CN1100863A - Method and device for monitoring and testing non-voice business in telephone metwork - Google Patents
Method and device for monitoring and testing non-voice business in telephone metwork Download PDFInfo
- Publication number
- CN1100863A CN1100863A CN 94107071 CN94107071A CN1100863A CN 1100863 A CN1100863 A CN 1100863A CN 94107071 CN94107071 CN 94107071 CN 94107071 A CN94107071 A CN 94107071A CN 1100863 A CN1100863 A CN 1100863A
- Authority
- CN
- China
- Prior art keywords
- signal
- unit
- voice
- telephone
- monitoring
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Monitoring And Testing Of Exchanges (AREA)
Abstract
Thje equipment is comprised of interface unit of user's line, signal collection unit, signal classification unit and management unit, it uses a nerve network pattern recognition device Adaline with functions of self-learning and self-adaption, and recognizes the non-talking service according to its short time energy of consecutive five frame signal, number of zero crossing, energy zero crossing product, energy zero crossing ratio and twelve order LPC parameters, its monitoring capacity is 512-16384 lines, recurring time for ten thousand voice channel is less than 1 min., illegal signal detection rate is 99.9999%, false alarm rate is less than 0.00001, the recording time for illegal signal is accurate in millisecond order.
Description
The present invention relates to a kind ofly in telephone communication network, non-voice service be carried out the device systems of monitoring management and the technical method of employing thereof, belong to and in telephone call office, be used for monitoring and testing apparatus technical field subscriber's line and that test signal can be provided, and the telephone communication network business statistics is analyzed and the computerization field of managing.
In recent years, high speed development along with Chinese national economy, the online open fax of public telephone, various non-voice services such as fax, compunication have obtained great development, the relevant art file that the competent departments concerned has also made the service management and the technical management of suitable China's actual conditions to non-voice service open on the telephone network.But in concrete execution and supervision and management, run into tangible problem, be that a large amount of " unregistered households " fails to carry out registration formalities, load onto facsimile machine and other modem network access privately and open use, it is serious to cause postal and telecommunications service to run off, influenced the income of relevant department, also the communication quality to whole phone network produces harmful effect.How the non-voice service in the telephone network is carried out the effective supervision management, becomes the problem that solution is badly in need of in China postal and telecommunication entities.Domestic do not have gratifying measure of practical function and equipment so far as yet, and all there are many weak points in all kinds of non-voice service monitoring equipments of exploitation at present.For example the described product of patent application prospectus CN1046425A is its representational product, its shortcoming mainly is 1, design principle is perfect inadequately, whether to surpass 10 seconds signal duration as difference speech and non-voice foundation, will cause a large amount of special voice signals, the erroneous judgement of very noisy signal, for example " phone requesting song " etc.; 2, the terminal decision signal type depends on artificial monitoring, thereby has determined that this plant automation degree is relatively poor, and efficient is lower, a signal is made judgement generally take 20 seconds; 3, be the mode of a monitoring means of every line configuring, it is higher to make that this installs average every line cost, is unsuitable for being applied in the jumbo telephone office; 4, back-stage management and statistical function a little less than, the adaptability of the demand that the user is changed is relatively poor.
The purpose of this invention is to provide a kind of big capacity real-time system of specially non-voice service on the telephone network being monitored and managing, thereby stop up the postal and telecommunications service leakage effectively, guarantee the communication quality of phone network.
Another goal of the invention of the present invention is to design a kind of technical method based on non-voice signal on the identification telephone network of network mode recognition technology, and is both general, reliable, quick and convenient, cheap and practical again.
Above-mentioned purpose adopts following proposal to realize, the non-voice service monitoring-testing device is by the subscriber line interface unit in this telephone network, the signal sampling unit, signal taxon and background management unit are formed, user interface section comprises with 512 lines being the extendible external connection module of unit, capacity reaches 16384 lines, with the observation circuit that the circuit busy-idle condition is monitored, the signal sampling unit contains the telephone path signal sampler, signal power amplifier, the signal frequency overlapped-resistable filter, the signal taxon is provided with the mode discriminator Adallne that distinguishes speech and non-voice signal, the aforementioned identification result is carried out the comprehensive decision device of Adaptive synthesis conclusion and the signal classifier that the signal recognition result is sorted out, and background management unit is one can carry out data management, statistical analysis, the graphical interfaces application program of System self-test and printing function.
The technical method of discerning non-voice signal adopts according to what the non-linear time-frequency domain parameter of signal was adjudicated has a self study, the network mode identifier Adallne of adaptive ability, the short-time energy of continuous five frame signals of foundation, weak point advanced the remainder, energy nil product, energy zero ratio and ten second order LPC parameters and discerned.
Telephone network non-voice service monitoring-testing device of the present invention, the Digital Signal Processing of uses advanced, thereby the auto monitoring and measurement and the management of big capacity online in real time have been realized, for non-voice service charge, network management policy-making provide the reliable technique means, owing to adopt ten thousand working methods of patrolling and examining, thereby make system monitoring broad covered area and average every line cost very cheap.
