CN103247175B - Road congestion monitoring method based on idling sound frequency spectrums of automobiles - Google Patents
Road congestion monitoring method based on idling sound frequency spectrums of automobiles Download PDFInfo
- Publication number
- CN103247175B CN103247175B CN201310153034.XA CN201310153034A CN103247175B CN 103247175 B CN103247175 B CN 103247175B CN 201310153034 A CN201310153034 A CN 201310153034A CN 103247175 B CN103247175 B CN 103247175B
- Authority
- CN
- China
- Prior art keywords
- road
- judged
- energy
- spectrum
- circuit
- 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.)
- Expired - Fee Related
Links
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention discloses a road congestion monitoring method based on idling sound frequency spectrums of automobiles, which is characterized in that the method is based on the principle that idling sound of automotive vehicles accounts for a higher proportion in the environment in case of a traffic jam. FFT conversion for road sound collected within a certain time is performed, the frequency domain energy spectrums within the low frequency range (20 Hz-40 Hz) under the states of traffic jam and unimpeded traffic are obviously distinguished, and an obvious peak exists at the idling frequency in case of the traffic jam; and the steeping degree of the peak is converted into a factor k through an algorithm, and a road condition judgment standard corresponding to the k value is built. Experiments show that the road congestion can be effectively detected through the calculation of the k value.
Description
Technical field
The present invention relates to a kind of monitor congestion in road method and corresponding monitoring device.
Background technology
Judge that traffic congestion has a lot of method, ground induction coil detection method, camera detection method, microwave monitoring method about how to monitor both at home and abroad.Be mainly shooting and monitor differentiation in real time, these method operating costs and equipment Requirement high, uneconomical at actual motion.Though also have some additive methods to be in laboratory stage or technology maturation in addition, be subject to expense in the application, run the restriction of the factors such as interference, be not therefore widely applied.
Summary of the invention
The object of this invention is to provide a kind of congestion in road monitoring method based on auto idle speed phonetic analysis and device, the method uses the principle of spectrum analysis, processes, and then realize the judgement of road congestion conditions to the distinctive frequency range of auto idle speed sound.
For reaching above object, the present invention takes following technical scheme to be achieved:
Based on a congestion in road monitoring method for auto idle speed sound spectrum, it is characterized in that, comprise the steps:
(1) the idling sound that driving vehicle on road sends is gathered, and be converted into electric signal, sampling time interval is 10 seconds, by audio amplifier circuit, these electric signal are amplified again, after low-pass filter circuit by frequency control contained by electric signal at below 100HZ, export electric signal to dsp processor finally by level lifting circuit;
(2) dsp processor by containing idling energy spectrum number of signals mode convertion after, in sampling time interval, within every 2 seconds, carry out a Fourier transform, and the energy spectrum superposition after FFT is converted, form the characteristic energy spectrum that accumulative resolution reaches below 0.6Hz, find peak value in the scope of the 20Hz-50Hz composed in this characteristic energy, and the energy of each 0.2Hz scope of this peak value and left and right, peak value place and the accumulative characteristic energy in 10 seconds are composed be business, draw coefficient k of blocking up;
(3) as k<10, be judged as that vehicle is in current X state; Be judged as that as k > 10 vehicle is in the T state of stagnation;
(4) within every five minutes, once add up, use N
xrepresent the number of X state in this period, N
trepresent the number of T state in this period, when:
A, N
t< N
x* 0.3; Then be judged as the coast is clear;
B, N
x* 0.3≤N
t< N
x* 0.7 is judged as that road is busy;
C, N
x* 0.7≤N
t; Then be judged as congestion in road.
In said method, described sampling carries out with 572Hz sampling rate, and sampled point is 1024.After signal is carried out Fourier transform, energy compensating is carried out to the result of FFT, namely add up peak point in characteristic energy spectrum and the ratio of the second peak point according to gained, determine corresponding penalty coefficient.
The present invention compares existing traffic congestion monitoring method, and its advantage is, by judging the jam situation of road to the analysis of auto idle speed sound spectrum, it is skillfully constructed, apply convenience, operating cost is cheap.After experimental products networking, by the analysis of coefficient k value of blocking up to each measurement point, simply can judge road congestion conditions, for city manager's heart understanding road conditions in the controlling, making traffic control in time has important reference role.In addition, monitoring device of the present invention can adopt cheap, the simple chip of interface, DSP and custom circuit manufacture, and the collection as sound can adopt microphone circuit.In the sample and transform of digital signal, DSP shows its distinctive superiority, speed is fast, cost is low, low in energy consumption, high performance processing power, there is powerful external communication interface (SCI, SPI, CAN) be convenient to form large control system, the result wanted can be obtained fast with its Treatment Analysis frequency spectrum data; Little, lightweight, the portable and environmental protection of whole device volume.
Accompanying drawing explanation
Below in conjunction with the drawings and the specific embodiments, the present invention is described in further detail.
Fig. 1 is the overall major part circuit theory diagrams of apparatus of the present invention.
Fig. 2 is that the microphone of apparatus of the present invention enters to hold audio amplifier circuit figure.
Fig. 3 is the low-pass filtering amplification circuit diagram of apparatus of the present invention.
Fig. 4 is the level lifting circuit of apparatus of the present invention.
Fig. 5 is the schematic flow sheet of the inventive method.
Fig. 6 is the energy spectrum of " X " state.
Fig. 7 is the energy spectrum (measurement point is based on dolly) of " T " state.
Fig. 8 is the energy spectrum (measurement point is based on cart) of " T " state.
Embodiment
Based on a congestion in road monitoring device for auto idle speed sound spectrum, comprise front-end circuit and dsp processor.
Integrated circuit is see Fig. 1, and integrated circuit figure major part is that the minimum system of DSP2812 engineering comprises: power circuit, adc circuit, and clock circuit and idling sound transducer comprise: sound collection amplifying circuit, low-pass filter, level lifting circuit.
Microphone adopts floating ground mode to access SSM2019 input end 2-3, and SSM2019 output terminal pin 6 is by 33K resistance access ADOP07 resistance capacitance second-order filter circuit, and ADOP07 output terminal 6 connects AD8607 pin 3, AD8607 output terminal 1 and meets DSP ADCINA1 and hold.
The function of front-end circuit mainly gathers the sound that driving vehicle on road sends, and the signal collected is carried out preliminary process after amplifying.In order to reach the requirement for voice signal, front-end circuit comprises microphone input circuit, audio amplifier circuit, filtering circuit and level lifting circuit four part, and this device can be powered by independent current source.First by microphone input circuit voice signal collected and be converted into electric signal, then by audio amplifier circuit, electric signal being amplified, after low-pass filter circuit by frequency control contained by electric signal at below 100HZ.Finally, there is certain level demand because front-end circuit exports the DSP connect for the collection of signal, devise a level lifting circuit at the afterbody of front-end circuit, signal voltage amplitude is raised to necessary requirement.
(1) audio amplifier circuit
Acoustical signal is converted to electric signal by this input circuit, uses the microphone that antijamming capability is strong.Exporting electric signal amplitude is 0-10mv.This circuit adopts audio product to be SSM2019, and it act as and is amplified by the electric signal of microphone, and enlargement factor determines by accessing resistance Rg, and physical circuit figure is shown in Fig. 2.
(2) filter amplification circuit
Process because post-processed needs that simulating signal dress is changed to digital signal, effective information is lower than 60Hz, in order to avoid aliasing effect, carry out analog filtering herein, cutoff frequency 100Hz, chip is adopted to be 8 pin OP07, adjustable second-order low-pass filter (more easily choosing resistance capacitance).The filtering circuit used is common low-pass filter circuit.Physical circuit figure is shown in Fig. 3.
(3) level lifting circuit
DSP enters terminal voltage and requires 0-2.8V, therefore needs the alternating signal lifting after by amplification to be direct current signal, and 1 AD0807 can realize.Physical circuit figure is shown in Fig. 4.
