WO2005048552A1

WO2005048552A1 - Methods and apparatuses for dc offset estimation in ofdm systems

Info

Publication number: WO2005048552A1
Application number: PCT/IB2004/052413
Authority: WO
Inventors: Yifeng Zhang; Charles Razzell
Original assignee: Koninklijke Philips Electronics NV
Current assignee: Koninklijke Philips NV
Priority date: 2003-11-13
Filing date: 2004-11-13
Publication date: 2005-05-26
Anticipated expiration: 2006-05-13

Abstract

Systems and methods according to the present invention address provide communication systems that perform sequential detection and compensation of frequency offset and DC offset. This can be accomplished, for example, using symbol in a short preamble of an OFDM signal. Sequential detection and compensation of FRO/DCO according to the present invention provides an accurate and fast mechanism for removing DCO that results in a low noise floor and accurate and fast mechanism for removing DCO that results in a low noise floor and accurate received symbol demodulation.

Description

METHODS AND APPARATUSES FOR DC OFFSET ESTIMATION IN OFDM SYSTEMS The present invention relates generally to wireless communication systems and, more particularly, to direct current (DC) offset estimation in orthogonal frequency division multiplexed (OFDM) wireless communication systems. Technologies associated with the communication of information have evolved rapidly over the last several decades. For example, over the last two decades wireless communication technologies have transitioned from providing products that were originally viewed as novelty items to providing products which are the fundamental means for mobile communications. Perhaps the most influential of these wireless technologies were cellular telephone systems and products. Cellular technologies emerged to provide a mobile extension to existing wireline communication systems, providing users with ubiquitous coverage using traditional circuit-switched radio paths. More recently, however, wireless communication technologies have begun to replace wireline connections in almost every area of communications. Wireless local area networks (WLANs) are rapidly becoming a popular alternative to the conventional wired networks in both homes and offices. Many of today's WLAN systems operate in accordance with the IEEE 802.1 lb standard. As will be appreciated by those skilled in the art, IEEE 802.11 specifies that WLAN devices will use one of two spread spectrum access methodologies, specifically either frequency-hopping or code spreading. In frequency hopping systems, a wireless connection between two WLAN units will periodically change frequencies according to a predefined hop sequence. In code spreading (also sometimes referred to as "direct sequence spreading"), the wireless data signal is spread across a relatively wideband channel by, for example, multiplication with a pseudorandom noise (PN) sequence. Other WLANs are designed in accordance with the IEEE 802.11 a or 802.11 g standards. These standards provide for the transmission of signals using orthogonal frequency division multiplexing (OFDM). In OFDM systems, a signal is split into several narrowband channels each of which is transmitted at a different frequency. The detection of an OFDM signal is based on the orthogonality of the sub-carriers. Orthogonality, in turn, depends upon both the precision at which the sub-carrier frequencies are transmitted and received and the integration interval. For a given integration interval, variance between the ideal sub-carrier frequencies and the sub-carrier frequencies used by the transmitter and receiver will introduce noise which is unrecoverable after the performance of a Fast Fourier Transform (FFT) at the receiver side to recover the sub- carriers. Such variances are unavoidable due to the fact that transmitters and receivers will inherently have local oscillators (LOs) with different frequencies due to differences in, e.g., the crystals used to fabricate the LOs. To compensate for this problem, OFDM systems can perform a frequency de-rotation after mixing to remove the frequency offset (FRO) between the transmitter and receiver. In addition to the aforedescribed FRO, OFDM systems may also suffer from DC offset which can be introduced due to, for example, self-mixing of the LO, even-order IM products of signal/interference and mismatch in analog circuitry. If a DC offset is present, then it will be mixed up to the offset frequency by the de-rotator and act as a jammer tone to the OFDM