Field of the invention

[0001] The present invention is related to the field of wireless communication in vehicular environments and in particular to the channel tracking in fast vehicular environments.

Background of the invention

[0002] In wireless communication systems, wherein transmission from at least one transmission device to at least one receiving device is considered, knowledge of the channel between the transmitting device and the receiving device, can be necessary for or can be used for improving the transmitted signal at the receiving device. This need is more apparent in fast vehicular environments where there is an increasing requirement to provide communication means to vehicles, either vehicle-to-vehicle or vehicle-to-infrastructure, for several applications such as road safety, route planning, congestion avoidance, etc... To this end, the IEEE802 standardization committee has recently developed the IEEE802.11p standard that defines the physical (PHY) and medium-access layers of the wireless communications. Other professional bodies have started developing the standards for the higher layers, to support Intelligent Transportation Systems applications. Some examples include the IEEE 1609 WAVE standard and the ISO TC204 WG16 CALM standard. The IEEE802.11p has been developed within the IEEE802.il Working Group, which also developed previously some very successful WLAN standards such as the IEEE802.11a/g, based on coded OFDM. As a result, there is a strong focus from the Working Group to reuse as much as possible of the PHY layer from the IEEE802.11a/g standards. To that extent, the IEEE802.11p PHY layer is actually identical to that of the IEEE802.Ha/g except for the baseband sampling rate, which is 5 or 10 M Hz instead of 20 MHz. This has been achieved by keeping all parameters of the OFDM modulation and dividing the sample rate by two or four. The frame structure, containing a long training sequence in the preamble for channel estimation and pilot subcarriers in the data portion of the frame, is also unchanged in the IEEE802.11p.

[0003] However, doing so, the waveform design is not very well adapted to fast time-varying channel conditions as may occur in high speed traffic conditions. Indeed, at higher speeds, the channel varies so fast that it changes between the start and the end of the frame. Hence, the equalizer, which is computed at the start of the frame, is not performing well as the channel gets outdated. Because the number of pilots is too small, there are no means provided by the PHY to suitably track the channel coefficient variations until the end of the frame. This then translates into a lot of symbol and bit errors at the demodulator output and ruins the packet error rate (PER) performance.

[0004] Existing literature has identified and partially addressed this issue. Results of BPSK and

QPSK in urban and motorway channel have been presented in the paper "Physical layer simulations of IEEE802.11a for vehicle-to-vehicle communications" (J. Maurer et al., 62nd IEEE Vehicular Technology Conference, VTC-2005-Fall, vol. 3, Sep. 2005), whereby a significant flooring of the performance is reported. Other work has proposed a mid-amble to better track the channel evolution, but this frame structure has not been adopted by the IEEE802.11p standard. A coded decision-directed channel estimation has been proposed for terrestrial digital multimedia broadcasting (DMB). This approach however is not suited for IEEE802.11p because it introduces a large latency at the receiver. Also a decision-directed channel estimation for IEEE802.11a in a mobile channel has been suggested, but the proposed solution does not provide means to reduce the impact of wrong decisions.

[0005] US2009/268782 presents a multi-stage channel frequency response estimation method for multi-band OFDM-based UWB systems. A channel estimation sequence in the preamble is employed to obtain an estimate of the channel frequency response, which is then smoothed. A further estimate is obtained using some OFDM symbols of the frame header. Again a smoothing operation is performed. However, the proposed method only performs in an acceptable way as long as channel variations occur at a slower pace than the duration of a frame structure.

[0006] In US2008/232497 channel tracking in an OFDM wireless receiver for indoor Wi-Fi is discussed. The application is concerned with finding a solution for improved channel estimation. The proposed method exploits that a packet of information includes a known transmitted part having known subcarrier values. On channel estimations derived from said known data a smoothing is performed (see e.g. fig.8 of the document). Channel correction is performed and a hard decision is made. Based on comparing the pre-decision and post-decision constellation values a measure of the channel drift is derived, which is exploited in the channel estimation update. Although this approach indeed leads to improved channel estimation, its application is limited to e.g. indoor Wi-Fi as the channel changes then occur at a relatively slow pace.

[0007] Hence, there is a need for a technique to derive channel knowledge that is adapted for use in case of fast varying channel conditions.

Summary of the invention

[0008] It is an object of embodiments of the present invention to provide for a method and receiver structure for performing channel tracking that can be used in fast time-varying channel conditions. [0009] The above objective is accomplished by a method and receiver structure according to the present invention.

[0010] In a first aspect the invention relates to a method for tracking a time-variant communication channel in fast vehicular environment. The method comprises the steps of :

a) receiving a data structure transmitted over the time-variant communication channel, wherein the received data structure comprises a preamble having a first portion of known data symbols and a plurality of payload data signals, each of the payload data signals containing a second portion of known data symbols and a portion of unknown data symbols,

b) performing an initial estimation of the communication channel using the first portion of known data symbols of the preamble and taking the initial estimation as current channel estimation,

c) equalizing a payload data signal of the plurality taking into account the current channel estimation, yielding an estimate of the corresponding transmitted signal,

d) slicing the estimate of the corresponding transmitted signal,

e) performing a further estimation of the communication channel taking into account the sliced estimate of the corresponding transmitted signal and the second portion of known data symbols of the payload data signal,

f) performing a smoothing of said further estimation, thereby obtaining a new current channel estimation,

g) repeating steps c, d, e and f for a subsequent received payload data signal of the plurality using the current channel estimation.

[0011] The proposed channel tracking method employs an initial channel estimation derived from the known pilot tones present in the preamble. This initial estimate is exploited when processing the first payload data signal after the preamble. For operating on the further data signals (i.e. the other payload data in the data structure) the solution relies both on data-aided tracking (i.e. on the known pilots in the data signals) and on decision-directed tracking (i.e. on the unknown data symbols themselves). Based on the available estimation of the communication channel the received data signal being processed is equalized. After that hard decisions on the data symbols in the signal has been obtained, they are used together with the known symbols on the pilot tones to perform a further, updated estimation of the channel. However, before using this estimate with a next data signal, the updated further estimation undergoes a smoothing operation. This is an essential element to achieve a good performance. Indeed, by exploiting channel smoothing the channel estimation is considerably improved, especially at low signal-to-noise ratios. [0012] In a preferred embodiment the smoothing of the further estimation is performed in the time domain.

[0013] Preferably channel length is taken into account when performing the smoothing of said further estimation of the communication channel.

[0014] In an embodiment of the invention a smoothing operation is performed on the initial estimation. Although the overall effect is smaller than that of smoothing the decision-directed further estimation, such smoothing of the initial estimation is beneficial for noise reduction.

[0015] In a preferred embodiment the data structure is received with a plurality of antennas.

In an embodiment a measurement indicative of the received power is then performed on each of the antennas. Advantageously maximum ratio combining is performed.

[0016] In another embodiment at least one of the unknown data symbols is not considered in the step of smoothing.

[0017] In a preferred embodiment the payload data signals are OFDM signals.

Advantageously the payload data signals follow the I EEE 80211a standard or the I EEE 80211p standard.

[0018] In a second aspect the invention relates to a receiver structure for tracking a time- variant communication channel in a fast vehicular environment. The receiver structure comprises

- means for receiving a data structure transmitted over the time-variant communication channel, said received data structure comprising a preamble having a first portion of known data symbols and a plurality of payload data signals, each of said payload data signals containing a second portion of known data symbols and a portion of unknown data symbols,

- computation means for performing an initial estimation of the communication channel using said first portion of known data symbols and taking said initial estimation as current channel estimation,

- equalizing means for equalizing a payload data signal of the plurality taking into account the current channel estimation, yielding an estimate of the corresponding transmitted signal,

- a slicer for slicing the estimate of the corresponding transmitted signal,

the computation means further being arranged for performing a further estimation of the com munication channel taking into account the sliced estimate of the corresponding transmitted signal and the second portion of known data symbols of said payload data signal, and for performing a smoothing of the further estimation, thereby obtaining a new current channel estimation from the further estimation.

[0019] For purposes of summarizing the invention and the advantages achieved over the prior art, certain objects and advantages of the invention have been described herein above. Of course, it is to be understood that not necessarily all such objects or advantages may be achieved in accordance with any particular embodiment of the invention. Thus, for example, those skilled in the art will recognize that the invention may be embodied or carried out in a manner that achieves or optimizes one advantage or group of advantages as taught herein without necessarily achieving other objects or advantages as may be taught or suggested herein.

[0020] The above and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter.

Brief description of the drawings

[0021] The invention will now be described further, by way of example, with reference to the accompanying drawings, wherein like reference numerals refer to like elements in the various figures, and in which:

[0022] Fig.l illustrates the data frame structure of the IEEE802.11p PHY (time values for the

10MHz channel spacing; for the 5M Hz channel spacing, all time values must be doubled). The equalizer is computed based on the channel estimation performed at the beginning of the frame.

[0023] Fig.2 represents a typical channel response plotted for all symbol indexes within a burst.

[0024] Fig.3 represents one embodiment of the proposed method of combined decision- directed channel estimation and channel smoothing.

[0025] Fig.4 shows the performance comparison of the conventional and advanced decision- directed (DD) receiver with and without smoothing, for the SISO, AS and M C case. Simulation conditions are Q.PSK, highway channel at 150km/h.

[0026] Fig.5 illustrates a performance comparison of the conventional and advanced (DD) receiver with and without smoothing, for the SISO, AS and M C case. Simulation conditions are Q.PSK, urban channel at 50km/h.

[0027] Fig.6 represents the performance comparison of the conventional and advanced (DD) receiver with and without smoothing, for the SISO, AS and M RC case. Simulation conditions are 16QAM, highway channel at 150km/h.

[0028] Fig.7 represents a performance comparison of the advanced receiver with smoothing, for ns = 52, 28 and 16, for the SISO, AS and M RC case. Simulation conditions are QPSK, highway channel at 150km/h.

Detailed description of illustrative embodiments

[0029] The present invention will be described with respect to particular embodiments and with reference to certain drawings but the invention is not limited thereto but only by the claims. [0030] Furthermore, the terms first, second and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a sequence, either temporally, spatially, in ranking or in any other manner. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other sequences than described or illustrated herein.

[0031] It is to be noticed that the term "comprising", used in the claims, should not be interpreted as being restricted to the means listed thereafter; it does not exclude other elements or steps. It is thus to be interpreted as specifying the presence of the stated features, integers, steps or components as referred to, but does not preclude the presence or addition of one or more other features, integers, steps or components, or groups thereof. Thus, the scope of the expression "a device comprising means A and B" should not be limited to devices consisting only of components A and B. It means that with respect to the present invention, the only relevant components of the device are A and B.

[0032] Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment, but may. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner, as would be apparent to one of ordinary skill in the art from this disclosure, in one or more embodiments.

[0033] Similarly it should be appreciated that in the description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. This method of disclosure, however, is not to be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.

[0034] Furthermore, while some embodiments described herein include some but not other features included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention, and form different embodiments, as would be understood by those in the art. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0035] It should be noted that the use of particular terminology when describing certain features or aspects of the invention should not be taken to imply that the terminology is being redefined herein to be restricted to include any specific characteristics of the features or aspects of the invention with which that terminology is associated.

[0036] In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the invention may be practiced without these specific details. In other instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

[0037] A solution is proposed for performing channel tracking in fast vehicular environments for OFDM systems, comprising an advanced receiver scheme that updates the channel during the burst in a decision-directed fashion. The proposed scheme is combined with channel smoothing and optionally with spatial diversity for improving the receiver performance for performing well at low SN s.

[0038] Regarding notational conventions, normal latin characters for time-domain signals (a)

(fl

and tilde characters for frequency-domain signals ^v " ^■ ' ; vectors and matrices are denoted by a single and double under-bar respectively ^nC ' =^ ; the superscripts ^T and ^H denote the matrix

(A^ and A^) t transpose and complex conjugate transpose respectively = = ' ; the superscript is used to indicate the pseudo-inverse —

[0039] A vehicular communication system is considered that uses OFDM as defined in the

IEEE802.11p standard with a data frame structure as illustrated in Fig. 1. The frame structure includes a preamble (known symbols) for initial channel estimation and a payload comprising both unknown data and a few pilots (known symbols). The first payload data block in Fig.l ('payload data #1') is in the IEEE802.11p standard known as the SIGNAL field. It is a packet that is modulated at a relatively low data rate. Information on the carriers in the Signal field indicates amongst other things which type of modulation must be used for the subsequent communication. The communication standard is intended for vehicle-to-vehicle or vehicle-to-roadside communications in the upper 5GHz band. At each OFDM symbol time interval, a symbol ^{* q} is transmitted over the channel, wherein the subscript q denotes the OFDM symbol index. Note that this symbol has a size N but consists of N _D data-sub-carriers, N _P pilot sub-carriers and N _GL +N _GR +1 sub-carriers with zero amplitude for the left and right guard bands and DC sub-carrier. This leads to N = N _D + N _P + N _GL + N _GR + 1. At every time index q, the frequency-domain symbol vector ^q is converted by a size N inverse discrete Fourier Transform (IDFT) into a size N vector of time-domain samples, which is then prepended with a length N _C p cyclic prefix and parallel-to-serial converted, resulting in the discrete-time transmitted signal

₁ N- l

S _q (n) = —=∑ £ _q (k)e ^j27Tnk ^ ^N , -NcP≤n<N- l (1)

ν ^Λ k=0

The linear time-variant (LTV) channel is modelled by the quasi-static time-variant discrete impulse response h _q (l) assuming that the channel remains constant during one OFDM symbol but varies from one symbol to the next. In h _q (l), I is the delay index and q is the OFDM symbol index.

The time-domain received samples for the q-th OFDM symbol are given by :

L-l

'i ^' q (n) =∑ h _q (l)s _q (n - I) + n _q (n) , -N _C p≤n<N- i (2)

1=0 where the additive noise n _q (n) are complex samples of a circularly white complex Gaussian noise process with variance σ%- After serial-to-parallel conversion, cyclic prefix removal and discrete Fourier

y

transform (DFT), the elements of the frequency-domain received signal vector ^— ' are expressed as :

1 ^{iV _ 1}

y _q (m) = -J= _t r _q (n)e- ^j27Tnm/N , o<m<N- i (3)

All these operations can be compactly written in matrix form:

2 g ±— ^" ll p ^" [ ^jj -g ^* — p ^' E=z ' — 9 — <? '

= H · X + W„ (4)

T E F

where =CP and =CP are the matrices for cyclic prefix addition and removal respectively , = is the unitary Fourier matrix and — is the frequency-domain noise vector with the same statistics as the noise n because of the unitary nature of = F . When the channel is static during the frame, the channel can be estimated at the first transmitted OFDM symbol and this estimation can be used throughout the frame for equalization. However, when the channel is not static, the channel must be estimated - or at least tracked - at every OFDM symbol.

Equalization

[0040] With the assumption that the channel is static during one OFDM symbol, the matrix

H_ = F · /? · H_ ^■ T , · F ^H

9 ⁽ ' ⁽ l — CP — is exactly diagonal and the equalization can be performed as in conventional static OFDM systems with a single complex multiplication per sub-carrier. Hence, the per sub-carrier equalizer coefficient 9 ^ ' is simply given by

The per sub-carrier equalization process is then simply: x _q (m) = e _q (m) ^■ y _q {m) . (6)

In this case both the equalizer coefficients computation (5) and the equalization itself (6) are low complexity operations since the equalizer coefficients (5) computation involves N _D + N _P complex divisions at the preamble time and the equalization (6) involves N _D complex multiplications per OFDM symbol.

Preamble-based channel estimation

[0041] In the IEEE802.11a and IEEE802.11p standards, the channel estimation is based on a preamble P defined in the frequency-domain; it occupies N _s = N _D + N _P sub-carriers (corresponding to the data and pilot sub-carriers) and uses BPSK modulation. Based on the signal model outlined in equation (4), the received preamble can be expressed in the frequency-domain as z_ = H · p + n

= P ' h + n

= E - E - h + n (7) where = H and =P are diagonal matrices containing the entries of vectors — h and p respectively, and — and — are the frequency and time-domain channel responses to be estimated, respectively. From a frequency-domain perspective only, the maximum-likelihood (M L) estimate of the channel is given by

¾ = Ζ ^{_ 1} - £ (8)

where the inverse of = p is trivial since = p (P

is a diagonal matrix of BPSK symbols — -— P) .

[0042] Generally speaking, in IEEE802.11p systems at typical urban (50km/h) or highway (up to 150km/h) velocities, the channel changes quite rapidly. This is illustrated in Fig. 2, which shows the frequency responses at all OFDM symbol indices from 1 to 30 for a typical highway channel at 150km/h. As can be seen, the frequency response measured on the first OFDM symbol (as drawn with a thick line) cannot be used after only a few OFDM symbols. Considering a typical system operation at 5.9GHz and at a speed of 150km/h, the maximum Doppler shift is approximately 800Hz, which is about 0.5% of the sub-carrier spacing. This creates a little amount of inter-carrier interference (ICI) and requires channel tracking. Unfortunately, the I EEE802.11p standard has not foreseen to increase the number of pilot signals to track the channel response variations. This causes a performance degradation of the pilot-based tracking system, which in such circumstances is unacceptable. Thus, the proposed system and method have been developed to address the problem of channel tracking at high velocities without compromising the receiver performance.

Channel Tracking

[0043] To cope with channel variations, a channel tracking method has been devised that relies both on data-aided tracking (i.e. on the known pilots) and on decision-directed tracking (i.e. on the unknown data themselves). The principle of operation is as follows. At the beginning of the burst, the received preamble is used to perform the initial channel estimation as outlined above, providing

= h0. Then, for every OFDM symbol with index q, the following operations are performed :

the current OFDM s mbol is equalized with the channel estimate of the previous OFDM the equalized vector is sliced (to obtain a hard decision) providing ^~ 9 the new channel — is estimated, using the known pilot values ¹ 5 and the estimated x

hard symbols Ί . Advanced channel tracking - smoothing

[0044] In order to improve the receiver performance at low SNR, the proposed channel tracking method is combined with a channel smoothing technique. This is because a decision-directed mechanism suffers from wrong decisions occurring more frequently at low SNR. As a result, errors occurring in the data decision process create localized peaks in the channel response. When channel smoothing is combined with decision-directed based channel tracking, those peaks are strongly attenuated (smoothed out) and the channel tracking behaves correctly at much lower SNR values. The smoothing removes part of the estimation noise by exploiting the propagation physics: the time duration of the channel impulse response is limited because of the short distances involved (a few hundreds of nanoseconds due to reflections). However, the OFDM symbol duration - and also that of the preamble - is 64 samples or 8s for a 10MHz bandwidth and 16s for a 5M Hz bandwidth. Hence, the estimated impulse response duration can be limited to a certain number L, typically lesser or equal to 10. The improved channel estimate with smoothing is derived as follows, whereby the last line in expression (7) can be rewritten as z_ = Ρ_ · F_ - h - - n = Ρ_ · F_ ^■ h _L + n (9)

Where = FL is a reduced Fourier matrix resulting from — F by taking its first L columns and its N _s rows corresponding to the N _s sub-carrier positions used and — L is a reduced — vector where only the L h

first taps are non-zero. From (9), the time-domain ML estimate — L with the time-domain constraint becomes:

L

= (F_ · F ) ^{_1 ■} F_ ^■ P · i (10) and it can easily be transformed into the frequency-domain ML estimate —L with the time-domain constraint :

L= F _T > h _Tt = F _r · ( - _r ) ^_1 - Ff · Ρ·Ι (1 1)

The combination of the decision-directed channel estimation and the channel smoothing is illustrated in Fig.3.

Advgnced chgnnel tmcking - spgtigl diversity [0045] The performance of the proposed method is ultimately limited by deep nulls in the frequency response where the channel response is less accurate and the noise enhancement in the equalization process is dominant. When the channel undergoes deep fades, a very effective way to further improve the performance of the receiver is to resort to multiple antenna reception. Receive antenna diversity techniques can provide significant gains (without any modification in the transmitted signal) provided that the antennas are sufficiently decorrelated. There are several forms of antenna diversity. The simplest one is antenna selection (AS) and a more sophisticated one is maximum-ratio combining (MRC). In AS the power received by every receive antenna is compared and the antenna receiving the strongest signal is selected for reception. In M RC the signal received by each antenna is weighted by a factor proportional to the channel coefficient and all weighted signals are summed up in phase. In an OFDM system, this can be more advantageously performed in the frequency domain with a maximum ratio combining per sub-carrier.

[0046] The two spatial diversity methods, AS and M RC, are explained in more details below.

The principle of antenna selection is very simple: it involves measuring the power received (or SNR) on each receive antenna in the time domain (e.g. on the preamble) and switching to the antenna with the highest received power (or SNR) and then continuing as for the single antenna case. This requires some buffering. Hence, only one receiver chain is needed. For OFDM system, this can be performed in the time-domain or in the frequency-domain. However, frequency-domain AS is not very meaningful since it requires two complete receiver chains, at least up to the FFT output. Hence, only time-domain AS is considered for the proposed channel tracking method.

For Maximum Ratio Combining the system model in (2) and (3) has to be expanded as follows, to account for the use of several receive antennas. The receive signal vector at sample time n is given by: and the frequency-domain vector at sub-carrier m is given by:

The MRC equalizer coefficients are computed as follows:

Hence, one needs to update the channel estimation per antenna (as in the single antenna case), then compute the equalizer coefficients (this is now a vector per sub-carrier as in equation (14) and then multiply per-sub-carrier this equalizer vector with the received vector (13), this providing the equalized received signal per sub-carrier (similar to equation (6) but e and y are vectors, x is still a complex scalar).

Complexity reduction of the smoothing

[0047] The smoothing operation can be very demanding, since it requires the multiplication of a large matrix with a vector (in IEEE802.11a or IEEE802.11p systems, specifically a 52x52 matrix times a 52x1 vector) at each OFDM symbol. Hence, two methods can be applied to reduce such com plexity:

1) Lossless complexity reduction : For a 52x52 matrix times a 52x1 vector the number of operations required is " S complex multiple-accumulate (CMAC). This matrix multiplication can be significantly reduced by exploiting the special structure of the projection matrix p _, ( pH TP \-l . pH

—— i —— —— / I

Indeed, by defining the singular value decomposition of L :

This leads to:

F - (F ^■ F Γ ¹ · F ¹¹ = U _L · ί (16) because the L singular values of F _L are all equal to 1. Hence and quite interestingly, one can decompose the smoothing operation into two multiplications, the first one being a (L x N _s ) matrix times a (N _s x 1 ) vector and the second being a (N _s x L) matrix times a (L x 1) vector.

Doing so results in 2N _S L CMACs, which is smaller than " S since L is typically equal to 10 or j r"2 less while NS is equal to 52. Hence, the total number of operations is reduced from ⁱ S = 2704 CMACs to 2N _S L = 1040. Note that this simplification does not incur any performance loss. Several other lossless complexity reduction techniques can be applied. One method exploits the structure of U _L . Another option is to exploit the fact that multiplication with (16) can be implemented by means of a DFT, a small matrix multiplication and an IDFT.