Wherein, 1. its anti-aliasing effect of circuit output capacitance C11, C12 selects large bulk capacitance, and 2. circuit needs this electric capacity equally, provides steady voltage to keep AD conversion voltage stabilization (sampling keeps).
The chip selection major consideration of this front-end circuit is power supply aspect, and positive and negative 5V, to make portable equipment.Can meet for the most chip of other performance index.
DSP processing section adopts TI company 2812 type DSP to gather exporting electric signal by front-end circuit herein, carries out fft analysis after digital-to-analog conversion to signal.After voice signal is gathered by mimic channel, send into DSP, within every 2 seconds, carry out a FFT conversion.
With reference to figure 5, circuit module process flow diagram, is mainly collection, filtering, lifting, computing 4 parts.
With reference to figure 6, DSP (digital signal processor) program design.Program entirety is made up of 5 major parts, is initialization, sampling, FFT conversion, FFT correction and overlap-add procedure respectively.
Sampling section was mainly a cycle timing sampling with 10 seconds, carried out sampling 1024 points with 572Hz sampling rate.In order to make reliable results, need guarantee two conditions, one is that the sampling time is suitable, and the too short sampling time can not the jam situation of actual response road, and this procedure Selection was carry out in one section (sampling period) with 10 seconds; Two is that sample frequency is enough large, and sample frequency is too low, can reduce the resolution of frequency spectrum, is unfavorable for obtaining accurate result.In order to reach higher resolution, we utilize the superposability of acoustic energy to be divided into 5 parts 10 seconds, be accumulated in together, finally make resolution reach 0.558Hz after within every 2 seconds, once carrying out Fourier transform.
FFT conversion fraction mainly utilizes TI company to carry FFT storehouse to carry out Fourier transform.When sampling rate is the quadratic power of sampling signal frequency time, and when sampling precision is very high, the result precision of FFT also height very.But, when sampling rate is the quadratic power of sampling signal frequency, or when sampling precision is very high, have one ungratified time, the precision of FFT is had a greatly reduced quality very much, at this moment revises the result of FFT with regard to needing, and the present invention adopts conventional TI company to carry FFT correction card and revises.
Owing to there is the situation of spectral leakage in FFT conversion, cause the interference appearance of harmonic wave and the energy attenuation at characteristic frequency place.But because front-end circuit carries out good low-pass filtering, and window function is chosen more suitable, therefore harmonic influence is disturbed almost can to ignore, but the energy attenuation at characteristic frequency place can not be ignored, therefore after signal is carried out FFT, provide following algorithm to compensate the decay of energy, namely according to the ratio of the peak-peak point in gained frequency spectrum and the second peak value point, determine corresponding penalty coefficient.The determination of each penalty coefficient is that the correction for native system has good effect by obtain and the lot of experimental data result processed is summed up out.
The frequency spectrum (energy spectrum) that overlap-add procedure part draws after mainly dividing 5 FFT conversion according to the stackable principle of energy sampled signal in 10 seconds stacks up and obtains last accumulative characteristic frequency spectrum (energy spectrum), finds idling frequency (peak value) in the scope of the 20Hz-50Hz of accumulative characteristic frequency spectrum.Coefficient k of blocking up is that after being converted by the energy of 0.2Hz scope near idling frequency amplitude and this amplitude and FFT in 10 seconds, accumulative characteristic frequency spectrum compares got, its reaction be the projecting degree (relative to whole energy spectrum curve steep) of idling frequency.Have obviously outstanding idling frequency in characteristic frequency region when blocking up, and time unobstructed, characteristic frequency spectrum distribution is average.
Show via great many of experiments, in the coast is clear situation sound energy spectrum in, energy is comparatively average in the distribution of individual frequency range, and now curve is milder, or in " burr " shape but curve totally presents smooth, corresponding k value is less.Or due to environmental impact etc., can be concentrated to some extent at 1-2Hz, but due to beyond the span of peak value calculating k value, so the k value calculated is still very little.When road enters congestion status, be namely stopping of not stopping working at measurement point vehicle, the energy spectrum obtained has obvious spike within the scope of 20Hz-40Hz, and curve is very precipitous, and corresponding k value is more much larger than first two situation.Through experiment statistics, when compact car is more, the frequency corresponding to the spike of energy spectrum is lower, between 20Hz-30Hz, wherein particularly little with private car, and taxi and less than 9 small-sized goods, passenger vehicle are larger.And for large car, corresponding to its energy spectrum peak value frequency higher, generally at 30Hz-40Hz, and the k value analyzed is comparatively greatly, and namely concentration of energy situation is more obvious.The calculated value of k has difference clearly in varied situations, and the energy spectrum corresponding to different road conditions has a great difference.As shown in Figs. 7-8.
The inventive method is adopted to judge road congestion conditions
After first set device assembles, We conducted and go out to measure, measurement result presents with the form of k value and energy spectrum maximal value place frequency.Measurement data is pressed the priority serialization number of Measuring Time by us, measures X state (vehicle is in current state) data 400 groups altogether, T state (vehicle stays cool) data 352 groups.Measure place in crossing, Xing Qing road, West Road, Xianning, Xi'an, hand over large north gate.Time main at noon with the traffic peak in afternoon.101-125 group data are as shown in table 1.
Table 1
Find out have invalid data to exist from upper table, as the 102nd group of data of T state, but its probability is lower.
Do not add up with data measured not being added any screening and process, obtain table 2.
Table 2
Therefore 10 are selected as the resolution value of k value.Namely when k<10 is judged as X state; When k > 10 is judged as T state.In surveyed data, the accuracy of this standard reaches 99.601%.Once differentiate because final output is determined as 5 minutes, include 30 groups of data in 5 minutes, make accuracy rate higher.
Within every five minutes, once add up, differentiate any in " block up, busy, unimpeded " of state belonging to it by following standard.Use N
xrepresent the number of X state in this period, N
trepresent the number of T state in this period.
(1) N
t< N
x* 0.3; Then be judged as unimpeded.
(2), N
x* 0.3≤N
t< N
x* 0.7 is judged as busy,
(3) N
x* 0.7≤N
t; Then be judged as blocking up.
Then the LED of respective color is lighted according to judged result.Output can adopt three-color LED light, and reddish yellow is green represents congestion in road, busy, unimpeded respectively.
Claims (3)
1., based on a congestion in road monitoring method for auto idle speed sound spectrum, it is characterized in that, comprise the steps:
(1) the idling sound that driving vehicle on road sends is gathered, and be converted into electric signal, sampling time interval is 10 seconds, by audio amplifier circuit, these electric signal are amplified again, after low-pass filter circuit by frequency control contained by electric signal at below 100HZ, export electric signal to dsp processor finally by level lifting circuit;
(2) dsp processor by containing idling energy spectrum number of signals mode convertion after, in sampling time interval, within every 2 seconds, carry out a Fourier transform, and the energy spectrum superposition after FFT is converted, form the characteristic energy spectrum that accumulative resolution reaches below 0.6Hz, find peak value in the scope of the 20Hz-50Hz composed in this characteristic energy, and the energy of each 0.2Hz scope of this peak value and left and right, peak value place and the accumulative characteristic energy in 10 seconds are composed be business, draw coefficient k of blocking up;
(3) as k < 10, be judged as that vehicle is in current X state; Be judged as that as k > 10 vehicle is in the T state of stagnation;
(4) within every five minutes, once add up, use N
xrepresent the number of X state in this period, N
trepresent the number of T state in this period, when:
A, N
t< N
x* 0.3; Then be judged as the coast is clear;
B, N
x* 0.3≤N
t< N
x* 0.7 is judged as that road is busy;
C, N
x* 0.7≤N
t; Then be judged as congestion in road.
2., as claimed in claim 1 based on the congestion in road monitoring method of auto idle speed sound spectrum, it is characterized in that, described sampling carries out with 572Hz sampling rate, and sampled point is 1024.
3. as claimed in claim 1 based on the congestion in road monitoring method of auto idle speed sound spectrum, it is characterized in that, after signal is carried out Fourier transform, energy compensating is carried out to the result of FFT, namely add up peak point in characteristic energy spectrum and the ratio of the second peak point according to gained, determine corresponding penalty coefficient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310153034.XA CN103247175B (en) | 2013-04-27 | 2013-04-27 | Road congestion monitoring method based on idling sound frequency spectrums of automobiles |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310153034.XA CN103247175B (en) | 2013-04-27 | 2013-04-27 | Road congestion monitoring method based on idling sound frequency spectrums of automobiles |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103247175A CN103247175A (en) | 2013-08-14 |
| CN103247175B true CN103247175B (en) | 2015-01-21 |
Family
ID=48926672
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310153034.XA Expired - Fee Related CN103247175B (en) | 2013-04-27 | 2013-04-27 | Road congestion monitoring method based on idling sound frequency spectrums of automobiles |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103247175B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105702028B (en) * | 2016-02-16 | 2018-07-03 | 中山大学 | A kind of traffic state judging system based on motor vehicle audio signal |
| CN106408947A (en) * | 2016-08-29 | 2017-02-15 | 迈锐数据（北京）有限公司 | Vehicle identification method, vehicle identification device and buried vehicle detector |
| CN113281053B (en) * | 2021-05-21 | 2023-09-12 | 中北大学 | Non-contact acoustic monitoring and LoRa transmission system and method based on stereo garage |
| CN114327040A (en) * | 2021-11-25 | 2022-04-12 | 歌尔股份有限公司 | Vibration signal generation method, device, electronic device and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8457768B2 (en) * | 2007-06-04 | 2013-06-04 | International Business Machines Corporation | Crowd noise analysis |
-
2013
- 2013-04-27 CN CN201310153034.XA patent/CN103247175B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN103247175A (en) | 2013-08-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103247175B (en) | Road congestion monitoring method based on idling sound frequency spectrums of automobiles | |
| CN104149687A (en) | Active noise reduction device suitable for car | |
| CN107483120A (en) | Underwater sound signal gathers and processing system | |
| CN110231117A (en) | A kind of drag-line fundamental frequency feature identification method based on S-transformation | |
| CN205510002U (en) | Amplification and filtering circuit of special signal of telecommunication | |
| CN207215896U (en) | A kind of current detection circuit calculated based on SOC | |
| CN206096489U (en) | Microseism signal pickup assembly | |
| CN103759807B (en) | A kind of airport noise monitoring device based on Internet of Things framework | |
| CN103101493B (en) | The reception device and method of reseptance of parking assisting system | |
| CN110278040A (en) | A kind of high speed weak optical signal receiving module | |
| CN204244188U (en) | Signal based on lock-in amplifier amplifies and testing circuit | |
| CN104597322B (en) | Spectrum analyzer | |
| CN104483547B (en) | The filtering method and system of electric power signal | |
| CN202649436U (en) | Battery power sensing control device | |
| CN106303879A (en) | A kind of detection device based on time-domain analysis and detection method | |
| CN112540240A (en) | Temperature compensation system and method for electric field strength measurement | |
| CN205029873U (en) | Detection device based on time domain analysis | |
| CN214669650U (en) | Cold atom gravimeter ground low-frequency vibration signal processing circuit | |
| CN103714827B (en) | Based on the recognition device of sending a car that voice signal detects | |
| CN110332993A (en) | A kind of high speed is adjustable time domain impulse type balanced homodyne detector and its collecting method | |
| CN206312366U (en) | High accuracy road traffic volume detection device | |
| CN203038025U (en) | Optical phase-shifting interferometer vibration resistance experiment platform intelligent control system | |
| CN102495409B (en) | Speed checking device for vehicles | |
| CN205786850U (en) | A kind of mains by harmonics signals collecting pretreatment circuit | |
| CN209692779U (en) | It is a kind of for capturing the receiving module of high speed weak optical signal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20150121 Termination date: 20190427 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |