Coded data generation or conversion – Analog to or from digital conversion – Nonlinear
Reexamination Certificate
1999-11-10
2001-05-29
Williams, Howard L. (Department: 2819)
Coded data generation or conversion
Analog to or from digital conversion
Nonlinear
C341S155000, C455S138000, C455S308000
Reexamination Certificate
active
06239730
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a method and a device for maximizing the ratio between signal and quantization noise when converting a multi-carrier signal from analogue to digital form with a limited number of bits within a particular quantization range. The invention can also be used for maximizing the ratio between signal and quantization noise when converting a multi-carrier signal from digital to analogue form.
CURRENT TECHNOLOGY
A technique which is often used in modern digital telecommunication systems is so-called multi-carrier systems, in other words a system where a number of carrier waves at different frequencies are used to transmit information within one and the same interval of time. An advantage of multi-carrier systems is that as the same information is transmitted at different frequencies, the effect of multi-wave propagation is counteracted, which is particularly desirable especially in very built-up areas, for example in inner-cities.
For transmission in a multi-carrier system the signal has to be converted from digital to analogue form, which is carried out in a so-called digital-to-analogue converter, usually called an ADC (Analogue to Digital Converter). In an ADC a certain amount of so-called quantization noise always arises, which is caused by the ADC only having access to a limited number of bits. Multi-carrier systems utilize signals with relatively large dynamics, which means that in a digital signal a relatively large number of bits is required in order to obtain the required resolution. This in turn means that the quantization noise can become particularly troublesome in a multi-carrier system.
The distance between the bits in an ADC is called the quantization interval. The bits to which the ADC has access define in addition a so-called quantization range which must be correctly placed in relation to the signal received in order to minimize the quantization noise.
A common method of minimizing the quantization noise is to assume that the amplitude of the received signal is equally divided between all the quantization intervals, and that each quantization interval thus contributes equally towards the quantization noise. According to this assumption, the noise decreases monotonically with the quantization range. As this method does not take into account the error contribution from the part of the signal which ends up outside the quantization interval, the method is not able to minimize the noise correctly. Nor can the method be used to minimize the noise as a function of the quantization range.
SUMMARY OF THE INVENTION
The problem which is solved by the present invention is thus to minimize the quantization noise which arises in conversion between analogue and digital form of a multi-carrier signal. The signal comprises an AC component which has a certain RMS value, and the conversion can be carried out with a particular limited number of bits within a quantization range.
The signal may possibly also comprise a DC component, which can be wanted or unwanted. A wanted DC component means that the DC level is used to transmit information, an unwanted DC component means that the level is only to be regarded as interference which should be removed.
According to the invention, the quantization noise is minimized by creating an ideal ratio between the quantization range and the RMS value of the AC component (RMS=Root Mean Square).
Unwanted DC components in the signal are removed, and if the signal comprises a wanted DC component the signal is centred by the wanted DC component being placed in the centre of the quantization range. Before the conversion takes place, a scaling factor is calculated, which is used to create an ideal ratio between the RMS value of the AC component and the quantization range of the ADC. Thereafter, the AC component is multiplied by the scaling factor, which means that the required ideal ratio between the quantization range and the RMS value of the AC component is obtained.
In a preferred embodiment, the scaling factor is calculated as a factor which minimizes the ratio between the quantization noise and the AC component, taking into account the number of bits which can be used for the conversion.
In an alternative embodiment of the invention, the scaling factor is calculated as a factor which minimizes the expected value of the quantization noise for the type of distribution of the AC component taking into account the number of bits which can be used for the conversion. In this embodiment, the type of distribution of the AC component is also determined.
If the type of distribution of the signal is Gaussian, the logarithmic value of the ideal ratio between the quantization range and the RMS value of the AC component can be expressed as 20*log
10
(V
QR
/V
RMS
)=6.26+0.71*b, where V
QR
is the accessible quantization range, V
RMS
is the RMS value of the AC component, and b is the accessible number of bits for the conversion.
The invention can be used to minimize the quantization noise both when converting a multi-carrier signal from digital form to analogue and when converting a multi-carrier signal from analogue to digital form. The invention is mainly intended to the used in an OFDM system (Orthogonal Frequency Division Multiplex), but can in principle be used in all multi-carrier systems.
REFERENCES:
patent: 5546081 (1996-08-01), Baumgartl
patent: 5557638 (1996-09-01), Fisher et al.
patent: 5687285 (1997-11-01), Katayanagi et al.
patent: 5841385 (1998-11-01), Xie
patent: 6002352 (1999-12-01), El-Ghoroury et al.
patent: 58-29212 (1983-02-01), None
patent: 2-13112 (1990-01-01), None
EP0420545 A (General Electric Co) 1991-04-03 (abstract) World PatentsIndex (online), London, UK, Derwent Publications, Ltd. (retrieved on 1999-08-18), retrieved from EPO WPI Database DW9114.
Nixon & Vanderhye P.C.
Telefonaktiebolaget LM Ericsson (publ)
Williams Howard L.
LandOfFree
Method and device for maximizing the ratio between signal... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and device for maximizing the ratio between signal..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and device for maximizing the ratio between signal... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2458967