CN109951413B - PM2.5 pollution detection method based on multi-antenna WLAN - Google Patents
PM2.5 pollution detection method based on multi-antenna WLAN Download PDFInfo
- Publication number
- CN109951413B CN109951413B CN201910206253.7A CN201910206253A CN109951413B CN 109951413 B CN109951413 B CN 109951413B CN 201910206253 A CN201910206253 A CN 201910206253A CN 109951413 B CN109951413 B CN 109951413B
- Authority
- CN
- China
- Prior art keywords
- csi
- data
- subcarrier
- value
- phase
- 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.)
- Active
Links
Images
Landscapes
- Mobile Radio Communication Systems (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The whole detection method comprises a training stage and a testing stage, CSI data of different PM2.5 concentrations in an indoor environment are respectively collected, statistical characteristics are taken as characteristics of machine learning after data preprocessing, and classification detection is carried out on the CSI data by utilizing a machine learning algorithm. In order to further improve the detection accuracy, a method of synthesizing CSI amplitude and phase information on different antenna pairs is adopted to obtain the final PM2.5 pollution detection. The invention can effectively realize the detection of the indoor PM2.5 with low cost and has certain application value.
Description
Technical Field
The invention relates to the technical field of wireless sensing and signal processing technology, in particular to a PM2.5 pollution detection method based on a multi-antenna WLAN, which is suitable for an indoor health safety system in a complex environment.
Background
Fine particles refer to particles having an aerodynamic equivalent diameter of 2.5 microns or less in ambient air. The higher the content concentration, the more serious the air pollution is. The greater the impact on human health and atmospheric environment.
In the research directions of a plurality of environment perception fields, indoor PM2.5 pollution detection is still an unrelated field, and relevant measures can be taken in time as long as the pollution condition of the indoor PM2.5 can be detected, so that the health of a human body is guaranteed, and the method has important research significance and value in the fields of indoor air environmental protection, atmospheric pollution, human health and the like.
Disclosure of Invention
In order to overcome the defect of poor accuracy of the conventional PM2.5 pollution detection method, the invention provides the PM2.5 pollution detection method based on the multi-antenna WLAN with good accuracy.
In order to solve the technical problems, the invention provides the following technical scheme:
a PM2.5 pollution detection method based on a multi-antenna WLAN comprises the following steps:
step 1: configuring a wireless local area network;
step 2: carrying out orthogonal modulation on original data by adopting a plurality of subcarriers at a transmitting end;
and step 3: estimating the channel state information of each subcarrier, and receiving an OFDM channel state information matrix CSI matrix at a receiving end:
wherein, the CSIi,jThe channel information value of the jth subcarrier on the ith antenna is obtained;
and 4, step 4: performing linear transformation on CSI phase data on all subcarriers by utilizing subcarrier symmetry;
and 5: processing abnormal points by adopting a Lauda criterion, respectively making difference values between each subcarrier value of the acquired data and the subcarrier mean values of other data samples, and if the difference values are more than 3 times of standard deviation, considering the sample value as an abnormal value, and replacing the abnormal value with the mean value;
step 6: carrying out smooth filtering processing by adopting a weighted moving average filter;
and 7: extracting features by using the processed data, wherein the features comprise a maximum value MAX, a minimum value MIN, a MEAN value MEAN, a standard deviation STD, a median absolute deviation MAD, a skewness SKEW and a KUR of 1 to 30 subcarriers;
and 8: processing data characteristics by adopting a normalization method;
and step 9: repeating steps 5 and 6 for the amplitude information;
step 10: combining the amplitude and phase information and repeating step 7;
step 11: detecting each data sample of the test data by using a machine learning algorithm, testing the classification accuracy of the data sample and classifying the pollution;
step 12: and recording the classification result.
Further, in step 4, in performing linear transformation on CSI phase data on all subcarriers by using subcarrier symmetry, the linear transformation method is as follows:
wherein the content of the first and second substances,is the actual measured phase, phiiIs the true phase, kiIndicating the sequence number of the sub-carrier.
Furthermore, in step 5, the implementation method of the ralada criterion is as follows:
wherein the content of the first and second substances,indicating the CSI value, mean, of the f-th sub-carrier of the i-th sample in a certain packetfRepresents the average value, std, of the f-th sub-carrier in the packetfIndicating the standard deviation of the f-th subcarrier.
Still further, in step 6, the method for performing smoothing filtering by using weighted moving average includes:
where n is 1,2, 3.., 30, which indicates a subcarrier number, m indicates a degree of smoothing, and m is set to 9.
The invention provides an indoor PM2.5 pollution detection method based on Channel State Information (CSI) and machine learning of a wireless local area network. The whole detection method mainly comprises a training stage and a testing stage, CSI data with different PM2.5 concentrations are respectively collected, then the CSI data are preprocessed, statistical characteristics of 30 subcarriers on each pair of antennas are taken as characteristics of machine learning, and the characteristics are classified by a machine learning algorithm. Meanwhile, in order to further improve the accuracy of classification, a method of integrating CSI amplitude and phase information on different antenna pairs is adopted, and finally the detected PM2.5 pollution is obtained.
The invention has the following advantages:
1. the advantages of simple deployment, strong anti-interference capability and low price of the wireless local area network equipment are fully utilized;
2. the invention does not need human body carrying equipment, electronic tags and the like, and has certain application value in the fields of indoor air environmental protection, human health research and the like;
3. the WiFi sensing idea is applied to indoor PM2.5 pollution detection, and a new research idea is provided for a PM2.5 detection method;
4. the machine learning algorithm is mainly applied, and the method has good classification effect on the condition of large data volume and multi-concentration pollution;
5. the higher accuracy in multiple environments illustrates the utility of the inventive method.
Drawings
FIG. 1 is a functional block diagram of the present invention;
FIG. 2 is a schematic diagram of an embodiment of a data acquisition and detection system in an environment 1;
FIG. 3 is a schematic diagram of an embodiment of a data acquisition and detection system in environment 2;
FIG. 4 is the indoor PM2.5 pollution detection performance of the present invention in Environment 1;
FIG. 5 is the indoor PM2.5 pollution detection performance of the present invention in Environment 2;
FIG. 6 is a polar plot of CSI amplitude and phase information before and after phase cancellation offset;
FIG. 7 is a graph of CSI magnitude information before and after de-exception;
fig. 8 is a diagram of CSI magnitude information before weighted moving average filtering;
fig. 9 is a diagram of CSI magnitude information after weighted moving average filtering.
Detailed Description
The following detailed description of the preferred embodiments of the present invention, taken in conjunction with the accompanying fig. 1-9, will make the advantages and features of the invention easier to understand by those skilled in the art, and thus will clearly and clearly define the scope of the invention.
Referring to fig. 1, a PM2.5 pollution detection method based on a multi-antenna WLAN includes the following steps:
step 1: configuring a wireless local area network;
step 2: carrying out orthogonal modulation on original data by adopting a plurality of subcarriers at a transmitting end;
and step 3: estimating the channel state information of each subcarrier, and receiving an OFDM channel state information matrix CSI matrix at a receiving end:
wherein, the CSIi,jThe channel information value of the jth subcarrier on the ith antenna is obtained;
and 4, step 4: performing linear transformation on CSI phase data on all subcarriers by utilizing subcarrier symmetry;
and 5: processing abnormal points by adopting a Lauda criterion, respectively making difference values between each subcarrier value of the acquired data and the subcarrier mean values of other data samples, and if the difference values are more than 3 times of standard deviation, considering the sample value as an abnormal value, and replacing the abnormal value with the mean value;
step 6: carrying out smooth filtering processing by adopting a weighted moving average filter;
and 7: extracting features by using the processed data, wherein the features comprise a maximum value MAX, a minimum value MIN, a MEAN value MEAN, a standard deviation STD, a median absolute deviation MAD, a skewness SKEW and a KUR of 1 to 30 subcarriers;
and 8: processing data characteristics by adopting a normalization method;
and step 9: repeating steps 5 and 6 for the amplitude information;
step 10: combining the amplitude and phase information and repeating step 7;
step 11: detecting each data sample of the test data by using a machine learning algorithm, testing the classification accuracy of the data sample and classifying the pollution;
step 12: and recording the classification result.
Further, in step 4, in performing linear transformation on CSI phase data on all subcarriers by using subcarrier symmetry, the linear transformation method is as follows:
wherein the content of the first and second substances,is the actual measured phase, phiiIs the true phase, kiIndicating the sequence number of the sub-carrier.
Furthermore, in step 5, the implementation method of the ralada criterion is as follows:
wherein the content of the first and second substances,indicating the CSI value, mean, of the f-th sub-carrier of the i-th sample in a certain packetfRepresents the average value, std, of the f-th sub-carrier in the packetfIndicating the standard deviation of the f-th subcarrier.
Still further, in step 6, the method for performing smoothing filtering by using weighted moving average includes:
where n is 1,2, 3.., 30, which indicates a subcarrier number, m indicates a degree of smoothing, and m is set to 9.
The PM2.5 pollution detection method comprises the following steps:
1. the equipment platform comprises an Access Point (AP) and a Monitoring Point (MP), wherein the AP is a common wireless router, and the MP is a notebook computer provided with an Intel 5300 network card and channel state information extraction software;
2. the specific implementation site is two different indoor environments, one is a spacious and narrow space, and as spaciousness and silence are realized, interference is relatively small, so that the multipath effect is less, as shown in fig. 2;
3. the second experimental environment is an office in a normal and complex environment, as shown in fig. 3, people interfere with the second experimental environment, more articles exist, the environment is noisy, and the multipath effect is relatively serious;
4. placing two notebooks at proper positions, recording the distance between the AP and the MP, and recording the change of air environment by using a sensor;
5. and collecting data packets of channel state information of PM2.5 with different concentrations, wherein the collection time of each concentration point is 10 seconds, and recording the concentration of PM2.5 in the current air environment. After the collection is finished, a dat file can be obtained in each PM2.5 environment with each concentration;
6. extracting channel state information data of each pair of antennas from each dat file;
7. preprocessing the data;
8. obtaining usable phase information using phase cancellation offsets is shown in fig. 6;
9. removal of outliers using the Lauda method is shown in FIG. 7;
10. normalizing the data;
10-1, smoothing filtering is carried out on the fluctuation signals, as shown in FIGS. 8 and 9;
10-2, the normalization is realized by the following method:
wherein V represents the original characteristic value,denotes the normalized value, VminAnd VmaxRespectively representing the minimum value and the maximum value of a certain characteristic;
11. taking the other two pairs of antennas to repeat the steps;
12. combining the processed amplitude and phase characteristic information together by using a Matlab tool;
13. detecting each data sample of the test data under the environment 1 by using a machine learning algorithm, wherein the algorithm comprises a support vector machine, a random forest and a k nearest neighbor test, the classification accuracy is tested, and pollution classification is carried out, and the classification effect is shown in figure 4;
14. the environment 1 operation was repeated for the test data in environment 2, with the classification effect shown in fig. 5.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes performed by the present specification and drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (1)
1. A PM2.5 pollution detection method based on a multi-antenna WLAN is characterized by comprising the following steps:
step 1: configuring a wireless local area network;
step 2: carrying out orthogonal modulation on original data by adopting a plurality of subcarriers at a transmitting end;
and step 3: estimating the channel state information of each subcarrier, and receiving the channel state information matrix CSI of OFDM at a receiving endmatrix:
Wherein, the CSIi,jThe channel state information value of the jth subcarrier on the ith antenna is obtained;
and 4, step 4: performing linear transformation on CSI phase data on all subcarriers by utilizing subcarrier symmetry;
and 5: removing data outliers by adopting a Lauda criterion;
step 6: carrying out smooth filtering processing by adopting a weighted moving average filter;
and 7: extracting features by using the processed data, wherein the features comprise a maximum value MAX, a minimum value MIN, a MEAN value MEAN, a standard deviation STD, a median absolute deviation MAD, a skewness SKEW and a KUR of 1 to 30 subcarriers;
and 8: processing data characteristics by adopting a normalization method;
and step 9: repeating steps 5 and 6 for the amplitude information;
step 10: combining the amplitude and phase information and repeating step 7;
step 11: detecting each data sample of the test data by using a machine learning algorithm, testing the classification accuracy of the data sample and classifying the pollution;
step 12: recording the classification result;
in step 4, in performing linear transformation on CSI phase data on all subcarriers by using subcarrier symmetry, the linear transformation method is as follows:
wherein the content of the first and second substances,is the actual measured phase, phiiIs the true phase, kiA sequence number indicating a subcarrier;
in the step 5, the implementation method of the Layouda criterion is as follows:
wherein, Vi fIndicating the CSI value, mean, of the f-th sub-carrier of the i-th sample in a certain packetfRepresents the average value, std, of the f-th sub-carrier in the packetfRepresents the standard deviation of the f-th subcarrier;
in step 6, the method for performing smoothing filtering by using the weighted moving average includes:
(m·CSIn+(m-1)·CSIn-1+…+1·CSIn-m+1);
where n is 1,2, 3.., 30, which indicates a subcarrier number, m indicates a degree of smoothing, and m is set to 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910206253.7A CN109951413B (en) | 2019-03-19 | 2019-03-19 | PM2.5 pollution detection method based on multi-antenna WLAN |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910206253.7A CN109951413B (en) | 2019-03-19 | 2019-03-19 | PM2.5 pollution detection method based on multi-antenna WLAN |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109951413A CN109951413A (en) | 2019-06-28 |
CN109951413B true CN109951413B (en) | 2021-11-23 |
Family
ID=67010213
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910206253.7A Active CN109951413B (en) | 2019-03-19 | 2019-03-19 | PM2.5 pollution detection method based on multi-antenna WLAN |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109951413B (en) |
Family Cites Families (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105611627B (en) * | 2016-01-08 | 2019-07-19 | 重庆邮电大学 | The estimation method of WLAN access point AOA based on double antenna |
US10302613B2 (en) * | 2016-10-11 | 2019-05-28 | International Business Machines Corporation | Estimation of concentration of particulate matter in atmosphere |
KR101962013B1 (en) * | 2017-06-09 | 2019-03-25 | 재단법인 다차원 스마트 아이티 융합시스템 연구단 | Method and apparatus for measuring fine particulate matters |
CN108200644B (en) * | 2018-01-10 | 2020-06-02 | 浙江工业大学 | Passive indoor positioning method based on network visible view |
CN108469404B (en) * | 2018-02-01 | 2020-06-02 | 滨州学院 | PM2.5 concentration estimation system and method |
CN108399470B (en) * | 2018-02-09 | 2021-10-08 | 苏州科技大学 | Indoor PM2.5 prediction method based on multi-example genetic neural network |
CN109344901A (en) * | 2018-10-11 | 2019-02-15 | 浙江工业大学 | A kind of driver drives vehicle motion detection method based on wireless channel |
CN109460776A (en) * | 2018-10-11 | 2019-03-12 | 浙江工业大学 | A kind of driver's differentiating method based on channel status detection |
CN109357980A (en) * | 2018-11-02 | 2019-02-19 | 钟祥博谦信息科技有限公司 | PM2.5 data monitoring system and method |
-
2019
- 2019-03-19 CN CN201910206253.7A patent/CN109951413B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN109951413A (en) | 2019-06-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109672485B (en) | Indoor personnel real-time invasion and movement speed detection method based on channel state information | |
Zhou et al. | Device-free presence detection and localization with SVM and CSI fingerprinting | |
Zhou et al. | CSI fingerprinting with SVM regression to achieve device-free passive localization | |
CN108200644B (en) | Passive indoor positioning method based on network visible view | |
CN110475221B (en) | Personnel action identification and position estimation method based on channel state information | |
CN108038419B (en) | Wi-Fi-based indoor personnel passive detection method | |
CN111698258B (en) | WiFi-based environmental intrusion detection method and system | |
CN110072186B (en) | Weighted naive Bayes indoor positioning method based on independent attributes | |
CN105678273A (en) | Initial point detection algorithm of transient signal in radio frequency fingerprint identification technology | |
Yang et al. | Wi-Wheat: Contact-free wheat moisture detection with commodity WiFi | |
CN112333653A (en) | Identity intelligent identification method and system based on WiFi channel state information | |
CN105873212A (en) | Indoor-environment-person detection method based on channel state information | |
CN106685590A (en) | Channel state information and KNN (K-Nearest Neighbor)-based indoor human body orientation recognition method | |
CN110933628B (en) | Fingerprint indoor positioning method based on twin network | |
CN111901028B (en) | Human body behavior identification method based on CSI (channel State information) on multiple antennas | |
CN109698724A (en) | Intrusion detection method, device, equipment and storage medium | |
CN106357575A (en) | Multi-parameter jointly-estimated interference type identification method | |
CN115294615A (en) | Radio frequency fingerprint identification method based on machine learning | |
Tsai et al. | Refined autoencoder-based CSI hidden feature extraction for indoor spot localization | |
CN111465093A (en) | Fingerprint positioning method and device, electronic equipment and computer storage medium | |
CN109951413B (en) | PM2.5 pollution detection method based on multi-antenna WLAN | |
CN112883895B (en) | Illegal electromagnetic signal detection method based on self-adaptive weighted PCA and realization system thereof | |
AbdelAziz et al. | The diversity and scale matter: Ubiquitous transportation mode detection using single cell tower information | |
CN112954791B (en) | Channel State Information (CSI) positioning method based on subcarrier screening | |
CN111934797B (en) | Collaborative spectrum sensing method based on covariance eigenvalue and mean shift clustering |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |