Image analysis – Pattern recognition – Template matching
Reexamination Certificate
1999-02-23
2002-05-28
Johns, Andrew W. (Department: 2621)
Image analysis
Pattern recognition
Template matching
C375S150000
Reexamination Certificate
active
06396953
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to communication systems. More particularly, the present invention relates to a data pattern correlator or data pattern recognizer for use in a communication system.
BACKGROUND OF THE INVENTION
Data pattern recognizers or correlators are generally utilized in various communication systems to determine if a particular pattern has been received. Once the data pattern is recognized, the correlator provides a time registration or synchronization signal. The synchronization signal is utilized to coordinate the reception of data or information relative to the particular pattern.
Correlators, such as, binary correlators, are utilized for A/J (spread spectrum) message and other data message synchronization. In an exemplary radio system application, a data pattern correlator analyzes the received radio signal to determine if a particular data pattern is present. The data pattern can be comprised of a number of symbols, such as, digital bits, or other indicators, provided in series on a particular radio channel. After the radio system determines that the particular pattern is present, it can then receive further information, such as, data, voice, or other symbols, in accordance with a synchronized communication technique. The data pattern and information can be modulated on the radio signal via any suitable technique, including amplitude modulation, phase modulation, frequency modulation, quadrature amplitude modulation, frequency shift key modulation, or other technique.
Conventional data pattern recognizers have utilized match-filtering techniques which can be susceptible to false pattern recognition due to noise, especially when the data pattern is short. Typically, conventional data pattern recognizers or correlators have increased the size of the data pattern to decrease the occurrence of false pattern recognition. However, increasing the data pattern size presents other problems with pattern recognition performance. Longer data patterns require increased hardware overhead and communication time for recognition. More communication time adversely affects minimum signal dwell times (maximum hop rates) message access times, and message turn-around time periods. In addition, longer correlators become susceptible to missed recognitions due to symbol rate differences between the received signal and the correlator symbol rate clock.
Conventional data pattern recognizers or correlators oversample the incoming signal and compare only one sample of the incoming signal to each symbol in the data pattern (e.g., samples are compared to symbols on a one-to-one basis). With this conventional approach, the only way to improve performance at a constant false recognition rate is to increase the number of characters in the data pattern. As stated above, such a solution has significant disadvantages.
In a conventional system, if all samples of the oversampled incoming signal are compared to each symbol (e.g., multiple samples per symbol or character), any performance gained is reduced by inaccuracies due to channel non-linearities and to doppler shift (difference between the received symbol rate and the correlator's symbol rate). Therefore, comparing all of the samples to each symbol also has significant disadvantages.
Heretofore, current time registration techniques also suffer in deficiencies and accuracy (generally from ±½ character or more), due to software processing time, character uncertainty, and low accuracy correlation phase sorting. Further, several data characters can be lost immediately after the pattern recognition due to the slow response time of current correlator techniques.
Thus, there is a need to improve data pattern recognition performance for a given number of pattern characters and maintain a low constant false match rate due to noise. Further, there is a need to perform data correlation utilizing multiple samples per character. Further still, there is a need for greater synchronization reliability in communication applications.
SUMMARY OF THE INVENTION
The present invention relates to a data pattern recognizer for determining a presence of a data pattern in a signal. The data pattern is comprised of a plurality of symbols. The data pattern recognizer includes a sampler and a comparator. The sampler is configured to oversample the signal and to provide at least two samples (normally 4 or more) per each symbol associated with the data pattern. The comparator is coupled to the sampler and compares one (conventional mode) or more (new modes) samples per symbol, for each of the pattern symbols, to the respective symbols in the data pattern. The comparator indicates the presence of the data pattern if a threshold of comparison samples match respective symbols of the data pattern.
The present invention further relates to a communication system including a signal input, a demodulator coupled to the signal input, a matched filter and symbol slicer, coupled to the demodulator, and a correlator coupled to the slicer. The correlator determines a presence of a data pattern in a signal from the slicer. The data pattern has a number of sequential symbols. The correlator samples the signal at least two times (normally 4 or more) per sequential symbol to provide at least two phases per sequential symbol. The correlator determines the presence of the data pattern by comparing one (conventional mode) or more (new modes) middle phases associated with each symbol to each symbol.
The present invention still further relates to a method for determining a presence of a data pattern in an incoming signal. The data pattern has N sequential symbols. The method includes sampling the incoming signal at least two times (normally 4 or more) per period associated with each of the N sequential symbols to obtain at least two signal samples per symbol and comparing one (conventional mode) or more (new modes) middle samples for each of the N pattern symbols to the corresponding N sequential symbols. The method also includes comparing the total number of matches of the middle samples to the corresponding symbols to a threshold and indicating the presence of the data pattern in the incoming signal in response to the comparison to the threshold.
According to one exemplary embodiment of the present invention, the data correlator takes advantage of typical radio and channel noise being uncorrelated with received signal data patterns. Increasing the number of samples per character that forms the data pattern for matching improves the speed and the accuracy of pattern matching and time registration. A unique phase processing and combining circuit is utilized to compare the number of samples to each character. Preferably, the system utilizes only samples in the more linear/maximum amplitude part of the characters for comparison (e.g., middle samples).
In accordance with another exemplary aspect of the present invention, correlators or digital pattern recognizers can be utilized in almost any high-performance, data-signaling synchronization application, such as, radios, modems, encryption devices, or other communication systems. The symbols can be for binary data, quadrature phase shift key (QPSK), quadrature amplitude modulation (QAM), or any type of data symbol. The symbols can represent data, voice, or other information. The data pattern sequence may be a continuous string of symbols or an interrupted string of symbols. Such an interrupted string of symbols may contain unknown symbols within the known pattern. Such unknown symbols can be masked over so that they contribute no matches or mismatches in the correlation process.
REFERENCES:
patent: 3879581 (1975-04-01), Schlosser et al.
patent: 4009344 (1977-02-01), Flemming
patent: 5862172 (1999-01-01), Sugita et al.
patent: 6002709 (1999-12-01), Hendrickson
Azarian Seyed
Eppele Kyle
Jensen Nathan O.
Johns Andrew W.
Rockwell Collins, Inc.
LandOfFree
Data pattern correlator does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data pattern correlator, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data pattern correlator will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2833590