CN102157148A - DTW (dynamic time warping) voice recognition-based truck examining method - Google Patents
DTW (dynamic time warping) voice recognition-based truck examining method Download PDFInfo
- Publication number
- CN102157148A CN102157148A CN2010106189540A CN201010618954A CN102157148A CN 102157148 A CN102157148 A CN 102157148A CN 2010106189540 A CN2010106189540 A CN 2010106189540A CN 201010618954 A CN201010618954 A CN 201010618954A CN 102157148 A CN102157148 A CN 102157148A
- Authority
- CN
- China
- Prior art keywords
- template
- sound
- steps
- speech recognition
- dtw
- 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.)
- Pending
Links
Images
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
本发明公开了一种基于DTW的语音识别的货车车辆检查方法，包括有训练模板和将语音识别模板用于车辆检查。本发明以DTW语音识别算法为核心，通过训练货车通过轨道发出的正常声音以及有损坏的时候发出的声音形成识别模板，在识别时可以判别货车损坏部件的情况，然后向控制中心发送结果，从而做出早期诊断预报处理，为货车的行驶安全提供保障。本发明具有故障检测识别率高、操作系统简单的优点。
The invention discloses a DTW-based speech recognition inspection method for trucks and vehicles, which includes a training template and uses the speech recognition template for vehicle inspection. The present invention takes the DTW speech recognition algorithm as the core, and forms a recognition template by training the normal sound that the truck passes through the track and the sound when it is damaged. During the recognition, the damaged parts of the truck can be judged, and then the result is sent to the control center, thereby Make early diagnosis and forecast processing to provide guarantee for the driving safety of trucks. The invention has the advantages of high fault detection and recognition rate and simple operating system.
Description
Technical field
The present invention relates to dynamic time programming (DTW) audio recognition method, particularly relate to a kind of truck vehicle inspection method based on the DTW speech recognition.
Background technology
Along with the raising of the lorry speed of a motor vehicle, lorry safety just shows especially more and more important.Vehicle inspection is mainly checked wheel, axletree, spring, brake block, and wheel is most important carrying and operation parts in all parts of lorry, and its quality is the important indicator of lorry safety.The reason of many lorry accidents after all or wheel occurs due to the crackle.The railway operation safety problem is the key factor that the restriction railway develops towards quick, heavily loaded direction.Along with High-Speed Freight Vehicle more and more puts into operation, the traffic safety problem also more and more obtains the attention of railway interests.
The truck vehicle detection method of prior art mainly contains following several:
1. whether traditional truck vehicle detection method mainly is that the workman impacts each parts of wheel with hammer, has judged whether slight crack by sound and working experience, have dislocation; Perhaps the workman patrols and examines the parts such as wheel of each vehicle that arrives at a station.These manual detection methods take time and effort and can't satisfy at present the needs of logistics fast.
2. freight car rolling bearing initial failure rail limit acoustics diagnose system (TADS) is cooperated by Brunswick company and U.S. TTCI company and develops, and is the important component part of rolling stock safety precaution early warning system.TADS gathers the rattle signal of the vehicle bearing of high-speed cruising train by acoustic sensor array is installed on the rail limit, adopt the modern acoustics diagnostic techniques, time-domain signal is carried out energy spectrum, power spectrumanalysis, adopt fuzzy diagnosis and wavelet analysis scheduling theory, set up complicated mathematical model and more and more perfect expert system, according to different bearing fault signal frequencies, energy, the amplitude and the relevant speed of a motor vehicle, the factor of load, determine various bearing fault type and accident defect degree, thereby realize the rolling bearing initial failure is carried out early warning, take precautions against, guarantee traffic safety.It is expensive but this system is very complicated.
Therefore, need badly the simply truck vehicle inspection method by speech recognition of a kind of discrimination height, operating system will be provided.
Summary of the invention
The objective of the invention is to avoid weak point of the prior art and a kind of truck vehicle inspection method based on the DTW speech recognition is provided, have fault detect discrimination height, the simple advantage of operating system.
Purpose of the present invention realizes by following technical measures:
A kind of truck vehicle inspection method based on the DTW speech recognition is provided, includes: training utterance recognition template and described speech recognition template is used for vehicle inspection;
Wherein, described speech recognition template is used for vehicle inspection, includes following steps:
Steps A 1: sound collection, at lorry during through the test zone track, gathering parts such as wheel has the sound of various damages, and gathers the sound of normal wheel through the test zone, forms speech samples;
Steps A 2: pre-service, the speech samples that collects is done pre-service, extract the MFCC parameter then;
Steps A 3: the MFCC parameter that steps A 2 is extracted is as the input of training module, and training forms template, and preserves template and finish training;
Wherein, described training template is used for vehicle inspection, includes following steps:
Step B1: gather the sound that lorry sends through out-of-date conductor rail road by voice acquisition device;
Step B2: the sound that collects by pre-service, is extracted the MFCC parameter then;
Step B3: the MFCC parameter DTW speech recognition algorithm with step B2 extracts, obtain recognition result, judge which kind of damage whether the parts such as wheel of vehicle damage and belong to, and the record check result;
Step B4: the check result of step B3 is sent to truck vehicle Surveillance center, by truck vehicle Surveillance center lorry is handled then.
