Pulse or digital communications – Receivers – Particular pulse demodulator or detector
Reexamination Certificate
1999-03-09
2002-07-30
Pham, Chi (Department: 2731)
Pulse or digital communications
Receivers
Particular pulse demodulator or detector
C375S261000, C375S320000
Reexamination Certificate
active
06426981
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to radio communication, and, more particularly, to the decoding of a radio frequency transmission channel conveying, by quadrature modulation of at least one carrier, coded digital information.
BACKGROUND OF THE INVENTION
Terrestrial digital broadcasting, such as defined in the European DVB-T (Digital Video Broadcasting-terrestrial) specification, is based on the MPEG-2 compression standard. In addition, mobile or portable terrestrial transmission may use digital quadrature modulation on a large number of “orthogonal” carriers (OFDM modulation: “orthogonal Frequency Division Multiplexing”). After a first coding phase, commonly referred to as source coding by those skilled in the art, the digital signals to be transmitted are transformed into a succession of packets of a certain number of bytes. The number of bytes is typically 188 bytes in the DVB-T specification based on the MPEG-2 compression standard which are intended to be transmitted by a radio frequency pathway or channel to a receiver. However, radio frequency transmission channels are generally corrupted with errors because of disturbances which effect the useful signal transmitted, such as, for example noise, interference, echo.
It is therefore necessary, before performing the modulation of the carrier or carriers by the signal, to perform a specific coding processing, commonly referred to as channel coding. This allows the detection and correction in the receiver of most of the errors caused by the radio frequency transmission channel.
This channel coding generally includes re-introducing a calculated redundancy into the signal, generally by using algebraic and/or convolutional error codes. In transmissions which meet the DVB-T specification, so-called Reed-Solomon external coding is used. This may be followed, in particular, by an internal coding or convolutional coding which is better known as Viterbi coding by those skilled in the art.
On exiting the channel coding phase a stream of digital data (symbols) is available. These symbols will be used to modulate one or more (in the case of OFDM modulation) carriers with a view to transmission to the receiver.
Instead of using simple digital modulation, modulating the carrier directly either in terms of amplitude or in terms of frequency by the serial bit train representing the information transported, so-called quadrature modulation is generally used which is more effective under given conditions of signal
oise ratio. Thus, for quadrature modulation the input symbols, coded over n bits, are converted into two signals I and Q each coded over n/2 bits, thus yielding 2
n
/2 states for each of the signals I and Q. The signals I and Q then quadrature-modulate one or more carriers. This results in a constellation of points in the space (I, Q), these corresponding to the various values which the signals I and Q may take.
OFDM modulation includes quadrature-modulating, not a single carrier, but a number N of carriers, typically 2048 or 8192, respectively 2K or 8K depending on the application, by symbols of duration T
s
. The frequency difference between two consecutive carriers is 1/T
s
. OFDM modulation has the advantage of exhibiting very good behavior in the case of multipath reception. Multipath effects are frequent during terrestrial reception, fixed or mobile, if the delay of the multiple paths remains less than the time separating two OFDM symbols (the guard interval).
On receiving the radio signal, after UHF demodulation and digitization, a demodulation processing is performed. The demodulation comprises, in the case of OFDM modulation, an N-point fast Fourier transform processing so as to extract each carrier, as well as a determination of phase in a phase-locked loop, to retrieve digital signals representative of the transmitted signals coded as I and Q. These digital signals will then undergo a first decoding phase, termed channel decoding, followed by a second decoding phase termed source decoding, so as to obtain the initial digital information at the end of the chain.
One of the steps of the channel decoding includes, for each signal pair I and Q received, determining the digital words transmitted (2 bits for quadrature modulation of the QPSK type, 4 bits for quadrature modulation of the 16QAM type, 6 bits for quadrature modulation of the 64QAM type) and consequently the code word associated with the corresponding point of the constellation and representative of the corresponding coded information received. This determination is also accompanied by the determination of a confidence word, or more simply “confidence criterion” or “soft decision”, assigned to the code word. The code word thus determined, assigned by its confidence criterion, will then be used in the remainder of the channel decoding, especially in a convolutional Viterbi decoder.
The confidence criteria associated with the code words makes it possible, for each code word, to afford a greater or lesser guarantee as to the accuracy of the value determined for this code word. Currently, the various confidence criteria are stored in a memory table and chosen as a function of the incoming pair I,Q. The table is precalculated on the basis of a unique law whose threshold values are predetermined.
Moreover, the various stored values of the confidence criteria have been preestablished for a certain type of transmission channel exhibiting predefined characteristics in terms in particular of noise or echo. This leads, upon a modification of the characteristics of the transmission channel, such as, in mobile television applications, to unsuitable choices of the “code word/confidence criterion” pair, and consequently, to an increase in the rate of erroneous bits on completion of the channel decoding.
SUMMARY OF THE INVENTION
In view of the foregoing background, the present invention seeks to reduce the rate of erroneous bits during the decoding of a radio frequency reception of channels, using OFDM modulation in particular.
It has been observed, through measurements, that the decoding step which maps each carrier to the code word of the constellation is sensitive to the type of route traversed by the transmission signal. The decoding process according to the invention makes it possible to adapt this step of the decoding to the type of reception channel.
The invention therefore proposes a process for decoding a radio frequency transmission channel conveying, by quadrature modulation of at least one carrier, coded digital information. In the process a succession of digital input blocks is received, each comprising two digital input words (I and Q) representative of the value of the coded information, and a code word (formed of two digital words hereinafter denoted XI and XQ of j bits each, with j=1 for QPSK, j=2 for 16QAM and j=3 for 64QAM). The confidence is determined from each pair of input words and from a mapping law.
According to a general characteristic of the invention, at least one mapping rule or law is derived which can be parametrized by at least one parameter. This makes it possible, for each value of the parameter, to determine the corresponding confidence word for each of the two digital input words. The mapping law can make it possible to derive the confidence words directly. However, when the input blocks comprise, apart from the words I and Q, an initial confidence word emanating from the demodulator, it is particularly advantageous to multiply this initial confidence criterion by the confidence criterion derived by the law so as to obtain the digital confidence word associated with the code word.
The value of the parameter is varied and for each current value of the parameter and for a predetermined number of bits received, a rate of erroneous bits is determined corresponding to this current value. Those values having given rise to a minimum rate of erroneous bits are selected from among all the current values of the parameter, and the selected current value is assigned to the parameter for the remainder of the channel decoding
Allen Dyer Doppelt Milbrath & Gilchrist, P.A.
Pham Chi
STMicroelectronics S.A.
Tran Khai
LandOfFree
Process and device for decoding a radio frequency... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Process and device for decoding a radio frequency..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Process and device for decoding a radio frequency... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2868501