sub-carriers. This, in turn, will negatively impact the orthogonality of the subcarriers and result in unrecoverable inter-carrier interference (Id). One conventional technique which has been used to compensate for DC offset in FM communication systems is to employ a notch filter to remove the DC component. Either an analog or a digital notch filter can be used. However the notch filter suffers from, among other things, (1) its own introduction of ICI, (2) attenuation to the first several OFDM sub-carriers due to FRO and (3) the potential for an unacceptably long signal level settling time. Another conventional technique for compensating for DC offset is dynamic adjustment of the receiver's LO. In this technique, the difference between the LO frequency of the transmitter and the LO frequency of the receiver is estimated and the estimated FRO is then used to adjust the LO frequency at the receiver. However, this scheme will have difficulty in ad-hoc operation mode where all the stations work as access points and alignment of each other's LO will lead to confusion to the system. Also, such a scheme requires LOs to have fine frequency resolution which may not always be possible. Accordingly, it would be desirable to provide techniques and devices for providing frequency and DC offset compensation which avoid the problems of conventional techniques. Systems and methods according to the present invention address this need and others by providing communication systems that perform sequential detection and compensation of frequency offset and DC offset. This can be accomplished, for example, using symbols in a short preamble of an OFDM signal. Sequential detection and compensation of FRO/DCO according to the present invention provides an accurate and fast mechanism for removing DCO that results in a low noise floor and accurate received symbol demodulation. According to one exemplary embodiment of the present invention, a method for wireless communication includes the steps of receiving a signal having a preamble, determining a frequency offset using the preamble, removing the frequency offset from the signal to generate a frequency corrected signal, determining a DC offset associated with the frequency corrected signal, removing the DC offset from the frequency corrected signal to generate a DC corrected signal; and demodulating symbols from the DC corrected signal. According to another exemplary embodiment of the present invention, a receiver includes a mixer for mixing a received signal down to a DC signal, a sequential frequency offset/DC offset compensation unit for (a) determining, and compensating for, a frequency offset associated with the received signal to generate a frequency corrected signal, and (b) determining, and compensating for, a DC offset associated with the frequency corrected signal to generate a DC offset corrected signal and a demodulator for determining symbols associated with the received signal based on the DC offset corrected signal. The accompanying drawings illustrate exemplary embodiments of the present invention, wherein: FIG. 1 depicts a WLAN system in which the present invention can be implemented; FIG. 2 depicts an exemplary OFDM frame format which can be used in conjunction with an exemplary embodiment of the present invention; FIG. 3 depicts a transceiver which provides sequential FRO/DCO detection and compensation according to an exemplary embodiment of the present invention; FIG. 4 is a flowchart which illustrates a method for performing sequential FRO/DCO detection and compensation according to an exemplary embodiment of the present invention; FIG. 5 shows simulation results of an exemplary embodiment of the present invention; and FIG. 6 is a histogram showing additional simulation results associated with FIG.5. The following detailed description of the invention refers to the accompanying drawings. The same reference numbers in different drawings identify the same or similar elements. Also, the following detailed description does not limit the invention. Instead, the scope of the invention is defined by the appended claims. In order to provide some context for this discussion, an exemplary WLAN system will first be described with respect to Figure 1. Those skilled in the art will appreciate, however, that the present invention is not restricted to implementation in WLAN systems. Therein, a wireline network 10 (e.g., an Ethernet network) has a file server 12 and workstation 14 connected thereto. Those skilled in the art will appreciate that