A specification language speech recognition system and method based on dual ac automata

Abstract

A kind of specification language speech recognition system and method based on double AC automaton, including microphone module and collect the sound of equipment, and send sound data to speech recognition module, speech recognition module generates speech recognition text sequence according to sound data, and send to word element extraction AC automaton, word element extraction AC automaton generates word element number sequence according to speech recognition text sequence and word element list, and send to specification language AC automaton, specification language AC automaton generates specification language sequence according to word element number sequence and specification language list, and send to compliance detection module, compliance detection module detects specification language sequence by compliance list, wherein specification language decomposition module provides the corresponding list generated according to specification language for each module, the present application is accelerated by using double AC automaton, respectively to word element and specification language respectively implement detection, while enhancing method robustness, avoid mood word and other invalid words, realize the normal detection of similar words.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of speech recognition, in particular to a standard language speech recognition system and method based on double AC automata. BACKGROUND

[0002] AC automata is an algorithm for string search, and the background of the algorithm is to solve the problem of finding multiple pattern strings (or called "keywords") in a main text string, which can complete the search in linear time, very efficient;

[0003] The existing AC automata is efficient in solving the multi-pattern string matching problem of string, which can complete the search in linear time, is particularly suitable for processing the matching problem of a large number of keywords, can also match multiple pattern strings at the same time, without matching each pattern separately, greatly improving the efficiency, and is suitable for various scenes requiring multi-pattern matching, such as DNA sequence analysis in bioinformatics, keyword filtering in web crawling, etc., but it is difficult to be used for standard language detection in speech recognition. The standard language recognized by speech recognition contains ineffective words such as mood words, which affects the pattern string matching using AC automata, and the similar words in the standard language recognized by speech recognition make it difficult to construct a pattern string, so it is necessary to propose a standard language speech recognition system and method based on double AC automata. SUMMARY

[0004] The purpose of the present application is to provide a standard language speech recognition system and method based on double AC automata, which can enhance the robustness of the method, avoid ineffective words such as mood words, and realize the normalized detection of similar words, so as to solve the technical defects of the existing technology and the technical requirements that cannot be met.

[0005] To achieve the above purpose, the present application provides the following technical scheme: a standard language speech recognition system based on double AC automata, comprising:

[0006] A microphone module is used to collect the sound of the device and send the sound data to the speech recognition module;

[0007] The speech recognition module is used to receive the sound data sent by the microphone module, extract the speech features of the sound data, generate a speech recognition text sequence according to the speech features, and then send the speech recognition text sequence to the word element extraction AC automata;

[0008] The word element extraction AC automata is used to receive the speech recognition text sequence sent by the speech recognition module and the word element list sent by the standard language decomposition module, match the speech recognition text sequence with the word element list, obtain a complete word element, generate a corresponding word element number sequence, and then send the word element number sequence to the standard language AC automata;

[0009] The specification language AC automaton receives the token number sequence sent by the token extraction AC automaton and the specification language list sent by the specification language decomposition module, matches the token number sequence with the specification language list, matches to a complete specification language, generates a corresponding specification language sequence, and then generates the specification language sequence to the compliance detection module (5);

[0010] The compliance detection module (5) receives the specification language sequence sent by the specification language AC automaton and the compliance list sent by the specification language decomposition module, and detects the specification language sequence through the compliance list;

[0011] The specification language decomposition module generates a token list, a specification language list and a compliance list according to the specification language, and sends them to the token extraction AC automaton, the specification language AC automaton and the compliance detection module (5) respectively.

[0012] A specification language speech recognition method based on double AC automata, comprising:

[0013] 1) The microphone module collects the sound of the device and converts the sound into PCM data, and then sends the PCM data to the speech recognition module;

[0014] 2) The speech recognition module generates a speech recognition text sequence according to the sound data

[0015] 2.1) The speech recognition module receives the PCM data sent by the microphone module;

[0016] 2.2) The speech recognition module performs speech feature extraction on the PCM data;

[0017] 2.3) The speech recognition module generates a speech recognition text sequence according to the speech features through the language model;

[0018] 2.4) The speech recognition module sends the speech recognition text sequence to the token extraction AC automaton;

[0019] 3) The token extraction AC automaton generates a token number sequence according to the speech recognition text sequence and the token list