Beneficial effect of the present invention:
The truck vehicle inspection method of a kind of speech recognition based on DTW of the present invention, include the training template and will train template to be used for vehicle inspection, its principle is to send the purpose that different sound reaches vehicle detection when utilizing the wheel that crack or damage are arranged to contact with the junction of track.The present invention is based on the truck vehicle inspection method of the speech recognition of DTW technology, the difference that is applied to the truck vehicle inspection with the TADS technology is: the TADS technology is according to different bearing fault signal frequencies, energy, amplitude and the relevant speed of a motor vehicle, the factor of load, determine various bearing fault type and accident defect degree, and the present invention be according to normal wheels with damage phonetic feature that wheel sends different phonetic and differentiate the fault of parts such as wheel; Hardware system required for the present invention on the other hand belongs to the simple embedded system, and TADS system complex costliness.Therefore, compare the TADS technology, the present invention has fault detect discrimination height, the simple advantage of operating system.
Description of drawings
The present invention will be further described to utilize accompanying drawing, but the content in the accompanying drawing does not constitute any limitation of the invention.
Fig. 1 is the system and device schematic diagram of embodiment 1 of the truck vehicle inspection method of a kind of speech recognition based on DTW of the present invention.
The structural representation of the speech recognition vehicle inspection system of the embodiment 1 of the truck vehicle inspection method of Fig. 2 a kind of speech recognition based on DTW of the present invention.
Include among Fig. 1 to Fig. 2:
Sound collection module s1, pretreatment module s2, characteristic parameter extraction module s3,
Identification module s5, the s7 of truck vehicle Surveillance center, training module s8, template s9;
Track 1, voice acquisition device 4.
Embodiment
Below in conjunction with accompanying drawing technical scheme of the present invention is described.
DTW is a dynamic time programming, be a kind of pattern match and model training technology early, the thought of its applied dynamic programming has successfully solved the difficult problem that the duration when the phonic signal character argument sequence compares does not wait, and has obtained superperformance in the alone word voice recognition system.Speech recognition based on DTW has isolated word detection discrimination height, and system's characteristic of simple is used very extensive.Based on the advantage of DTW technology, the present invention detects the DTW speech recognition application in truck vehicle.
Use a kind of speech recognition based on DTW of the present invention the truck vehicle inspection method a kind of speech recognition vehicle check system an embodiment as depicted in figs. 1 and 2, include sound collection module s1, pretreatment module s2, characteristic parameter extraction module s3, identification module s5, the s7 of truck vehicle Surveillance center, training module s8 and template s9.
Method of the present invention include the training utterance recognition template and with the speech recognition template applications that trains in the inspection of truck vehicle.
In the present embodiment, the training utterance recognition template is to realize by following steps:
Steps A 1: sound collection, at lorry during through test zone track 1, gathering parts such as wheel has the sound of various damages, and gathers the sound of normal wheel through the test zone, forms speech samples.
Concrete, prepare freight car wheel and some the normal freight car wheel damaged before some.The wheel of the damage of preparing comprises: radial fissure wheel, surface crack wheel, the severe wheel of pit wheel, friction ratio is arranged, free wheel is arranged; The freight car wheel of common parts damages such as axletree, spring, brake block.
Steps A 2: pre-service, the speech samples that collects is done pre-service, extract MFCC parameter (Mel frequency cepstral coefficient) then.
Concrete, allow lorry by the test zone track 1 of voice acquisition device 4 has been installed, all wheels connect the sound that 1 place of connecing sends through track among the acquisition step A1 respectively, and form sample sound, with the sample sound of gathering, extract sound end again by pretreatment module s2.
Steps A 3: the MFCC parameter that steps A 2 is extracted is as the input of training module s8, and training forms template s9, and preserves template s9 and finish training.
Concrete, the efficient voice of the sound end that step s2 is extracted extracts the characteristic parameter of speech samples by characteristic parameter extraction module s3, preserves the characteristic parameter that proposes by training module s8 then, and forms template s9 and finish training process.
In the present embodiment, the speech recognition template applications that trains is realized by following steps in the inspection of truck vehicle:
Step B1: will need the track 1 of lorry that detect, and gather the sound that all wheels of this lorry contact with track 1 by voice acquisition device 4 is installed.
Step B2: all each voice of gathering sound, s2 obtains sound end by pretreatment module, and each voice of the sound end that obtains by characteristic parameter extraction module s3, are extracted the MFCC characteristic parameter of voice.
Step B3: the MFCC characteristic parameter that extracts by identification module s5, is obtained recognition result with the DTW speech recognition algorithm.According to recognition result, can draw wheel and be normally or damage, and can obtain the type that wheel damages.The record check result is parts damages such as wheel damage, axletree, spring, brake block.
Step B4: the result checking, send to the s7 of truck vehicle Surveillance center, allow the s7 of truck vehicle Surveillance center that lorry is processed.
Claims (1)
1. the truck vehicle inspection method based on the DTW speech recognition is characterized in that, includes: training utterance recognition template and described speech recognition template is used for vehicle inspection;
Described training utterance recognition template includes following steps:
Steps A 1: sound collection, at lorry during through the test zone track, gathering parts such as wheel has the sound of various damages, and gathers the sound of normal wheel through the test zone, forms speech samples;
Steps A 2: pre-service, the speech samples that collects is done pre-service, extract the MFCC parameter then;
Steps A 3: the MFCC parameter that steps A 2 is extracted is as the input of training module, and training forms template, and preserves template and finish training;
Described speech recognition template is used for vehicle inspection, includes following steps:
Step B1: gather the sound that lorry sends through out-of-date conductor rail road by voice acquisition device;
Step B2: the sound that collects by pre-service, is extracted the MFCC parameter then;
Step B3: the MFCC parameter DTW speech recognition algorithm with step B2 extracts, obtain recognition result, judge which kind of damage whether the parts such as wheel of vehicle damage and belong to, and the record check result;
Step B4: the check result of step B3 is sent to truck vehicle Surveillance center, by truck vehicle Surveillance center lorry is handled then.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010106189540A CN102157148A (en) | 2010-12-31 | 2010-12-31 | DTW (dynamic time warping) voice recognition-based truck examining method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010106189540A CN102157148A (en) | 2010-12-31 | 2010-12-31 | DTW (dynamic time warping) voice recognition-based truck examining method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102157148A true CN102157148A (en) | 2011-08-17 |
Family
ID=44438616
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010106189540A Pending CN102157148A (en) | 2010-12-31 | 2010-12-31 | DTW (dynamic time warping) voice recognition-based truck examining method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102157148A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102682765A (en) * | 2012-04-27 | 2012-09-19 | 中咨泰克交通工程集团有限公司 | Expressway audio vehicle detection device and method thereof |
| CN103065627A (en) * | 2012-12-17 | 2013-04-24 | 中南大学 | Identification method for horn of special vehicle based on dynamic time warping (DTW) and hidden markov model (HMM) evidence integration |
| CN103364202A (en) * | 2012-04-03 | 2013-10-23 | 财团法人资讯工业策进会 | Vehicle fault diagnosis method and system |
| CN103863188A (en) * | 2014-04-03 | 2014-06-18 | 安徽师范大学 | Vehicle voice recognition signal online self-diagnosis method |
| CN106153335A (en) * | 2015-02-10 | 2016-11-23 | 中国科学院声学研究所 | A kind of train bearing acoustics online system failure diagnosis and method |
| CN110087969A (en) * | 2016-10-19 | 2019-08-02 | 罗伯特·博世有限公司 | Device and method for examining the wheel flat of rail vehicle |
| CN110299151A (en) * | 2019-05-20 | 2019-10-01 | 菜鸟智能物流控股有限公司 | Detection method, detection model generation method and device |
| CN112883078A (en) * | 2021-02-07 | 2021-06-01 | 江西科技学院 | Track dynamic inspection historical data matching method based on DTW and least square estimation |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101063629A (en) * | 2006-04-25 | 2007-10-31 | 汤京乐 | Judging functional failure of electromechanical by speech recognition technology |
| CN101145280A (en) * | 2007-10-31 | 2008-03-19 | 北京航空航天大学 | Vehicle Sound Recognition Method Based on Independent Component Analysis |
