Multiplex communications – Communication techniques for information carried in plural... – Combining or distributing information via frequency channels
Reexamination Certificate
1998-09-11
2002-05-14
Chin, Wellington (Department: 2664)
Multiplex communications
Communication techniques for information carried in plural...
Combining or distributing information via frequency channels
C370S295000, C370S319000, C455S136000, C375S350000
Reexamination Certificate
active
06389040
ABSTRACT:
TECHNICAL FIELD
This invention relates to communication systems and, more particularly, to burst communication systems.
BACKGROUND OF THE INVENTION
Communication channels, and particularly, wireless communication channels, are subject to channel impairments such as multipath propagation, i.e., spread, and fading, in addition, to additive noise. Carrier frequency offset that typically occurs because of transmitter and receiver oscillator mismatch in such systems is further compounded by Doppler shifts in mobile communication systems. Rapid frequency acquisition and tracking are crucial for accurate decoding of the information being received.
In receivers, frequency lock loops have typically been used to generate a carrier frequency offset estimate, which, in turn, is used to compensate a locally generated carrier frequency. However, in a burst communication system, e.g., a time division multiple-access (TDMA) system, in a fading environment, it may be required to used a so-called open loop offset frequency estimator based on a data burst preamble in order to avoid so-called “hang-up” effects. A frequency offset estimate may be required to be generated at the start of each burst before decoding the information. Rapid acquisition of the carrier frequency and, consequently, rapid generation of the carrier frequency offset, is required in burst communication systems because of the small number of training symbols available in the burst preamble. Additional known frequency acquisition techniques based on phase locked loops tend to have acquisition times longer than the duration of a burst. While there are open loop techniques for fast frequency or phase estimation, these estimation techniques generate estimates having a large variance in the presence of strong multipath spread. Adaptive equalizers, typically used in a multipath environment, are capable of adequately tracking small frequency offsets. However, with a large initial frequency offset the adaptive equalizer is incapable of tracking the frequency offset satisfactorily. Consequently, it is necessary to estimate the frequency offset and perform frequency correction before equalization.
A specific technique that has been proposed to generate frequency offset estimates is the so-called maximum likelihood estimation technique. This technique compensates for the phase changes caused in a received signal because of data modulation and generates an average over a number of symbols to remove the effect of noise. However, the maximum likelihood estimation technique fails in an environment including multipath spread because the received signal, due to data modulation, depends on more than one data symbol and, therefore, cannot be compensated by merely generating the conjugate of the training sequence, namely, x(n)*.
Additionally, it is known that frequency offset estimation in a frequency selective fading channel can be obtained jointly with the channel estimation. However, adaptive equalizer coefficients that are used for canceling intersymbol interference are often obtained directly without generating an explicit channel estimation. Consequently, joint frequency offset estimation and channel estimation for such systems results in additional complexity that is not necessary.
SUMMARY OF THE INVENTION
These and other problems and limitations of prior known frequency offset estimation arrangements and techniques are addressed by generating a frequency offset estimation without explicitly generating a channel estimation for a frequency selective fading communication channel. This is realized by recognizing that, in the absence of additive noise, the channel output at a time n depends only on the last previous predetermined number, L, of data symbols, and that a “state” is a sequence of the last L symbols.
Specifically, in a receiver, a received signal is mixed with a locally generated frequency corresponding to a frequency offset to generate a mixed signal. A calculation is made on the mixed signal in which channel outputs of the same state are combined and accumulated. Then, a summation is made over all possible states of the combined and accumulated channel outputs to yield a so-called metric calculation value for that mixed signal. The metric calculation is then repeated for a plurality of different locally generated frequencies corresponding on a one-to-one basis with a plurality of frequency offsets. The frequency offset corresponding to the largest metric calculation value is selected as the desired frequency offset estimate.
In one embodiment of the invention, a representation of a frequency offset estimation value is obtained by employing an open loop arrangement. More specifically, the frequency offset estimation value is obtained by generating simultaneously a plurality of metric calculation values over a corresponding plurality of predetermined frequency values. This is realized by mixing an input signal with each of the plurality of predetermined frequency values to generate a corresponding plurality of mixed signals and, then, obtaining a separate metric calculation over each the mixed signals. The maximum metric calculation value is selected and the frequency offset estimate is the frequency that corresponds to the selected metric calculation value.
In another embodiment of the invention, a frequency offset estimation value is generated by employing a closed loop arrangement. The received signal is mixed with a frequency offset estimation value and employed in a metric calculation to yield an error signal. A filtered version of the error signal is used to control, in one example, a numerically controlled oscillator to generate the frequency offset estimation value. The metric calculation is made at both a positive step frequency from a frequency offset value and at a negative step frequency from the frequency offset value. The resulting metric calculation values are algebraically subtracted and filtered to yield the error signal.
REFERENCES:
patent: 4520493 (1985-05-01), Heard et al.
patent: 5150380 (1992-09-01), Okanoue
patent: 5255290 (1993-10-01), Anvari
patent: 5422917 (1995-06-01), Scott
patent: 5499268 (1996-03-01), Takahashi
patent: 5537443 (1996-07-01), Yoshino et al.
patent: 5563888 (1996-10-01), Parr et al.
patent: 5640431 (1997-06-01), Bruckert et al.
patent: 6263014 (2001-07-01), Chang et al.
Sabel, A Maximum Likelihood Approach to Symbol Timing Recovery in Digital Communications, University of South Australia, pp. 1-196, 1993.*
Luise et al, Carrier Frequency Recovery in All-digital Modems for Burst-Mode Transmissions, IEEE, pp. 11169-1178, 1995.*
Marco Luise and Ruggero Reggiannini, “Carrier Frequency Recovery in All-Digital Modems for Burst-Mode Transmissions”IEEE Transactions On Communications, vol. 43, No. 2/3/4, Feb., Mar., Apr. 1995, pp. 1169-1.
Michael P. Fitz and William C. Lindsey, “Decision-Directed Burst-Mode Carrier Synchronization Techniques,” IEEE Transactions On Communications, vol. 40, No. 10, Oct. 1992, pp. 1644-1653.
Chin Wellington
Duong Frank
Lucent Technologies - Inc.
Stafford Thomas
LandOfFree
Apparatus and method for generating a frequency offset... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Apparatus and method for generating a frequency offset..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Apparatus and method for generating a frequency offset... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2864933