[0020] 3.1) The token extraction AC automaton receives the speech recognition text sequence sent by the speech recognition module and the token list sent by the specification language decomposition module;

[0021] 3.2) The token extraction AC automaton generates a corresponding token dictionary tree according to the token list, and the token extraction AC automaton constructs a first failure pointer for the token dictionary tree;

[0022] 3.3), the word extraction AC automaton matches the speech recognition text sequence to a complete word in the word dictionary tree and generates a corresponding word number sequence;

[0023] 3.4), the word extraction AC automaton sends the word number sequence to the standard phrase AC automaton;

[0024] 4), the standard phrase AC automaton generates a standard phrase sequence according to the word number sequence and the standard phrase list

[0025] 4.1), the standard phrase AC automaton receives the word number sequence sent by the word extraction AC automaton and the standard phrase list sent by the standard phrase decomposition module;

[0026] 4.2), the standard phrase AC automaton generates a corresponding standard phrase dictionary tree according to the standard phrase list, and constructs a second failure pointer for the standard phrase dictionary tree;

[0027] 4.3), the standard phrase AC automaton matches the word number sequence to a complete standard phrase in the standard phrase dictionary tree and generates a corresponding standard phrase sequence;

[0028] 4.4), the standard phrase AC automaton generates the compliance detection module (5) with the standard phrase sequence

[0029] 5), the compliance detection module (5) receives the standard phrase sequence sent by the standard phrase AC automaton and the compliance list sent by the standard phrase decomposition module, and detects the compliance phrase sequence through the compliance list to determine whether there is a missing or out-of-order phrase;

[0030] 6), the standard phrase decomposition module provides corresponding lists generated according to the standard phrase for each module

[0031] 6.1), the standard phrase decomposition module generates a word list according to the standard phrase and sends the word list to the word extraction AC automaton;

[0032] 6.2), the standard phrase decomposition module generates a standard phrase list according to the standard phrase and sends the standard phrase list to the standard phrase AC automaton;

[0033] 6.3), the standard phrase decomposition module generates a compliance list according to the standard phrase and sends the compliance list to the compliance detection module (5).

[0034] Preferably, the method for generating a speech recognition text sequence in step 2.3) is:

[0035] The speech recognition module maps the speech features to the phonemes of the acoustic model, and uses the known word sequence of the language model to predict the probability of the next word, and finally outputs the most likely speech recognition text sequence.

[0036] Preferably, in the step 3.2), the word list comprises word and word number, the word is composed of one or more words, multiple words can share a word number, each word of the word is a node of the word dictionary tree, the path from the root node of the word dictionary tree to a leaf node of the word dictionary tree represents a word, and the tree node of the word dictionary tree comprises a next level tree node stored by using a balanced binary tree, and the failure pointer corresponds to the longest matching suffix of the tree node of the word dictionary tree.

[0037] The above content in the application can be understood as that the failure pointer of the AC automatic machine points to the longest suffix state of the current state. Note that when the AC automatic machine performs matching, multiple pattern strings can be matched on the same bit.

[0038] Preferably, in the step 3.3), the word number sequence is generated in the following manner:

[0039] The word extraction AC automatic machine matches the speech recognition text sequence in the word dictionary tree. If it is found that there is no match in the matching process, the matched prefix is discarded, the failure pointer is used to jump to the tree node of the longest matching state, and when the word extraction AC automatic machine matches a complete word, the word extraction AC automatic machine generates the word number corresponding to the word as the word number sequence.

[0040] Preferably, in the step 4.2), the standard language list comprises standard language and standard language number, the standard language is composed of multiple word numbers, multiple standard languages can share a standard language number, and each word number of the standard language is a node of the standard language dictionary tree, the path from the root node of the standard language dictionary tree to a leaf node of the standard language dictionary tree represents a standard language, and the failure pointer corresponds to the longest matching suffix of the tree node of the word dictionary tree.

[0041] Preferably, in the step 4.3), the standard language sequence is generated in the following manner:

[0042] The standard language AC automatic machine matches the word number sequence in the standard language dictionary tree. If it is found that there is no match in the matching process, the matched prefix is discarded, the failure pointer is used to jump to the tree node of the longest matching state, and when the standard language AC automatic machine matches a complete standard language, the standard language AC automatic machine generates the standard language number corresponding to the standard language as the standard language sequence.

[0043] Preferably, in the step 5), the compliance list comprises the standard language number detection sequence, and when detecting, the compliance detection module (5) matches the standard language sequence with the standard language number detection sequence to detect whether the standard language is omitted and whether the order is continuous.

[0044] Preferably, in the step 6.1), the standard language decomposition module generates a list of word elements from the keywords of the standard language.

[0045] Preferably, the standard language decomposition module encodes and numbers the standard language to generate a list of standard language, and the standard language decomposition module numbers the standard language to generate a compliance list.

[0046] Compared with the prior art, the present application has the following beneficial effects:

[0047] The present application realizes the detection acceleration of word elements and standard language respectively by adopting double AC automatic machines, enhances the robustness of the method, avoids invalid words such as modal particles, and realizes the normalized detection of similar words. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 The present application has the following beneficial effects:

[0049] In the figure: microphone module 1, speech recognition module 2, word element extraction AC automatic machine 3, standard language AC automatic machine 4, compliance detection module 5, standard language decomposition module 6. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the present application will be described below in conjunction with the accompanying drawings of the embodiments of the present application. Figure 1 The technical solutions in the embodiments of the present application will be described below in conjunction with the accompanying drawings of the embodiments of the present application.

[0051] Please refer to Figure 1 The embodiments of the present application:

[0052] Embodiments:

[0053] As Figure 1 shown: a standard language speech recognition system based on double AC automatic machines, comprising:

[0054] The microphone module 1 is used to collect the sound of the device and send the sound data to the speech recognition module 2.

[0055] The speech recognition module 2 is used to accept the sound data sent by the microphone module 1, perform speech feature extraction on the sound data, generate a speech recognition text sequence according to the speech features, and then send the speech recognition text sequence to the word element extraction AC automatic machine 3.

[0056] Word extraction AC automaton 3, for receiving the speech recognition text sequence sent by the speech recognition module 2 and the word list sent by the standard language decomposition module 6, and matching the speech recognition text sequence with the word list, matching to a complete word, and generating a corresponding word number sequence, and then sending the word number sequence to the standard language AC automaton 4;

[0057] Standard language AC automaton 4, for receiving the word number sequence sent by the word extraction AC automaton 3 and the standard language list sent by the standard language decomposition module 6, and matching the word number sequence with the standard language list, matching to a complete standard language, and generating a corresponding standard language sequence, and then sending the standard language sequence to the compliance detection module 5;

[0058] Compliance detection module 5, for receiving the standard language sequence sent by the standard language AC automaton 4 and the compliance list sent by the standard language decomposition module 6, and detecting the standard language sequence through the compliance list;

[0059] Standard language decomposition module 6, the standard language decomposition module 6 generates a word list, a standard language list and a compliance list according to the standard language, and sends the three to the word extraction AC automaton 3, the standard language AC automaton 4 and the compliance detection module 5 respectively.

[0060] A standard language speech recognition method based on double AC automaton, comprising:

[0061] 1), the microphone module 1 collects the sound of the device, and converts the sound into PCM data, and then sends the PCM data to the speech recognition module 2;

[0062] 2), the speech recognition module 2 generates a speech recognition text sequence according to the sound data

[0063] 2.1), the speech recognition module 2 receives the PCM data sent by the microphone module 1;

[0064] 2.2), the speech recognition module 2 performs speech feature extraction on the PCM data;

[0065] 2.3), the speech recognition module 2 maps the speech features to the phonemes of the acoustic model, and uses the known word sequence of the language model to predict the probability of the next word to constrain the output of the acoustic model, and finally outputs the most possible speech recognition text sequence.

[0066] 2.4), the speech recognition module 2 sends the speech recognition text sequence to the word extraction AC automaton 3;

[0067] 3), the word extraction AC automaton 3 generates a word number sequence according to the speech recognition text sequence and the word list

[0068] 3.1), the word extraction AC automaton 3 receives the speech recognition text sequence sent by the speech recognition module 2 and the word list sent by the standard language decomposition module 6;

[0069] 3.2), the word extraction AC automaton 3 generates the corresponding word dictionary tree according to the word list, and the word extraction AC automaton 3 constructs the first failure pointer for the word dictionary tree;

[0070] In the step 3.2), the word list contains words and word numbers, the word is composed of one or more characters, and multiple words can share a word number (for example, a gas card and a fuel treasure share a word number, and the gas card and the fuel treasure are regarded as the same word for unified processing in the word dictionary tree), each character of the word is taken as a node of the word dictionary tree, and a path from the root node of the word dictionary tree to a leaf node on the word dictionary tree represents a word, the tree node of the word dictionary tree contains the next level tree node stored by using a balanced binary tree, so that the access speed is improved when there are a large number of same prefixes among the words, and the failure pointer corresponds to the longest matching suffix of the tree node of the word dictionary tree.

[0071] 3.3), the word extraction AC automaton 3 matches the speech recognition text sequence to a complete word in the word dictionary tree and generates a corresponding word number sequence;

[0072] In the step 3.3), the word number sequence is generated in the following manner:

[0073] The word extraction AC automaton 3 matches the speech recognition text sequence in the word dictionary tree, and if it is found that there is no match in the matching process, the prefix of the matched part is discarded, the failure pointer is used to jump to the tree node of the longest matching state, the word matching efficiency is improved, and when the word extraction AC automaton 3 matches a complete word, the word extraction AC automaton 3 generates the word number corresponding to the word as the word number sequence.

[0074] 3.4), the word extraction AC automaton 3 sends the word number sequence to the standard language AC automaton 4;

[0075] In the embodiment, the content of the step 3) can process short words by using the AC automaton, avoid the influence of ineffective words such as modal particles, and has high performance.

[0076] 4), the standard language AC automaton 4 generates a standard language sequence according to the word number sequence and the standard language list

[0077] 4.1), the standard language AC automaton 4 receives the word number sequence sent by the word extraction AC automaton 3 and the standard language list sent by the standard language decomposition module 6;

[0078] 4.2), the specification language AC automaton 4 generates a corresponding specification language dictionary tree according to the specification language list, and constructs a second failure pointer for the specification language dictionary tree;

[0079] In the step 4.2), the specification language list contains specification language and specification language number, the specification language is composed of multiple word element numbers, multiple specification languages can share one specification language number, the specification language is composed of multiple word element numbers, multiple specification languages can share one specification language number, which improves the robustness of rule language detection, avoids the influence of mood words and other invalid words, and avoids the influence of similar words on the construction of specification language; multiple specification languages can share specification language numbers, and one step of compliance detection is realized to support different sentence types. And each word element number of the specification language is a node of the specification language dictionary tree, and the path from the root node on the specification language dictionary tree to a leaf node on the specification language dictionary tree represents a specification language, and the failure pointer corresponds to the longest matching suffix of the tree node of the word element dictionary tree.

[0080] 4.3), the specification language AC automaton 4 matches the word element number sequence in the specification language dictionary tree to a complete specification language, and generates a corresponding specification language sequence;

[0081] In the step 4.3), the specification language sequence is generated in the following manner:

[0082] The specification language AC automaton 4 matches the word element number sequence in the specification language dictionary tree, and if a mismatch is found in the matching process, the matched prefix is discarded, and the failure pointer is used to jump to the tree node of the longest matching state, thereby improving the specification language matching efficiency. When the specification language AC automaton 4 matches a complete specification language, the specification language AC automaton 4 generates a specification language sequence corresponding to the specification language number of the specification language.

[0083] 4.4), the specification language AC automaton 4 generates a compliance detection module 5

[0084] 5), the compliance detection module 5 receives the specification language sequence sent by the specification language AC automaton 4 and the compliance list sent by the specification language decomposition module 6, and detects the compliance language sequence through the compliance list to determine whether there is a missing statement and an incorrect sequence;

[0085] In the step 5), the compliance list contains a specification language number detection sequence, and when detecting, the compliance detection module 5 matches the specification language sequence with the specification language number detection sequence to detect whether there is a missing statement and an incorrect sequence, thereby realizing specification language detection.

[0086] 6), the specification language decomposition module 6 provides corresponding lists generated according to the specification language for each module

[0087] 6.1) the specification language decomposition module 6 generates a list of lexical units from the key words of the specification language and sends the list of lexical units to the lexical unit extraction AC automaton 3;

[0088] 6.2) the specification language decomposition module 6 encodes the specification language and generates a list of specification language numbers and sends the list of specification language numbers to the specification language AC automaton 4;

[0089] 6.3) the specification language decomposition module 6 generates a list of compliance from the specification language numbers and sends the list of compliance to the compliance detection module 5.

[0090] The above shows and describes the basic principles and main features of the present application and the advantages thereof. It is apparent to those skilled in the art that the present application is not limited to the details of the above-described exemplary embodiments, but can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be considered in all respects as illustrative and not restrictive, the scope of the present application being defined by the appended claims rather than the above description, and it is intended to embrace all changes and modifications that fall within the meaning and range of equivalents of the claims. Any reference signs in the claims should not be construed as limiting the claims to the figures in which the reference signs are used.

[0091] Furthermore, it should be understood that, although the present specification is described in terms of embodiments, not every embodiment contains only one independent technical solution, and the present specification is described in this way only for the sake of clarity, and those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that those skilled in the art can understand.

Claims

1. A standard speech recognition system based on dual-AC automata, characterized in that, include: The microphone module (1) is used to collect the sound of the device and send the sound data to the speech recognition module (2); The speech recognition module (2) is used to receive the sound data sent by the microphone module (1), extract the speech features from the sound data, generate a speech recognition text sequence based on the speech features, and then send the speech recognition text sequence to the word extraction AC automaton (3). The lexical extraction AC automaton (3) is used to receive the speech recognition text sequence sent by the speech recognition module (2) and the lexical list sent by the standard term decomposition module (6), and match the speech recognition text sequence with the lexical list to obtain a complete lexical, and generate the corresponding lexical number sequence, and then send the lexical number sequence to the standard term AC automaton (4). The standard term AC automaton (4) is used to receive the lexical number sequence sent by the lexical extraction AC automaton (3) and the standard term list sent by the standard term decomposition module (6), and match the lexical number sequence with the standard term list to match a complete standard term and generate the corresponding standard term sequence. Then, the standard term sequence is used to generate the compliance detection module (5). The compliance detection module (5) is used to receive the standard term sequence sent by the standard term AC automaton (4) and the compliance list sent by the standard term decomposition module (6), and to detect the standard term sequence through the compliance list; The standard term decomposition module (6) generates a list of lexical terms, a list of standard terms and a list of compliance terms based on the standard terms, and sends the three to the lexical extraction AC automaton (3), the standard term AC automaton (4) and the compliance detection module (5) respectively.

2. A method for standard usage speech recognition based on dual-AC automata, characterized in that it includes: 1) The microphone module (1) collects the sound of the device, converts the sound into PCM data, and then sends the PCM data to the speech recognition module (2); 2) The speech recognition module (2) generates a speech recognition text sequence based on the sound data. 2.1) The voice recognition module (2) receives PCM data sent from the microphone module (1); 2.2) The speech recognition module (2) extracts speech features from the PCM data; 2.3) The speech recognition module (2) generates a speech recognition text sequence based on speech features using a language model; 2.4) The speech recognition module (2) sends the speech recognition text sequence to the word extraction AC automaton (3); 3) Lexical Extraction AC Automaton (3) Generates a lexical number sequence based on the speech recognition text sequence and the lexical list. 3.1) The lexical extraction AC automaton (3) receives the speech recognition text sequence sent by the speech recognition module (2) and the lexical list sent by the standard term decomposition module (6); 3.2) The lexical extraction AC automaton (3) generates the corresponding lexical dictionary tree based on the lexical list, and the lexical extraction AC automaton (3) constructs invalid pointers for the lexical dictionary tree; 3.3) The AC automaton (3) extracts the lexical character and matches a complete lexical character in the lexical dictionary of the speech recognition text sequence and generates the corresponding lexical character number sequence; 3.4) The lexical extraction AC automaton (3) sends the lexical number sequence to the standard usage AC automaton (4); 4) Standardized Terms AC Automaton (4) generates a standardized terms sequence based on the lexical number sequence and the standardized terms list. 4.1) The standard term AC automaton (4) receives the lexical extraction AC automaton (3) and the standard term decomposition module (6) send the list of standard terms; 4.2) The standard term AC automaton (4) generates the corresponding standard term dictionary tree based on the standard term list and constructs an invalidation pointer for the standard term dictionary tree; 4.3) The standard term AC automaton (4) matches the word number sequence with a complete standard term in the standard term dictionary and generates the corresponding standard term sequence; 4.4) The standard terminology AC automaton (4) generates a compliance detection module (5) from the standard terminology sequence. 5) The compliance detection module (5) receives the sequence of standard terms sent by the standard term AC automaton (4) and the compliance list sent by the standard term decomposition module (6), and checks the sequence of standard terms through the compliance list to see if there are any omissions or inconsistencies in the order. 6) The standard terminology decomposition module (6) provides each module with a corresponding list generated based on the standard terminology. 6.1) The standard term decomposition module (6) generates a list of lexical elements based on the standard terms and sends the list of lexical elements to the lexical extraction AC automaton (3); 6.2) The standard term decomposition module (6) generates a list of standard terms based on the standard terms and sends the list of standard terms to the standard term AC automaton (4); 6.3) Standard term decomposition module (6) generates a compliance list based on the standard term and sends the compliance list to the compliance detection module (5).

3. The standard speech recognition method based on dual-AC automata according to claim 2, characterized in that, The method for generating the speech recognition text sequence in step 2.3) is as follows: The speech recognition module (2) maps speech features to phonemes of the acoustic model and uses the known word sequence of the language model to predict the probability of the next word, and finally outputs the most likely speech recognition text sequence.

4. The standard speech recognition method based on dual-AC automata according to claim 2, characterized in that, In step 3.2), the lexical list includes lexicals and lexical numbers. A lexical consists of one or more characters. Multiple lexicals can share a single lexical number. Each character of a lexical is a node in the lexical dictionary. The path from the root node of the lexical dictionary to a leaf node in the lexical dictionary represents a lexical. The tree nodes of the lexical dictionary contain the next-level tree nodes and are stored using a balanced binary tree. The invalid pointer corresponds to the longest matching suffix of the tree node in the lexical dictionary.

5. The standard speech recognition method based on dual-AC automata according to claim 4, characterized in that, In step 3.3), the method for generating the lexical number sequence is as follows: The lexical extraction AC automaton (3) matches the speech recognition text sequence in the lexical dictionary. If a mismatch is found during the matching process, the matched prefix is ​​discarded and the invalid pointer is used to jump to the tree node of the longest matching state. When the lexical extraction AC automaton (3) matches a complete lexical, the lexical extraction AC automaton (3) generates a lexical number sequence corresponding to the lexical number.

6. The standard speech recognition method based on dual-AC automata according to claim 2, characterized in that, In step 4.2), the list of standardized terms includes standardized terms and standardized term numbers. Each standardized term consists of multiple lexical numbers, and multiple standardized terms can share a single standardized term number. Each lexical number of a standardized term serves as a node in the standardized term dictionary tree. The path from the root node of the standardized term dictionary tree to a leaf node in the standardized term dictionary tree represents a standardized term. The invalidation pointer corresponds to the longest matching suffix of the tree node in the lexical dictionary tree.

7. The standard speech recognition method based on dual-AC automata according to claim 6, characterized in that, In step 4.3), the method for generating the standardized terminology sequence is as follows: The standard term AC automaton (4) matches the word number sequence in the standard term dictionary tree. If a mismatch is found during the matching process, the matched prefix is ​​discarded and the invalid pointer is used to jump to the tree node of the longest matching state. When the standard term AC automaton (4) matches a complete standard term, the standard term AC automaton (4) generates a standard term sequence based on the standard term number corresponding to the standard term.

8. The standard speech recognition method based on dual-AC automata according to claim 2, characterized in that, In step 5), the compliance list includes a sequence of standard terms numbering for detection. During detection, the compliance detection module (5) matches the sequence of standard terms with the sequence of standard terms numbering to detect whether any terms are omitted or not in sequence.

9. The standard speech recognition method based on dual-AC automata according to claim 2, characterized in that, In step 6.1), the standard term decomposition module (6) generates a list of lexical terms from the keywords of the standard term.

10. A method for standard speech recognition based on a dual-AC automaton according to claim 2, characterized in that, The standard term decomposition module (6) encodes and numbers the standard terms to generate a standard term list, and the standard term decomposition module (6) generates a compliance list by numbering the standard terms.

Images