Pulse or digital communications – Testing – Data rate
Reexamination Certificate
1999-06-29
2002-10-08
Le, Amanda T. (Department: 2734)
Pulse or digital communications
Testing
Data rate
C375S341000, C370S252000
Reexamination Certificate
active
06463097
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to methods of rate detection at a receiving end of a digital communications system, such as a code division multiple access (CDMA) system, in which system the information data rate is variably selected at the transmitting end from an applicable rate set including a full rate and lower rates, each lower rate being the full rate divided by a different integer, and data is repeated for the lower rates to maintain a constant apparent data transmission rate. In its particular aspects, the present invention relates to a rate detection method in which a rate determination or classification decision process uses measurements of repetition characteristics of data which has not been de-repeated.
2. Description of the Related Art
Such a rate detection method is generally known from Edith Cohen and Hui-Ling Lou, “Multi-Rate Detection for the IS-95 CDMA Forward Traffic Channels, IEEE Global Telecommunications Conference 1995.
“In 1992, a direct sequence code division multiple access (DS-CDMA) system was adopted as Interim Standard 95 (IS-95) by the Telecommunications Industry Association (TIA) for deployment in the cellular band at 800 MHz. After successful field tests and trial systems, the IS-95 system is now operating with tens of millions of subscribers.”
CDMA is based on spread spectrum technology originally developed by the Allies during World War II to resist enemy radio jamming. Spread spectrum signals are characterized by a bandwidth W occupied by signals in a channel much greater than the information rate R of the signals in bit/s. Thus, a spread spectrum signal inherently contains a kind of redundancy which can be exploited for overcoming several kinds of interference (including signals from other users in the same band and self-interference in the sense of delayed multipath components) introduced by the channel. Another key property of spread spectrum signals is pseudo-randomness. Therefore, the signal appears to be similar to random noise, making it difficult to demodulate by receivers other than the intended ones. In CDMA systems, users share a common channel bandwidth and users are distinguished by different code sequences. In the case of IS-95 each communication with a user is modulated or scrambled by long and short Pseudo Noise (PN) sequences and also modulated by a specific one of a set of orthogonal sequences, known as Walsh codes, which is assigned to the user. The latter modulation is known as applying a Walsh cover. Thus, a particular receiver can recover a certain transmitted signal by applying the PN sequences, and also the Walsh sequence used by the corresponding transmitter for the particular receiver.
In the IS-95 DS-CDMA system variable information data rates are used according to the voice activity detected by the voice encoder. This enables a reduction in transmitted power at the lower rates leading to a reduced average transmitted power per user and consequent increase in capacity of the system. Two sets of information data rates (Rate Sets
1
and
2
) can be encoded, depending on the implemented voice encoder each set comprising full rate, and lower rates of half rate, quarter rate, and eighth rate. For the lower rates, symbols are repeated to achieve the same apparent symbol transmission rate as when full rate is used. In Rate Set
2
, there are 50% more symbols in a frame than in Rate Set
1
, but prior to transmission one third of the Rate Set
2
symbols are punctured so that in both rate sets the same number of symbols in a frame are transmitted. The information data rate can change from frame to frame, but information indicating the currently used data rate is not transmitted along with the speech data. Therefore, the receiver has to detect the data rate by hypothesis testing. The algorithm implemented by rate classification or decision logic which determines which of the possible information data rates is utilized for the current frame is called a Rate Detection Algorithm (RDA).
In accordance with the known rate detection method utilizing repetition characteristics of data, prior to any de-repetition, measures are formed for Rate Set
1
which determine how well symbols match within successive groups of 2, 4, and 8 symbols. Such method, while requiring only a relatively small amount of computational resources, is not of sufficient reliability for an IS-95 DS-CDMA system that a rate decision could be based solely thereon. Further, such known method yields even poorer results when applied to Rate Set
2
, because it does not take into account the effects of puncturing.
Other information which could be used for rate detection includes CRC checking results (which in accordance with IS-95 are available for all data rates except quarter and eighth rates in Rate Set
1
), Viterbi decoder survivor metrics, and correlations between re-encoded data and data entering the decoder for each possible data rate. The latter two methods, which utilize data available from or after the Viterbi decoding at each of the possible data rates, and for each data frame, are inherently more reliable, but are computationally intensive. The method employing such correlations makes particularly intensive use of computational resources since, each frame, for each of the possible data rates the data must not only be Viterbi decoded (after de-puncturing and de-repeating as required), but also convolutional re-encoded in order to form correlations between the re-encoded data and the data entering the Viterbi decoding for each possible data rate.
Intensive use of computational resources is undesirable, particularly in wireless handsets, because battery life is generally reduced as the number of instructions per section required to be executed by a digital signal processor (DSP) within the handset increases. A significant power savings is gained when the DSP has a relatively high proportion of slack time, during which the DSP can go into an idle mode.
OBJECTS AND SUMMARY OF THE INVENTION
It is an object of the present invention to provide an improved method of rate detection at a receiving end of a digital communications system which is highly reliable but on average utilizes a relatively small amount of computational resources. It is a further object that the rate detection method take into account any pattern of puncturing used for an applicable rate set.
These and other objects of the present invention are satisfied by such a method of rate detection wherein the data rate of received convolutional encoded data is first determined by a coarse decision method which is computationally simple because it is based on measures of the data computed before any Viterbi decoding (“pre-decoding” measures). Then, using data available from or after the Viterbi decoding at the first determined data rate to obtain or form a “post-decoding” measure, an evaluation is made whether the first determined data rate should be selected as the actual data rate, and preferably, also whether there is high or low confidence in this selection. Only when, the evaluation does not result in the selection of the first determined data rate as the actual data rate, is a more accurate but computationally intensive fine decision method resorted to, using post decoding measures obtained or calculated at one or more other data rates.
In accordance with the invention, pre-decoding measures of repetition patterns in the data are calculated for each of the possible lower data rates, and the coarse rate decision is made using these measures and a first set of thresholds. In the calculation of the pre-decoding measures for Rate Set
2
, it is taken into account that the data has been punctured according to a predetermined pattern.
In order to evaluate the result of the coarse decision method, the convolutional encoded data is de-punctured and de-repeated where required, Viterbi decoded, and convolutional re-encoded, all with respect to the first determined data rate. Then there is first formed a correlation between the received convolutional en
Chen Albert
Held Ingolf
Koninklijke Philips Electronics , N.V.
Le Amanda T.
Slobod Jack D.
LandOfFree
Rate detection in direct sequence code division multiple... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Rate detection in direct sequence code division multiple..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Rate detection in direct sequence code division multiple... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2997301