2) Lossy complexity reduction: A further complexity reduction is possible at the price of some performance loss. It is possible to estimate the channel through a combined smoothing and interpolation operation. In this case some subcarriers are not taken into account in the smoothing process (and the number of rows in F _L is reduced accordingly), relying on the fact that the frequency response variations are smooth. Practically, the pilot sub-carriers are always used because there is no risk of error on the known pilots. For the data subcarriers, three possibilities can be defined: 12, 24 or 48 subcarriers are used. This result in ns = 16, 28 or 52 sub-carriers used for the channel estimation. In addition, when 12 or 24 sub-carriers only are used, one discards the subset(s) that contain the deepest nulls, where the risk of error is higher. The resulting number of operations is as follows:

a. for ns = 52 sub-carriers: NL + NL = 2NL = 1040 CMACs

b. for ns = 28 sub-carriers: NL + 28L = 800 CMACs

c. for ns = 16 sub-carriers: NL + 16L = 680 CMACs

3) Other complexity reduction: It is important to note that the proposed method for channel tracking can completely replace other forms of tracking commonly used in IEEE802.11a systems, namely tracking of carrier frequency offset, sample clock offset and phase noise. Indeed, those time-variant effects are completely taken care of by the proposed channel tracking scheme.

[0048] As an example, the proposed channel tracking method is applied to an IEEE802.11p system in a vehicular environment. The parameters used in this example are summarized in Table I. All parameters are selected as per the PHY layer definition of the IEEE802.11p standard. The channel model is based on the paper "Vehicle-vehicle channel models for the 5GHz band" (I. Sen and D. Matolak, IEEE Trans. Intelligent Transportation Systems, vol. 9, no. 2, pp.235-245, Jun. 2008), where a highway channel with a vehicle speed of 150km/h and an urban channel with a vehicle speed of 50km/h are used. A packet length of 300 bytes is assumed, which is a representative size for cooperative awareness or decentralized environment notification messages, when a security header is included. The packet error rate (PER) is used as the measure of system performance because it is more meaningful at the application level. The same trend can however also be observed on bit error rate curves. Carrier frequency 5.9GHz

Bandwidth 10MHz

FFT size (N) 64

Cyclic prefix size (N p) 16

OFDM symbol duration with CP 8//s

Number of data sub-carriers (ND ) 48

Number of pilot sub-carriers (Np) 48

sub-carrier spacing 156.25kHz

modulation QPSK and 16QAM

coding convoliitional, rate 3/4

channel type urban (50km/h)

highway (150km/h)

packet size 300 bytes (2400 bits)

number of channel realizations 5000

S IMULATION PARAMETERS

Comparison of receiver processing scheme

The PER performance for three different receiver designs and for highway and urban channels is illustrated in Figures 4 and 5. The different receiver designs and channels used in the simulation are described in more details below :

• Conventional receiver as for the IEEE802.11a (labelled as "conventional" in the legend). The receiver performs channel estimation on the preamble and uses the pilot sub-carriers to track any non-static behaviour (carrier frequency offset, sampling clock offset and phase noise).

• The proposed receiver design with decision-directed based tracking only (labelled as "DD, smooth OFF") that updates the channel estimation at every OFDM symbols.

• The proposed receiver design with decision-directed based tracking and channel smoothing (labelled as "DD, smooth ON") that updates the channel estimation at every OFDM symbols.

The results for SISO, AS and M C are shown in Figs. 4 and 5. From the figures the following observations can be made.

• First, the decision-directed based tracking without smoothing does improve the performance but only at high SNR, because the errors at low SNR hamper the decision-directed mode.

• Second, the decision-directed mode combined with the smoothing is the key technique to obtain good performance and avoid the flooring. • Third, the SISO performance can be further improved by using AS or MRC (M RC is the best but requires two complete receive chains up to the equalizer).

[0049] Fig.6 also illustrates the results for 16Q.AM. All parameters are identical to those used for the results shown in Figure 4 except that the modulation is 16Q.AM instead of QPSK. These results demonstrate that, first, the combined decision-directed and smoothing channel estimation also works for 16Q.AM although it has a non-constant envelope and, second, that there is a SNR difference of about 6dB between QPSK and 16QAM. This is a known result that is due to the constellation points of 16QAM that lie closer to each other. [0050] While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive. The foregoing description details certain embodiments of the invention. It will be appreciated, however, that no matter how detailed the foregoing appears in text, the invention may be practiced in many ways. The invention is not limited to the disclosed embodiments.

[0051] Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single processor or other unit may fulfil the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems. Any reference signs in the claims should not be construed as limiting the scope.