-
2010
- 2010-12-31 CN CN2010106189540A patent/CN102157148A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101063629A (en) * | 2006-04-25 | 2007-10-31 | 汤京乐 | Judging functional failure of electromechanical by speech recognition technology |
| CN101145280A (en) * | 2007-10-31 | 2008-03-19 | 北京航空航天大学 | Vehicle Sound Recognition Method Based on Independent Component Analysis |
Non-Patent Citations (3)
| Title |
|---|
| 《中国优秀硕士学位论文全文数据库》 20091113 周宴宇 基于音频信号的滚动轴承智能故障诊断方法研究 3-4 1 , * |
| 《微计算机信息》 20091130 周宴宇 杨胜跃 何正明 龙慧 《基于MFCC与LPCC的滚动轴承故障音频诊断法》 第1.2节 1 , 第31期 * |
| 《微计算机信息》 20091130 周宴宇 杨胜跃 何正明 龙慧 《基于MFCC与LPCC的滚动轴承故障音频诊断法》 第1.2节 1 , 第31期 2 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103364202A (en) * | 2012-04-03 | 2013-10-23 | 财团法人资讯工业策进会 | Vehicle fault diagnosis method and system |
| CN102682765A (en) * | 2012-04-27 | 2012-09-19 | 中咨泰克交通工程集团有限公司 | Expressway audio vehicle detection device and method thereof |
| CN103065627A (en) * | 2012-12-17 | 2013-04-24 | 中南大学 | Identification method for horn of special vehicle based on dynamic time warping (DTW) and hidden markov model (HMM) evidence integration |
| CN103065627B (en) * | 2012-12-17 | 2015-07-29 | 中南大学 | Special purpose vehicle based on DTW and HMM evidence fusion is blown a whistle sound recognition methods |
| CN103863188A (en) * | 2014-04-03 | 2014-06-18 | 安徽师范大学 | Vehicle voice recognition signal online self-diagnosis method |
| CN106153335A (en) * | 2015-02-10 | 2016-11-23 | 中国科学院声学研究所 | A kind of train bearing acoustics online system failure diagnosis and method |
| CN110087969A (en) * | 2016-10-19 | 2019-08-02 | 罗伯特·博世有限公司 | Device and method for examining the wheel flat of rail vehicle |
| CN110299151A (en) * | 2019-05-20 | 2019-10-01 | 菜鸟智能物流控股有限公司 | Detection method, detection model generation method and device |
| CN112883078A (en) * | 2021-02-07 | 2021-06-01 | 江西科技学院 | Track dynamic inspection historical data matching method based on DTW and least square estimation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102157148A (en) | DTW (dynamic time warping) voice recognition-based truck examining method | |
| CN106198062A (en) | The vehicle-mounted acoustic testing system of track vehicle fault and detection method | |
| CN102923164B (en) | High-speed rail health monitoring system based on ultrasonic guide wave and wireless network | |
| CN110658005B (en) | A method for identifying rail corrugation disease based on vehicle body acceleration | |
| CN104833534A (en) | Train running fault diagnosis device based on multi-source information fusion, and method | |
| CN114386467B (en) | Rail vehicle chassis fault abnormal sound monitoring system and method based on acoustic characteristics | |
| CN105527104A (en) | Bearing running-in abnormal sound detection system | |
| CN106218668B (en) | Wheel out of round degree detection method and device | |
| CN104833535A (en) | Railway vehicle tire tread scratch detection method | |
| CN105092467A (en) | Rapid monitoring device and method for high-speed railway wheel tread abrasion | |
| CN116252820B (en) | Polygonal quantitative detection method for high-speed train wheels driven by improved frequency domain integration method | |
| CN107167233B (en) | Method for evaluating wheel set turning of rail transit vehicle | |
| CN108318589A (en) | Fatigue crack length intelligence assessment method based on Gaussian process and monitoring method | |
| CN107403139A (en) | A kind of municipal rail train wheel flat fault detection method | |
| CN105929025B (en) | One kind is based on the continuous wheel tread of the time and space and rail deformation detection method | |
| CN107423692A (en) | A kind of rail corrugation fault detection method based on wavelet-packet energy entropy | |
| Kundu et al. | A review on condition monitoring technologies for railway rolling stock | |
| CN107782548B (en) | Rail vehicle part detection system | |
| CN103106811A (en) | Recognition method of motor vehicle valid traffic conflict based on crashing time of two vehicles | |
| CN105973618A (en) | Preliminary assessment method for security domain of wheel service state | |
| RU2667808C1 (en) | Method of acoustic-emission diagnostics of critical parts of freight-car trucks at operation | |
| CN105741855A (en) | Steward calling and responding detection method based on audio analysis | |
| CN115541001A (en) | Rail transit rail corrugation identification method based on train vibration characteristics | |
| CN105501250A (en) | Fault diagnosis method of scuffed train wheel treads based on vehicle-mounted detection device | |
| CN110579274A (en) | Vehicle chassis fault sound diagnosis method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20110817 |