typical wireline networks will serve numerous fixed workstations 14, however only one is depicted in Figure 1 for simplicity. The wireline network 10 is also connected to a WLAN 16 via router 18. The router 18 interconnects the access points (AP) of the WLAN 16 with the wireline network, through which the access points can, for example, communicate with the file server 12. In the exemplary WLAN system of Figure 1, three cells 20, 22 and 23 (also sometimes referred to as a Basic Service Set (BSS) or Basic Service Area (BSA) are shown each with a respective AP, although those skilled in the art will once again appreciate that more or fewer cells may be provided in WLAN 16. Within each cell, a respective AP serves a number of wireless stations (W) via a wireless connection. According to exemplary embodiments of the present invention, the transmission of signals between APs and respective wireless stations W is performed using OFDM signals, e.g., in accordance with IEEE 802.1 la or 802.1 lg. Devices and methods according to exemplary embodiments of the present invention provide techniques for compensating for frequency offset and DC offset present in such OFDM signals using sequential estimation of these two characteristics. Since the DC offset will, absent compensation, be mixed up to the offset frequency during de-rotation, the ICI which would be introduced by the DC component is predictable given prior knowledge of the frequency offset. To better understand sequential FRO/DCO offset compensation according to the present invention consider an OFDM signal transmitted between a transmitter (e.g., an AP or a W) and a receiver (e.g., a W or an AP) in Figure 1. For this example, the OFDM signal will have the repeating frame structure seen in Figure 2, although those skilled in the art will appreciate that other frame structures can be used in conjunction with the present invention. Therein, a frame 24 includes a short preamble 26 having ten symbols S, a long preamble 28 having two symbols S2 and a cyclic redundancy field, and a data portion 29. The short preamble field 22 and long preamble field 24 can both have a period of 8 μs. More detail regarding an exemplary transceiver 30 according to the present invention can be seen in Figure 3. Therein, the OFDM signal of Figure 2 can be received via antenna 32. If the antenna is shared between the transmit chain 34 and the receive chain 36, an antenna switch 38 can be used to switch the antenna 32 between the two processing chains. The transmit chain 34 includes lowpass filters 70 and 72, mixers 74 and 76, phase shifter 78, RF VGA 80, PA driver 82, highpass filter 84 and power amplifier 86. On the receive side, the OFDM signal first passes through a bandpass filter 40 to reject noise disposed outside of the desired frequency band and a low-noise amplifier (LNA) 42 to amplify the received signal. The output of LNA 42 is fed into a mixer which includes in-phase (I) and quadrature (Q) components 44 and 46, respectively. The I and Q mixers are supplied with local oscillating signals which are 90 degrees out of phase with respect to one another via synthesizer 48, voltage controlled oscillator 49 and phase shifter 50. The resulting I and Q signals which are output from mixers 44 and 46 are then input to a two stage automatic gain control (AGC) unit. In each of the I and Q signal processing chains, the respective signal is first input to an adjustable gain amplifier 52, 54 and then to a low pass filter (LPF) 56, 58. A second adjustable gain amplifier in each of the I and Q signal processing chain (60 and 62, respectively) then amplifies the output of the respective LPF 58, 56. The second adjustable gain amplifier in each of the signal processing chains performs a second AGC pass after the signal has been fitted to the window of the (downstream and not shown) A/D converter. In addition, a coarse DC offset cancellation is performed by unit 64. This is accomplished by, in this example, a look-up table which provides an adaptive DC offset cancellation based on the gain setting of the amplifiers 52 and 54. The coarse DC offset cancellation unit 64 provides sufficient DC offset compensation to avoid saturating the baseband modem 66. Sequential FRO/DCO according to exemplary embodiments of the present invention is performed by unit 68 in the baseband modem 66. To understand one exemplary technique for sequential FRO/DCO processing according to the present invention, consider that the signal s(k) received by receiver 12 can be represented as:

where c_m is the QAM symbol on the nth sub-carrier of the OFDM signal, //is the frequency offset, d(k) is the DC offset on the kth sample and n(k) is the all white Gaussian noise

(AWGN) on the kth sample. The DC offset and AWGN are complex values represented as: d(k) = di(k) +jdq(k) (2) n(k) = ni(k) +jnq(k) (3) In a noiseless environment, equation (1) can then be rewritten as:

where a = 2πΔfT is the normalized frequency offset. Various techniques for removing the frequency offset can be used at this stage. Examples of such techniques can be found in U.S. Patent Nos. 6,539,063 and 6,501,730, the disclosures of which are incorporated here by reference. After removing the frequency offset, the received signal becomes: 32 -_9rf ^m ■ k_ _Ja c e " + d(k)e^Ja«* s(k) = s(k)e^J « = ^ ^{m ( )} ₍₅₎

QAM symbols on the nth OFDM sub-carrier of received signal s(k) can be recovered by calculating: 1 « jlπn ⁰⁴ *=⁰ (6)

For DC components, n=0, so equation (6) then reduces to:

For a communication system with a constant DC offset or a relatively slowly varying. DC offset (such that the DC offset can be considered to be constant over one Fast Fourier Transform (FFT) period), d(k) =d = d_r+jd_q for all k such that:

Setting uo=u_r+ju_q, d_r and d_q then become: 1 63α 1 . 63α α„ = — u_r cos 1 — u sin β 128 ? ^g 128 (9) 1 . 63a 1 63a d„ = —u_r sm H — u„ cos " β ^r 128 β ^q 128 (10) where, . a sin — 2 β = 64 sin 128 (11)

Equations (9) and (10) can be used in signal processing unit 68 by, e.g., a suitably programmed microprocessor, to provide an estimation of the DC offset for a typical QAM/OFDM data symbol having a 64 FFT period. For the training symbols in the short preamble 26 of Figure 2, a similar calculation can be performed to determine the DC offset of a particular OFDM frame during the processing of the short preamble as: 1 15α 1 . 15a ^ ^M _™ ^C0S + τr^w ₀₅ ^sm d„ = β, 128 β, ^qs 128 (12)

where, . a sin — A — ⁸ 16 sin ¹²⁸ (14)

Therein d_rs and d_qs represent the DC offset of the short training symbol and u_rs and u_qs can be calculated by averaging, for example, 16 samples of the short training symbols after the frequency offset has been removed. Once the DC offset has been calculated for the OFDM frame, it can then be subtracted from the I and Q signals passed to baseband modem 66. This process can be repeated, for example, every frame during the latter portion of the short preamble. An exemplary technique for performing sequential FRO/DCO according to the present invention can be summarized as shown by the flowchart of Figure 4. Therein, at step 100 a signal, e.g., an OFDM signal having a short preamble 26, is received for processing. The frequency offset is first determined (step 102) and removed (step 104). Then the DC offset is calculated, e.g., using equations (12) and (13), above, at step 106. The DC offset can then be removed, e.g., via a subtractor, at step 108 and the symbols demodulated at step 110. Applicants have simulated sequential FRO/DCO techniques according to the present invention, results of which are reproduced as Figures 5 and 6. Therein, coarse frequency offset removal was performed, e.g., as described in the aforementioned patents, prior to performing the calculations described above to determine the DC offset. Figure 5 illustrates the real (I) and imaginary (Q) outputs in millivolts (y-axis) as a function of frame index (x- axis). The histogram of Figure 6 takes the data of Figure 5 and plots it as percentages (y- axis) as a function of millivolts (x-axis) for both the I and Q channels. As a result, it can be seen that the standard deviation of the detection error for this simulation applying an exemplary embodiment of the present invention is less than 0.35% of the peak of the signal level under 27dB SNR conditions, which will generate a very small ICI noise floor to the system. Applicants have further determined that the implementation loss of the afore- described exemplary algorithm is only 0.2dB for 10% PER under any channel conditions. According to other exemplary embodiments of the present invention, the afore- described techniques can also be used to generate a tracking loop to follow slowly varying DC offsets, if they exist. This can be accomplished by using the techniques described above and employing equations (9) and (10) for each OFDM symbol. The above-described exemplary embodiments are intended to be illustrative in all respects, rather than restrictive, of the present invention. Thus the present invention is capable of many variations in detailed implementation that can be derived from the description contained herein by a person skilled in the art. All such variations and modifications are considered to be within the scope and spirit of the present invention as defined by the following claims. No element, act, or instruction used in the description of the present application should be construed as critical or essential to the invention unless explicitly described as such. Also, as used herein, the article "a" is intended to include one or more items.

Claims

CLAIMS What is claimed is: 1. A method for wireless communication comprising the steps of: receiving (100) a signal having a preamble; estimating (102) a frequency offset using said preamble; removing said frequency offset (104) from said signal to generate a frequency corrected signal; determining a DC offset (106) associated with said frequency offset estimation and the corrected signal; removing said DC offset (108) from said frequency corrected signal to generate a DC corrected signal; and demodulating symbols (110) from said DC corrected signal.

2. The method of claim 1, wherein said signal is an orthogonal frequency division multiplexed (OFDM) signal.

3. The method of claim 1, wherein said frequency offset is associated with a difference between predetermined sub-carrier frequencies and sub-carrier frequencies used by a transmitter which transmitted said signal and a receiver which received said signal.

4. The method of claim 2, wherein said signal has a short preamble (26) and a long preamble (28) and wherein said preamble used to compensate for said frequency offset and said DC offset is said short preamble .

5. The method of claim 4, wherein said short preamble (26) includes 10 training symbols transmitted over an 8μS period and said long preamble (28) includes 2 training symbols transmitted over an 8μS period.

6. The method of claim 1, wherein said step of removing said frequency offset (104) further comprises the step of: de-rotating said signal by said determined frequency offset to generate said frequency corrected signal.

7. The method of claim 1, wherein said step of determining said DC offset

(106) further comprises the step of: . . , 1 15a , 1 . 15α , , 1 . 15a 1 15α calculating α, = — w cos + — w„_c sm and d_as = — w„ sin 1 M COS " β_s " 128 β_s ^qs 128 ^q β_s ^rs 128 β_s ^qs 128

where d_rs and d_qs are components of said DC offset, u_rs and u_qsd Q real and imaginary components of a short preamble symbol value, α is normalized frequency offset and

. a sm —

A — ⁸ a 16 sin 128

8. The method of claim 1, further comprising the step of: coarsely compensating for said DC offset prior to determining said frequency offset.

9. A receiver comprising: a mixer (44,46) for mixing a received signal down to a DC signal; a sequential frequency offset/DC offset compensation unit (68)for:

(a) determining, and compensating for, a frequency offset associated with said received signal to generate a frequency corrected signal; (b) determining, and compensating for, a DC offset associated with said frequency corrected signal to generate a DC offset corrected signal; and a demodulator (66) for determining symbols associated with said received signal based on said DC offset corrected signal.

10. The receiver of claim 9, wherein said signal is an orthogonal frequency division multiplexed (OFDM) signal.

11. The receiver of claim 9, wherein said frequency offset is associated with a difference between predetermined sub-carrier frequencies and sub-carrier frequencies used by a transmitter which transmitted said signal and said receiver.

12. The receiver of claim 10, wherein said signal has a short preamble (26) and a long preamble (28)and wherein said preamble used to compensate for said frequency offset and said DC offset is said short preamble (26) .

13. The receiver of claim 12, wherein said short preamble (26) includes 10 training symbols transmitted over an 8μS period and said long preamble (28) includes 2 training symbols transmitted over an 8μS period.

14. The receiver of claim 9, wherein said sequential frequency offset/DC offset compensation unit (68) includes a de-rotator for de-rotating said received signal by said determined frequency offset to generate said frequency corrected signal.

15. The receiver of claim 9, wherein said sequential frequency offset/DC offset compensation unit (68) determines said DC offset by calculating: , , _A. , 1 15α 1 . 15a , , 1 . 15α 1 15a calculating , = — « cos + — u_n sm and d_as = — w„ sιn h — u„ cos , ^S " β, ^rs 128 β, ^qs 128 ⁹ β_s ^rs 128 β, ^qs 128 where d_rs and d_qs are components of said DC offset, u_rs and u_qs are real and imaginary components of a short preamble symbol value, α is normalized frequency offset and . a sin —

A =- ⁸ a 16 sin 128

16. The receiver of claim 9, further comprising: a coarse DC offset compensation unit (64) for coarsely compensating for said DC offset between said mixer and said sequential frequency offset/DC offset compensation unit (68).