Fig. 1 is this telephone network non-voice service monitoring-testing device general structure block diagram.
Fig. 2 is that the signal taxon of this non-voice service monitoring-testing device constitutes block diagram.
Fig. 3 is a network mode identifier Adaline schematic diagram in this telephone network non-voice service monitoring-testing device.
Describe equipment of the present invention and technical method in detail below in conjunction with accompanying drawing.
Composition and the information flow direction that shows each processing unit of this equipment shown in Figure 1.1 is subscribers feeder among the figure, the 2nd, the subscriber line interface unit, this element is the external interface part of this equipment, can external 512-16384 line, with 512 lines is the extendible external connection module of unit, its function is to finish each speech channel voice signal reference potential conversion, the busy-idle condition monitoring, and the signal with all conversation speech channels send the subsequent treatment cell processing successively then.The 3rd, signal sampling unit, this unit comprise that dialogue road signal carries out the sampler of 20K high-speed sampling, in routine inspection mode down-sampling 10ms(200 point sample value), signal amplifies by power amplifier then, carries out anti-aliasing filter through filter again.The 4th, the signal taxon comprises and can distinguish the multiple concrete speech and the mode discriminator Adaline of non-voice signal, and the aforementioned identification result is carried out the comprehensive decision device of Adaptive synthesis conclusion and the signal classifier that the signal recognition result is sorted out.This unit adopts the nonlinear analysis scheme based on the artificial neural network, extracts the nonlinear parameter of signal, utilizes the good self-organizing of neural net, self study characteristics, distinguishes speech and non-voice signal.The 5th, background management unit, this unit is the graphical interfaces application program that can carry out data management, statistical analysis, System self-test and printing function, adopts general personal microcomputer, management software Microsoft Vlsualc
++Establishment works in Chinese Windows, under 3.1 graphical interfaces, finishes the court verdict registration, the statistics of every form, analysis and printing and System self-test work.
The formation that shows the signal taxon shown in Figure 2, this unit is all realized by software on microcomputer, wherein 6 is network mode identification module Adaline, the Adaline identifier of a plurality of parallel connections respectively according to the time and frequency domain characteristics of signal to speech, noise, call signal, broadcast receiver, background sound, the pulse dialing sound, double-tone frequently, the beeper call signal, signals such as Modem compunication and fax are automatically tentatively discerned, the 7th, comprehensive judging module, utilize an Adallne that the preliminary recognition result of module 6 outputs is done adaptive comprehensive conclusion, eliminate the contradiction in the preliminary identification, thereby improve non-words signal recall rate, reduce False Rate, the 8th, the neural network learning module, be used for the parameter of the Adallne operating characteristic of decision module 6 and module 7 is done adaptive adjusting, this adjustment process is called the training study process.The 9th, the signal sort module, according to relevant Policy Conditions, the signal recognition result that module 8 is exported is referred to speech and non-voice classification.
The invention allows for and a kind ofly highly effective signal on the telephone network is carried out the technical method of type identification, this method is based on the signal mode recognition technology of neural net.Fig. 3 shows the structural representation of network mode identifier Adaline.The input X of identifier wherein
1, X
2, X
3... X
nBe the time-frequency domain characterisitic parameter of signal, comprise the short-time energy of continuous five frame signals, cross the remainder in short-term, energy nil product, the zero ten second order LPC coefficients when of energy etc., X
0Be outside input, W
0, W
1, W
2... W
nIt is corresponding adjustable parameter, be called weights, adder outputs to threshold value device A after weighted sum is done in input, and the two-value output Y of threshold value device A has represented whether identifier is the judgement of certain signal to signal, and on behalf of "Yes", " 0 ", value " 1 " represent "No".In the learning training process of identifier, it is whether input signal reality is the 0-1 judgement of certain class signal that the tutor responds Y() be fed back to adaptive learning modules B with error between the current Y of output, this module is regulated weights W adaptively according to the learning algorithm-BP algorithm (error backpropagation algorithm) of classics
0, W
1, W
2W
n, so that the error minimize between Y and the Y (referring to Jiao Licheng, " nerve network system theory ", publishing house of Xian Electronics Science and Technology University), thereby make identifier reach the optimum state of specific types of signals being done the classification judgement.Technical method principle of the present invention is perfect, and method advanced person, applicability are strong.
The present invention has made the model machine network access and carried out testing experiment, device systems technical indicator and characteristics: 1, system monitoring capacity 512-16384 line can 512 lines be that the unit arbitrarily disposes; 2, ten thousand speech channel monitoring time≤1 minute; 3, non-words subscriber signal recall rate 99.9999%; 4, false alarm rate＜10-5; 5, the non-words of real time record user is Millisecond any nonrecoverable beginning and ending time.
Claims (3)
1, a kind of telephone network non-voice service monitoring-testing device, it is characterized in that this equipment is by signal sampling unit, subscriber line interface unit (2) (3), signal taxon (4) and background management unit (5) are formed, wherein user interface section comprises with 512 lines being the extendible external connection module of unit, capacity reaches 16384 lines, and the observation circuit that the circuit busy-idle condition is monitored, the signal sampling unit is by the sampler that telephone path signal is sampled, the amplifier that signal power is amplified, the filter that signal is carried out anti-aliasing filter constitutes, the signal taxon comprises can distinguish the multiple concrete speech and the mode discriminator Adaline of non-voice signal, the aforementioned identification result is carried out the comprehensive decision device of Adaptive synthesis conclusion and the signal classifier that the signal recognition result is sorted out, and background management unit is one can carry out data management, statistical analysis, the graphical interfaces application program of System self-test and printing function.
2, a kind of telephone network non-voice service monitoring-testing device is discerned the technical method of non-voice signal, it is characterized in that adopting the network mode identifier Adaline with self study, adaptive ability that adjudicates according to the non-linear time-frequency domain parameter of signal.
3, the technical method of the non-voice signal of identification as claimed in claim 2, it is characterized in that crossing according to the short-time energy of continuous five frame signals, in short-term the remainder, can nil product, can zero than and ten second order LPC parameters discern.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN94107071A CN1036171C (en) | 1994-06-17 | 1994-06-17 | Method and device for monitoring and testing non-voice business in telephone metwork |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN94107071A CN1036171C (en) | 1994-06-17 | 1994-06-17 | Method and device for monitoring and testing non-voice business in telephone metwork |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1100863A true CN1100863A (en) | 1995-03-29 |
| CN1036171C CN1036171C (en) | 1997-10-15 |
Family
ID=5032800
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN94107071A Expired - Fee Related CN1036171C (en) | 1994-06-17 | 1994-06-17 | Method and device for monitoring and testing non-voice business in telephone metwork |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1036171C (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111131627A (en) * | 2019-12-20 | 2020-05-08 | 珠海高凌信息科技股份有限公司 | Method, device and readable medium for detecting personal harmful call based on streaming data atlas |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101667252B (en) * | 2009-10-15 | 2011-07-20 | 哈尔滨工业大学 | Classification and identification method for communication signal modulating mode based on ART2A-DWNN |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1046475A (en) * | 1989-04-22 | 1990-10-31 | 孔庆安 | Multi-angle screw classifying lining board |
-
1994
- 1994-06-17 CN CN94107071A patent/CN1036171C/en not_active Expired - Fee Related
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111131627A (en) * | 2019-12-20 | 2020-05-08 | 珠海高凌信息科技股份有限公司 | Method, device and readable medium for detecting personal harmful call based on streaming data atlas |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1036171C (en) | 1997-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108389592B (en) | Voice quality evaluation method and device | |
| US7570751B2 (en) | System and method for real-time fraud detection within a telecommunication network | |
| Petr et al. | Priority discarding of speech in integrated packet networks | |
| CN105469785A (en) | Voice activity detection method in communication-terminal double-microphone denoising system and apparatus thereof | |
| CN106550155A (en) | Suspicious number is carried out swindling the method and system that sample screens classification and interception | |
| CN104410973B (en) | A kind of fraudulent call recognition methods of playback and system | |
| CN101686444B (en) | System and method for detecting spam SMS sender number in real time | |
| CN111917574A (en) | Social network topology model and construction method thereof, user confidence degree and intimacy degree calculation method and telecommunication fraud intelligent interception system | |
| CN111131627B (en) | Method, device and readable medium for detecting personal harmful call based on streaming data atlas | |
| CN1036171C (en) | Method and device for monitoring and testing non-voice business in telephone metwork | |
| CN106504524A (en) | A kind of method that express highway section is divided based on mobile signaling protocol dynamic | |
| CN108234435A (en) | A kind of automatic testing method based on IP classification | |
| CN101950564A (en) | Remote digital voice acquisition, analysis and identification system | |
| CN116597856B (en) | Voice quality enhancement method based on frogman intercom | |
| CN111554325B (en) | Voice recognition method and system | |
| EP0515277B1 (en) | Modular traffic analyser of digital signals | |
| CN115238867B (en) | Power fault positioning method based on intelligent customer service unstructured data identification | |
| CN116489275A (en) | AI intelligent response system and method based on voice recognition | |
| DE69634902T2 (en) | DEVICE AND METHOD FOR DIGITAL / ANALOGUE MESSAGE TRANSMISSION | |
| CN101355613A (en) | Method and apparatus for distinguishing call type in IP voice network | |
| CN100474871C (en) | Signal transmission channel detection method and calling control system | |
| CN115659217A (en) | Fraud recognition model training method and device, electronic equipment and storage medium | |
| CN104994203B (en) | A kind of speech quality test system and method | |
| Cohen et al. | Traffic analysis for different classes of users of land mobile communication systems | |
| CN114025041B (en) | System and method for rapidly identifying nuisance calls based on non-frequency characteristics of signaling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C06 | Publication | ||
| PB01 | Publication | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |