control circuit. In yet another embodiment, the subsequent processor 318 derives the time adjustment based on (1) a first (eg, linear) function of the error in time, if the magnitude of the error in time is within a predetermined range and ( 2) a second function (for example, non-linear) of the error in time if the magnitude is outside the previously determined range. The non-linear function can be a subjection or saturation, the wait and confirm scheme described above, or some other function. The above description assumes that the time settings sent by the base station are correctly received by the user's terminal. Depending on how the settings are sent in time, there may be errors in the received time settings. The timing control circuit can be designed to take these errors into account. For example, time adjustments can be sent more frequently and / or with a higher transmission power in order to reduce the error range. Figure 7 shows a process 700 for performing time synchronization using spectrum calculation. A frequency response estimate is obtained for the pilot symbols received in each group of subbands (block 710). Different frequency response calculations can be obtained for different groups of subbands. Subsequently, the spectrum calculation is carried out in the frequency response calculations to obtain a measured arrival time (block 712). The spectrum calculation determines a dominant frequency component in the frequency response calculations and derives the arrival time measured based on this dominant frequency component. An error in time between the measured arrival time and the desired arrival time is determined (block 714). The error in time can be filtered with a low pass filter. The filtered or unfiltered time error is scaled with the TLC gain, which may be a fixed value or an adjustable value that is determined based on the conditions of the channel and / or the selected mode of operation (block 716). Subsequently, a time adjustment is generated based on the error in scaled time (block 718). The adjustment in time can be limited to be within a range of values previously determined or generated based on the linear and / or non-linear functions to take that take into account the imprecision of the measure. In the above description, the measure of arrival time is based on the recognition that the unknown delay delays (di for i = 1, 2, ... L) that will be confirmed are unknown frequency components (CÚÍ para i = 1, 2, ... L) of the frequency-domain channel gains obtained from the received pilot symbols. The spectrum calculation (or spectrum analysis) is carried out subsequently to determine the unknown frequency components of the channel gains. These frequency components, once determined, serve as calculations for the unknown derivation delays for a channel impulse response calculation. To measure the time of arrival, the spectrum calculation technique provides the dominant frequency component of channel gains, which is then used to derive the measured arrival time. The spectrum calculation technique can provide accurate time-of-arrival measurements for both (1) a narrow band pilot that is sent only in a small fraction of the bandwidth of the system at a time, and (2) a band pilot wide that is sent in all or a large fraction of the bandwidth of the system. For a narrow band pilot, only a small portion of the bandwidth of the system can be observed at any given moment. The resolution at which the arrival time of a transmission can be determined, is limited by the bandwidth of the observations of the channel. For example, if the pilot is sent only in the sub-bands M at a time, where M can be much less than N, then a receiver can only observe the wireless channel in a relatively narrow band based on the received pilot at these sub-bands M. Consequently, a crude arrival time measurement with a time resolution of T0fdm / M, can be obtained based on a narrowband pilot received in any of the sub-band groups M, where T0fdm is the duration of the OFDM symbol without a cyclic prefix. The spectrum calculation technique can provide a more accurate measurement of arrival time with a time resolution of Tofdm / N. As illustrated by the previous example, T0fdm / N can be much finer than the Tofdm / M time resolution obtained with any narrowband pilot transmission if M is much smaller than N. The precise arrival time measurement is important for synchronization in time, to avoid or minimize interference between multiple transmitters, and also for processing by the receiver to capture as much energy as possible in each received OFDM symbol. For clarity, a specific spectrum calculation technique has been described above, which is often referred to as a multiple signal classification technique (MUSIC). Other spectrum calculation techniques can also be used to confirm the frequency components of the frequency response estimates, and therefore the derivation delays of the impulse response calculations, and this is within the scope of the present invention. . For example, the calculation of spectrum can be carried out based on a periodogram technique, a Prony calculator, a Pisarenko harmonic decomposition technique, etc. Each spectrum calculation technique normally uses some type of score, to obtain a good calculation of the frequency component (s) that is being observed. These various spectrum calculation techniques, including the MUSIC technique, are described in S. L. Marple Jr. in "A Tutorial Overview of Modern Spectral Estimation", Proc. IEEE, 1989, pages 2152-2157, and in B. D. Kao and K. S. Arun in "Model Based Processing of Signals: A State Space Approach", Proc. IEEE, Vol. 80, No. 2, Feb 1992, pages 283-309. Figure 8 shows a block diagram of a user terminal 120x and a base station 110x. Through the reverse link, the user terminal 120x, a transmission data processor (TX) 810, receives, formats, encodes, interleaves and maps traffic data symbols and provides modulation symbols (or data symbols). An OFDM modulator 820 receives the data symbols and pilot symbols, performs OFDM modulation, and provides a stream of OFDM symbols. The pilot and data symbols can be transmitted in various ways, for example, as shown in Figure 4A. A transmission unit (TMTR) 822 receives and conditions (for example, converts to an analog, amplifies, filters and upconverts the frequency) the OFDM symbol stream and generates a reverse link signal, which is transmitted through the an 824 antenna to the base station llOx. In the base station 11Ox, an antenna 852 receives the reverse link signal and provides a signal received in the receiving unit (RCVR) 854. The receiving unit 954 conditions (eg, filters, amplifies and downconverts the frequency ) the received signal, digitizes the conditioned signal and provides received chips for an OFDM 856 demodulator. The OFDM demodulator 856 performs OFDM demodulation on the received chips, performs data detection on the received data symbols, and provides detected data symbols, which are calculations of the transmitted data symbols. A symbol of the reception data processor (RX) 858, desmaps, deinterleaves and decodes the detected data symbols to recover the transmitted data. Processing through the OFDM demodulator 856 and the RX data processor 858 is complementary to processing through the OFDM modulator 820 and the data processor TX 810, respectively, at the user terminal 120x. In the direct link, a TX 882 data processor processes traffic data and control data (for example, for time adjustments) and provides data symbols. An OFDM modulator 884 receives and multiplexes the data symbols with pilot symbols, performs OFDM modulation and provides a stream of OFDM symbols. The same or different transmission schemes can be used for the direct and inverse links. For example, the transmission scheme shown in Figure 4A can be used for the reverse link, and the transmission scheme shown in Figure 4B can be used for the forward link. An 886 unit receives and processes the OFDM symbol stream and generates the direct link signal, which is transmitted through an antenna 852 to the user terminals.

At the user terminal 120x, the direct link signal from the base station HOx is received by the antenna 824, and processed by a receiving unit 842 to obtain received chips. An OFD modulator 844 processes the received chips and provides detected data symbols to an RX data processor 846. The RX data processor 846 processes the detected data symbols to retrieve the traffic data transmitted by the base station HOx. The controllers 830 and 870 direct the operation at the user terminal 120x and the base station HOx, respectively. The memory units 832 and 872 store program codes and data used by the controllers 830 and 870, respectively. The time control units 828 and 868 carry out applicable time synchronization functions for the user terminal 120x and the base station HOx, respectively. The controllers 830 and 870 can also implement time control units 828 and 868, respectively, and can perform time synchronization functions for the terminal 120x and the base station HOx, respectively. Figure 9 shows one mode of the OFDM demodulator 856 and the time control unit 868 in the base station 11Ox. Within an OFDM demodulator 856, a cyclic prefix elimination unit 912 removes the cyclic prefix adhered to each OFDM symbol based on the measured arrival time provided by the time control unit 868. Subsequently an FFT 914 unit transforms the Ns. received chips for each transformed symbol received for the frequency domain with an N-dot FFT, and obtains N received symbols for the subbands N. The FFT unit 914 provides received pilot symbols for the time control unit 868 and symbols received data for a data detector 916. The data detector 916 performs detection (eg, cross-filtering or equalization) on the received data symbols with a frequency response calculation and provides detected data symbols to a RX data processor 858. The time control unit 868 performs processing in the receiving part for the time control circuit of the reverse link from the user terminal 120x to the base station llOx. A time of arrival measurement unit 920 obtains pilot symbols received and measures the arrival time of the transmissions from the user terminal 120x. Within unit 920, a pilot processor 922 eliminates modulation in each group of received pilot symbols and provides a frequency response calculation for the reverse link of the user terminal 120. A spectrum estimator 924 performs the calculation of spectrum in the frequency response calculations for different groups of pilot symbols (e.g., sent in different jump periods) and provides the measured arrival time for the transmission from the user terminal 120x. An adder 930 subtracts the measured arrival time from the desired arrival time and provides the error in time. The controller 870 can determine the desired arrival time based on the time programmed for the user terminal 120x. Although not shown in Figure 9, a low pass filter can filter the error in time and provide and filter the error in time. A multiplier 932 multiplies the error in time of adder 930 (or the error in filtered time from the low pass filter) with the TLC gain and provides the error in scaled time. A subsequent processor 940 receives the error in scaled time and generates the time adjustment for the user terminal 120x. The back processor 940 can implement any of the techniques described above, such as, for example, saturation, wait and confirm, etc. The time adjustment directs the user terminal 120x to advance or delay its timing, so that its transmission arrives at the desired arrival time at the 110x base station. A channel detector 942 detects the channel conditions, for example, SNR of the received transmission from the user terminal 120x, if the reverse link is static or changes rapidly, and so on. A TCL gain selector 944 receives the conditions of the detected channel from the detector 942, the selected mode of operation (eg, acquisition or tracking mode) of the time control circuit from the controller 870 and / or other inputs. Gain selector 944 adjusts the TCL gain to achieve the desired performance of the time control circuit. A time control circuit for the forward link, if necessary, can be implemented in a manner similar to that described above for the reverse link. For this time control circuit, the user terminal 120x measures the arrival time of a transmission received from the base station 11Ox, for example, using spectrum calculation. Subsequently, the user terminal generates the time adjustment for the direct link based on the measured arrival time and a desired arrival time for the direct link transmission. The base station adjusts its transmission timing for the user terminal, so that the direct link transmission arrives at the desired arrival time at the user's terminal. The time synchronization techniques described in the present invention can be implemented through various means. For example, these techniques can be implemented in hardware, software or a combination thereof. For a hardware implementation, the processing units for time synchronization in a receiver can be implemented within one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), signal processing devices digital (DSPDs), programmable logic devices (PLDs), field programmable output formations (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to perform the functions described herein, or a combination of the same. The processing units for time synchronization in a transmitter can also be implemented within one or more ASICs, DSPs, etc. For a software implementation, time synchronization techniques can be implemented with modules (e.g., procedures, functions, etc.), which perform the functions described above. The software codes can be stored in a memory unit (e.g., memory unit 872 in Figure 8) and executed by a processor (e.g., controller 870). The memory unit can be implemented inside the processor or externally to it. The above description of the embodiments described is provided to enable any person skilled in the art to make or use the present invention. Those skilled in the art will appreciate various modifications to these embodiments, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the modalities shown therein, without being in accordance with the broader scope consistent with the principles and novel features described herein.

Claims (42)

1. NOVELTY OF THE INVENTION Having described the present invention, it is considered as a novelty and, therefore, the content of the following is claimed as property: R E I V I N D I C A C I O N S 1. A method for performing synchronization in time in a communication system, wherein the method comprises: measuring a time of arrival of a transmission received through a communication channel using spectrum calculation; determine the error in time between the arrival time measured and the desired arrival time for the transmission; and generate an adjustment in time based on the error in time. The method according to claim 1, wherein the measurement of the arrival time of the transmission, comprises: obtaining a calculation of the channel for the communication channel, and carrying out spectrum calculation in the channel estimate to obtain the arrival time measured for the transmission. 3. The method according to claim 2, wherein obtaining the calculation of the channel for the communication channel, comprises: obtaining a plurality of frequency response calculations for a plurality of groups of frequency sub-bands, wherein the estimate The channel comprises the plurality of frequency response estimates. The method according to claim 2, wherein carrying out the spectrum calculation in the channel calculation, comprises: determining a dominant frequency component in the channel estimate, and deriving the arrival time measured based on in the dominant frequency component. 5. The method according to claim 1, wherein the method further comprises: scaling the error in time with a gain, and wherein the time adjustment is generated based on the error in scaled time. The method according to claim 5, wherein the method further comprises: adjusting the gain based on at least one condition of the communication channel. The method according to claim 5, wherein the method further comprises: adjusting the gain based on an operating mode selected from a plurality of operating modes comprising an acquisition mode and a tracking mode. 8. The method according to claim 5, wherein the method further comprises: determining a range of displacement in a channel profile obtained from the spectrum calculation; and adjust the gain based on the displacement range in the channel profile. The method according to claim 1, wherein the generation of the time adjustment based on the error in time, comprises: filtering the error in time, and deriving the adjustment in time based on the error in filtered time. The method according to claim 1, wherein the method further comprises: limiting the adjustment in time so that it is within a previously determined range of values. 11. The method according to claim 1, wherein the generation of the time adjustment based on the error in time, comprises: determining the magnitude of the error in time, generating the adjustment in time based on a first function if the The magnitude of the error in time is within a previously determined range, and generate the adjustment in time based on a second function if the magnitude of the error in time is outside the previously determined range. 12. An apparatus that operates to perform synchronization in time in a communication system, wherein the apparatus comprises: a measurement unit that operates to measure a time of arrival of a transmission received through a communication channel using a spectrum calculation; an adder operating to provide error in time between the measured arrival time and the desired arrival time for the transmission; and a subsequent processor that operates to generate an adjustment in time based on the error in time. 13. The apparatus according to claim 12, wherein the measurement unit comprises a pilot processor operating to receive the pilot sent through the communication channel, and to derive a calculation from the channel for the communication channel, and a spectrum calculator operating to carry out the calculation of spectrum in the calculation of the channel to obtain the arrival time measured for the transmission. The apparatus according to claim 13, wherein the pilot processor operates to derive a plurality of frequency response estimates for a plurality of groups of frequency sub-bands, and wherein the calculation of the channel comprises the plurality of frequency sub-bands. Frequency response calculations. 15. The apparatus in accordance with the claim 14, wherein the pilot processor operates to receive pilot symbols in each plurality of groups of frequency sub-bands and to derive the frequency response calculation for each group of frequency sub-bands based on the pilot symbols received from the group of frequency sub-bands. The apparatus according to claim 14, wherein each group of subbands includes a subset of N total subbands in the system, wherein N is an integer greater than 1. 17. The apparatus in accordance with claim 14, wherein each group of sub-bands includes M consecutive sub-bands in the system, where M is an integer greater than 1. 18. The apparatus in accordance with claim 14, wherein each group of subbands includes subbands evenly distributed across the N total subbands in the system, where M and N are integers greater than, and M is less than N. 19. The apparatus according to claim 14, wherein the plurality of subband groups are determined by a pseudo-random sequence used for the frequency hop. The apparatus according to claim 13, wherein the pilot processor operates to receive a broadband pilot sent through the communication channel and to derive the channel calculation based on the broadband pilot. The apparatus according to claim 13, wherein the spectrum calculation operates to determine a dominant frequency component in the channel calculation and to derive the measured arrival time based on the dominant frequency component. 22. The apparatus according to claim 13, wherein the spectrum calculation operates to perform spectrum calculation in the channel calculation using a multiple signal classification technique (MUSIC), a periodogram technique, a Prony estimator. or a Pisarenko harmonic decomposition technique. 23. The apparatus in accordance with the claim 12, wherein the apparatus further comprises: a multiplier that operates to scale the error in time with a gain, and wherein the subsequent processor operates to generate the time adjustment based on the error in scaled time. 24. The apparatus according to claim 23, wherein the gain is adjusted based on at least one condition of the communication channel. 25. The apparatus according to claim 23, wherein the gain is adjusted based on a signal to noise ratio (SNR) of the transmission. 26. The apparatus according to claim 23, wherein the apparatus further comprises: a controller operating to select an operating mode from among a plurality of operating modes and to adjust the gain based on the selected operating mode. 27. The apparatus according to claim 12, wherein the apparatus further comprises: a filter that operates to filter the error in time. 28. The apparatus according to claim 12, wherein the subsequent processor operates to limit the time adjustment so that it is within a previously determined range of values. 29. The apparatus in accordance with the claim 12, wherein the subsequent processor operates to determine the magnitude of the error in time to generate the time adjustment based on a first function, if the magnitude of the error in time is within a previously determined range, and to generate the adjustment in time based on a second function, if the magnitude of the error in time is outside the previously determined range. 30. The apparatus according to claim 12, wherein the subsequent processor operates to generate the time adjustment based on the error in time and in accordance with a first order circuit for time synchronization. The apparatus according to claim 12, wherein the subsequent processor operates to generate the time adjustment based on the time error and according to a second order circuit for time tuning. 32. The apparatus according to claim 12, wherein the communication system utilizes orthogonal frequency division multiplexing (OFDM). 33. The apparatus according to claim 12, wherein the communication system is an orthogonal frequency division multiple access (OFDM) system. 34. An apparatus that operates to carry out synchronization in time in a communication system, wherein the apparatus comprises: means for measuring a time of arrival of a transmission received through a communication channel using a spectrum calculation; means for determining an error in time between the measured arrival time and a desired arrival time for the transmission; and means to generate an adjustment in time based on the error in time. 35. The apparatus in accordance with the claim 34, wherein the means for measuring the arrival time of the transmission comprises: means for obtaining a calculation of the channel for the communication channel, and means for carrying out spectrum calculation in the calculation of the channel, to obtain the time of Metered arrival for the transmission. 36. The apparatus in accordance with the claim 35, wherein the means for obtaining the channel calculation for the communication channel comprises: means for obtaining a plurality of frequency response calculations for a plurality of groups of frequency sub-bands, wherein the channel estimate comprises the plurality of frequency response calculations. 37. The apparatus according to claim 35, wherein the means for carrying out the calculation of spectrum in the calculation of the channel, comprises: means for determining a dominant frequency component in the calculation of the channel, and means for deriving the arrival time measured based on the dominant frequency component. 38. The apparatus according to claim 34, wherein the apparatus further comprises: means for scaling the error in time with a gain, and wherein the time adjustment is generated based on the scaled time error. 39. The apparatus according to claim 34, wherein the apparatus further comprises: means for limiting the adjustment in time so that it is within a range of values previously determined. 40. A method for measuring the arrival time in a communication system, wherein the method comprises: obtaining a channel estimate for a communication channel; carry out spectrum calculation in the channel calculation, to determine a dominant frequency component in the channel calculation; and derive a time of arrival measured based on the dominant frequency component. 41. The method according to claim 40, wherein obtaining the channel calculation for the communication channel comprises: obtaining a plurality of frequency response calculations for a plurality of groups of frequency sub-bands, and wherein the calculation of the channel comprises the plurality of frequency response calculations. 42. An apparatus that operates to measure the arrival time in a communication system, wherein the apparatus comprises: means for obtaining a channel calculation for a communication channel; means for carrying out the calculation of spectrum in the channel calculation to determine a dominant frequency component in the calculation of the channel; and means for deriving a time of arrival measured based on the dominant frequency component. R E S U M E To perform time synchronization using the spectrum calculation, a receiver obtains a frequency response, calculated from the pilot symbols received in each set of frequency sub-bands used for the pilot transmission. The receiver performs the calculation of the spectrum of the response in the frequency response calculations for the different sets of subbands to obtain a measured arrival time for a transmission of a transmitter. The spectrum calculation determines a dominant frequency component in the frequency response calculations and calculates the measured arrival time based on the dominant frequency component. A time error between the measured arrival time and the desired arrival time is calculated and possibly filtered. The filtered or unfiltered time error is scaled with a fixed or adjustable gain. A time adjustment is then generated based on the scaled time error and using the linear and / or non-linear functions. The time setting is sent to the transmitter and used to adjust the transmission time at the transmitter. The base station Time Time FIG. 2 300 FIG. 3 Symbols Symbols Data Symbols Data Pilot c co .O .Ó D 2 CO 1 2 3 4 5 6 7 Periods of OFDM Symbols ..-- '410 Assigned Sub-bands Time Jump Period FIG. 4A Pilot symbol 420 Time Period of OFDM Symbols FIG. 4B Time (seconds) FIG.6 FIG. 7 FIG. 8 FIG.9