CN106790395A - Towards the method for the filtering of numeric type sensing data and transmission of intelligent perception application - Google Patents
Towards the method for the filtering of numeric type sensing data and transmission of intelligent perception application Download PDFInfo
- Publication number
- CN106790395A CN106790395A CN201611060655.3A CN201611060655A CN106790395A CN 106790395 A CN106790395 A CN 106790395A CN 201611060655 A CN201611060655 A CN 201611060655A CN 106790395 A CN106790395 A CN 106790395A
- Authority
- CN
- China
- Prior art keywords
- data
- caching
- moving average
- value
- numeric type
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/06—Protocols specially adapted for file transfer, e.g. file transfer protocol [FTP]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/56—Provisioning of proxy services
- H04L67/568—Storing data temporarily at an intermediate stage, e.g. caching
- H04L67/5683—Storage of data provided by user terminals, i.e. reverse caching
Abstract
The invention discloses a kind of method of the filtering of numeric type sensing data and transmission towards intelligent perception application, including step：1) data are stored in caching by sensors for data；2) the new data for adding caching are processed one by one using moving average method, is filtered out the random error of data；3) judge the size of caching at present, if reaching given threshold, perform below step 4), otherwise return to previous step 1)；4) the data write-in file in caching is stored, caching is emptied afterwards；5) calculate from the last time and be transferred to now time interval, if reaching given threshold, perform below step 6), otherwise return to previous step 1)；6) with current time name step 4) in file, and according to setting packet transmission strategy transfer it to server end.The present invention is based on intelligent perception environment, general can use in the intelligent perception application for having used numeric type sensor, with practicality and versatility.
Description
Technical field
The present invention relates to intelligent perception and mobile computing field, a kind of numeric type towards intelligent perception application is referred in particular to
The method that sensing data is filtered and transmitted.
Background technology
In recent years, developing rapidly with wireless communication technology and Intelligent mobile equipment, the computer application of mobile terminal
Increasingly burning hot, " intelligent perception " concept is also known to more people.Sensor collection by multiple mobile terminals in colony is a large amount of
Data, suggestion and help that each individuality meets their needs are given after processing by analysis, are the new trends of mobile computing,
It is an important ring of smart city development.
Multiple sensors are installed, the reading of wherein numeric type sensor is numerical value on mobile terminal, are also all kinds of shiftings
Quantity, the sensor of most species in dynamic terminal, the physical quantity of measurement is including acceleration, rotating vector, light etc..Android
The reading of the numeric type sensor on platform is all float types, is stored in the float type arrays of entitled values.
Sensing data have refresh that rapid, individual data is scrappy, in the short time the characteristics of accumulation mass data, it is and specific
To mobile phone terminal, then also there are problems that hardware handles limited capacity,.With the hair of mobile phone sensing technology in recent years
Exhibition, the portfolio of related application, class of business and data volume are sharply increased, and the shortcoming of mobile phone terminal is more obvious, more exacerbates hand
Problem of the generator terminal in Data Collection, treatment and transmission.
The filtering and transmission of sensing data are related generally in the present invention.By optimizing the two in sensor application
Must be through link, there is provided a generic service for acceleration processing data, the data-handling efficiency of sensor application can be generally lifted,
So as to realize the optimization to whole intelligent perception network, with certain practicality and wide usage.
The content of the invention
A kind of shortcoming and defect it is an object of the invention to overcome prior art, there is provided number towards intelligent perception application
Value type sensing data filter and transmission method, using and to optimize complexity relatively low and can effectively reduce mobile device internal memory and account for
Filter algorithm, and the transmission plan of efficiency of transmission of the sensing data from mobile terminal to server can be significantly improved
Slightly, general can use in the intelligent perception application for having used numeric type sensor, with practicality and versatility, number can be strengthened
According to accuracy, and improve the efficiency of transmission of data.
To achieve the above object, technical scheme provided by the present invention is：Numeric type towards intelligent perception application is sensed
Device data filtering and the method for transmission, comprise the following steps：
1) data are stored in caching by sensors for data：Angle from exploitation is to be encapsulated as the data in array
Object, is then stored into linked list data structure；Wherein, the use of caching is in order to avoid because of main thread excessively frequently write operation
Cause Android program to be collapsed, first write data into caching in main thread function, then create sub-line journey periodically will be data cached
Write-in file, can be effectively prevented from program interim card, improve program operational efficiency；
2) the new data for adding caching are processed one by one using moving average method, is filtered out the random error of data；
3) judge the size of caching at present, if reaching given threshold, perform below step 4), otherwise return to step above
It is rapid 1)；
4) the data write-in file in caching is stored, caching is emptied afterwards；
5) calculate from the last time and be transferred to now time interval, if reaching given threshold, perform below step 6),
Otherwise return to previous step 1)；
6) with current time name step 4) in file, and according to setting packet transmission strategy transfer it to service
Device end.
In step 2) in the moving average method that uses it is specific as follows：
If the sensor reading of t is yt, wherein exact value is ft, and error is et, then having the time of N number of data
In section, have：
Yt=ft+et (1)
T=1,2 ..., N
In order to reduce the influence of error as far as possible, it is necessary to carry out smooth and filtering process to sensing data；Moving average
The basic thought of method is：For the data of non-stationary, it is considered as smoothly in the interval of preset length, then carries out part
Be averaged so that reduce error influence；And be one group of data of N for length, part is constantly carried out from the beginning to the end to be averaged,
Smoothing can be realized, and filters out fractional error；
If depending on data smoothing in the interval of m length, all data in this interval can be by interval average institute
Substitute, this average is to have filtered the reading value after error and noise, it is believed that be exact value；
If m=2n+1, then there is general type such as following formula：
K=n+1, n+2 ... N-n
E is obtained by formula (1)k=yk‐fkSuch that it is able to estimate the size of error；
It is exactly above simplified moving average method process, in this process it should be noted that preceding n of data with after
N data are due to adjacent data deficiencies, it is impossible to carry out local taking and averagely operate, it is therefore desirable to are supplied；
But it is frequently not that each data is made no exception actually used when taking average, but adds setting
Weights, the moving average method of weighting is calculated as follows formula：
K=n+1, n+2 ... N-n
Wherein piIt is weight coefficient, andDifferent weights values just bring different moving average methods,
When weights are 1/m, formula (3) is just equal to formula (2), is all to wait power moving average；
Obviously, the value of m and p will bring influence to the effect of moving average method, in order to obtain optimal pretreating effect,
It is accomplished by being calculated for the most suitable parameter value of numeric type sensor by experiment, or even needs for each class numeric type
Sensor calculates respective optimal value respectively；
Although moving average method can reduce EMS memory occupation in calculating process by recursion, but still need space
It is data cached for calculate, due to step 1) in employed first cache unify afterwards write method, this causes moving average
Method can also be calculated using same caching, so as to evade this shortcoming.
In step 6) in use packet transmission strategy be after data file is divided into some groups, with " reference value+some
The form of deviation " carries out data transmission, that is, allow server according to reference value and deviation restoring data；Its specific strategy is：
Data are grouped by interval of n, if data are y, then every group of separation is y0、yn、y2n..., the data on separation are directly passed
Server is defeated by, and for yknTo y(k+1)nData y between (k=0,1 ..., (N/n) -1)i, then will be in two kinds of situation：
If 1. | yi‐ykn|<Δ, Δ is the threshold value of setting, then it is assumed that data are in yiPlace's stabilization, abandons transmission, and services
Device end is then given tacit consent in the data for being not received by sometime putting and uses yknTo fill up；
If 2. | yi‐ykn|>Δ, then be sent to server by this difference, by server by utilizing yknValue and the difference are come
Calculate currency.
The present invention compared with prior art, has the following advantages that and beneficial effect：
1st, complexity is low, and cluster or the sensing data processing method of filter method, moving average method are used relative to other
Calculating thinking it is relatively simple, amount of calculation is smaller in the case of application recursion, can relatively well make up mobile device calculate energy
The poor shortcoming of power.
2nd, occupy little space, the caching that data are used when writing can be used for the calculating of moving average method, both reduce
The program interim card that high-frequency write-in is caused, the exceptional space needed for data filtering algorithm is reduced again alleviates mobile device and deposits
Store up the pressure in space.
3rd, with strong points, moving average method belongs to classic data processing algorithm in itself, and non-invention is created, but uses
Moving average method it is critical that data are being weighted with mean time, the weights assigned per number are factor data type and reality
Border demand it is different and different.The present invention will show that the slip run on all kinds of numeric type sensing datas is put down by experiment
The reasonable weights of equal method, can obtain more preferable effect in numerical value sensor environment.And in data transmission, packet transmission
Packet size in strategy is also directed to various kinds of sensors to determine, make use of the numeric type sensing data majority to be in well
The characteristics of more steady in linear change, local interval, can also effectively improve the efficiency of sensor data transmission.
4th, highly versatile, different from the Study on Data Processing carried out in specified context, the present invention is not related to specifically
Sensor application function, the data filtering for being optimized and transmission are also that various kinds of sensors application can all be experienced in data handling
Link, therefore can be applied for many mobile terminals for having used sensor with general.
Brief description of the drawings
Fig. 1 is the inventive method flow chart.
Fig. 2 is the algorithm flow chart of moving average method used in the present invention.
Fig. 3 is transmission strategic process figure of the invention.
Specific embodiment
With reference to specific embodiment, the present invention is described further.
As shown in figure 1, the filtering of numeric type sensing data and transmission towards intelligent perception application described in the present embodiment
Method, comprise the following steps：
1) sensors for data, data are stored in into caching, and (angle from exploitation is to be encapsulated as the data in array
Object, is then stored into linked list data structure)；Wherein, the use of caching is in order to avoid because of main thread, excessively frequently write-in is grasped
Work causes Android program to be collapsed, and caching first being write data into main thread function, then creates sub-line journey will periodically cache number
According to write-in file, program interim card can be effectively prevented from, improve program operational efficiency.
2) the new data for adding caching are processed one by one using moving average method, is filtered out the random error of data；
3) judge the size of caching at present, if reaching given threshold, perform below step 4), otherwise return to step above
It is rapid 1)；
4) the data write-in file in caching is stored, caching is emptied afterwards；
5) calculate from the last time and be transferred to now time interval, if reaching given threshold, perform below step 6),
Otherwise return to previous step 1)；
6) with current time name step 4) in file, and according to setting packet transmission strategy transfer it to service
Device end.
In step 2) in the moving average method that uses be described as follows：
If the sensor reading of t is yt, wherein exact value is ft, and error is et, then having the time of N number of data
In section, have：
Yt=ft+et (1)
T=1,2 ..., N
In order to reduce the influence of error as far as possible, it is necessary to carry out smooth and filtering process to sensing data.Moving average
The basic thought of method is：For the data of non-stationary, it is considered as smoothly in the interval of suitable length, then carries out part
Be averaged so that reduce error influence.It is one group of data of N for length, part is constantly carried out from the beginning to the end and is averaged, i.e.,
Smoothing is capable of achieving, and filters out fractional error.
For example, regarding the interior data smoothing in interval of m length, then all data in this interval can be by interval average
Substituted, this average is to have filtered the reading value after error and noise, it is believed that be exact value.It is right if taking m=5
In f3Have：
f3=1/5 (y1+y2+y3+y4+y5)
If m=2n+1, then there is general type such as following formula：
K=n+1, n+2 ... N-n
E is obtained by formula (1)k=yk-fk, such that it is able to estimate the size of error.
It is exactly above the moving average method process of simple version.In this process it should be noted that the preceding n of data and
N data are due to adjacent data deficiencies afterwards, it is impossible to carry out local taking and averagely operate, it is therefore desirable to are supplied.
But it is frequently not that each data is made no exception actually used when taking average, but adds certain
Weights.The moving average method of weighting is calculated as follows formula：
K=n+1, n+2 ... N-n
Wherein piIt is weight coefficient, andDifferent weights values just bring different moving average methods,
When weights are 1/m, formula (3) is just equal to formula (2), is all to wait power moving average.
Obviously, the value of m and p will bring tremendous influence to the effect of moving average method.In order to obtain optimal pretreatment
Effect, it is necessary to calculated for the most suitable parameter value of numeric type sensor by experiment, or even need for each class
Numeric type sensor calculates respective optimal value respectively, and this is also further work of the invention.
Although moving average method can reduce EMS memory occupation in calculating process by recursion, but still need certain space
Come data cached for calculating, this should be detrimental to its using on the mobile apparatus.But employed in step A
The method for unifying write-in afterwards is first cached, this causes that moving average method can also be calculated using same caching, so as to advise
This shortcoming is kept away, this is also one of the reason for present invention uses the method.
Each algorithm steps for taking the moving average method in the case of 5 weighted averages are carried out in detail with reference to Fig. 2
Description：
A. the number of mark position in caching is found.Mark position points to first not processed number.
B. to current number and four numbers adjacent with it (i.e. the first two and latter two), (each flexible strategy is because of reality to seek weighted average
Border demand and it is different).Therefore moving average method is cannot to process two numbers being stored in earliest in caching.
C. the number tried to achieve is the value after filtering, and current number is replaced with it, moves next to process after mark position
Number.
D. also accessible number is judged whether, criterion is whether mark position is more than 2 with a distance from caching end.
If then returning to step B, step E is otherwise carried out.
E. judge whether caching has expired, criterion is whether cache size has exceeded given threshold.If then entering step
Rapid F, otherwise terminates this time treatment.
F. caching is emptied.Here " emptying " is actually to remove all data in addition to caching the number of most end two,
The purposes for leaving the two numbers are that the data for allowing next new entrance to cache can meet treatment conditions.Terminate flow afterwards.
In step 6) in use packet transmission strategy be after data file is divided into some groups, with " reference value+some
The form of deviation " carries out data transmission, that is, allow server according to reference value and deviation restoring data.Specific strategy is：Will
Data with n be interval packet (most suitable n values need to be determined for the experiment of various kinds of sensors), if data are y, then every group
Separation is y0、yn、y2nDeng.Data on separation are transferred directly to server, and for yknTo y(k+1)n(k=0,1 ...,
(N/n) the data y between -1)i, then will be in two kinds of situation：
If 1. | yi-ykn|<Δ (Δ is the threshold value of setting), then it is assumed that data are in yiPlace is basicly stable, abandons transmission, and
Server end is then given tacit consent in the data for being not received by sometime putting and uses yknTo fill up, so as to improve efficiency of transmission.
If 2. | yi-ykn|>Δ, then be sent to server by this difference, by server by utilizing yknValue and the difference are come
Calculate currency.Do so and to be advantageous in that and reduce the data of transmission by sending the comparatively less difference of numerical value
Amount, so as to improve efficiency of transmission.
As shown in figure 3, the data packet transfer strategic process that the present invention is used is as follows：
A. the packet of current location in file is accessed, if it is separation packet, step B is carried out, is otherwise carried out
Step C.
B. packet is transferred to server.
C. the data difference of current data packet and a upper separation packet is calculated.If the absolute value of difference is more than setting
Threshold value, then carry out step D, otherwise carries out step E.
D. by the differential transmission to server, the separation packet that server end can be received by difference and before
Calculate the value of current data packet.
E. abandon transmit this packet, server end can with the separation packet for receiving before to substitute current when
Between put value.
F. judge whether to have arrived data file end, if then terminating flow, otherwise n will be moved behind current location, and return to
Step A.
Embodiment described above is only the preferred embodiments of the invention, not limits practical range of the invention with this, therefore
The change that all shapes according to the present invention, principle are made, all should cover within the scope of the present invention.
Claims (3)
1. towards the numeric type sensing data filtering of intelligent perception application and the method transmitted, it is characterised in that including following
Step：
1) data are stored in caching by sensors for data：Angle from exploitation is that the data in array are encapsulated as into object,
It is then stored into linked list data structure；Wherein, the use of caching is in order to avoid because of main thread, excessively frequently write operation causes
Android program is collapsed, and caching is first write data into main thread function, then create sub-line journey periodically by data cached write-in
File, can be effectively prevented from program interim card, improve program operational efficiency；
2) the new data for adding caching are processed one by one using moving average method, is filtered out the random error of data；
3) judge the size of caching at present, if reaching given threshold, perform below step 4), otherwise return to previous step
1)；
4) the data write-in file in caching is stored, caching is emptied afterwards；
5) calculate from the last time and be transferred to now time interval, if reaching given threshold, perform below step 6), otherwise
Return to previous step 1)；
6) with current time name step 4) in file, and according to setting packet transmission strategy transfer it to server
End.
2. according to claim 1 towards the numeric type sensing data filtering of intelligent perception application and the method transmitted,
Characterized in that, in step 2) in the moving average method that uses it is specific as follows：
If the sensor reading of t is yt, wherein exact value is ft, and error is et, then having the time period of N number of data
It is interior, have：
Yt=ft+et (1)
T=1,2 ..., N
In order to reduce the influence of error as far as possible, it is necessary to carry out smooth and filtering process to sensing data；Moving average method
Basic thought is：For the data of non-stationary, it is considered as smoothly in the interval of preset length, then carries out local taking
Averagely, so as to reduce error influence；And be one group of data of N for length, part is constantly carried out from the beginning to the end to be averaged, can
It is enough to realize smoothing, and filter out fractional error；
If depending on data smoothing in the interval of m length, all data in this interval can be substituted by interval average,
This average is to have filtered the reading value after error and noise, it is believed that be exact value；
If m=2n+1, then there is general type such as following formula：
K=n+1, n+2 ... N-n
E is obtained by formula (1)k=yk-fkSuch that it is able to estimate the size of error；
It is exactly above simplified moving average method process, in this process it should be noted that preceding n of data and rear n individual
Data are due to adjacent data deficiencies, it is impossible to carry out local taking and averagely operate, it is therefore desirable to are supplied；
But it is frequently not that each data is made no exception actually used when taking average, but adds the weights of setting,
The moving average method of weighting is calculated as follows formula：
K=n+1, n+2 ... N-n
Wherein piIt is weight coefficient, andDifferent weights values just bring different moving average methods, hold power
When value is 1/m, formula (3) is just equal to formula (2), is all to wait power moving average；
Obviously, the value of m and p will bring influence to the effect of moving average method, in order to obtain optimal pretreating effect, just need
To be calculated for the most suitable parameter value of numeric type sensor by experiment, or even need to be sensed for each class numeric type
Device calculates respective optimal value respectively；
Although moving average method can reduce EMS memory occupation in calculating process by recursion, but still need space to cache
Data for calculate, due to step 1) in employed first cache unify afterwards write method, this causes moving average method
Can be calculated using same caching, so as to evade this shortcoming.
3. according to claim 1 towards the numeric type sensing data filtering of intelligent perception application and the method transmitted,
Characterized in that, in step 6) in use packet transmission strategy be after data file is divided into some groups, with " reference value+
The form of some deviations " carries out data transmission, that is, allow server according to reference value and deviation restoring data；Its specific strategy
For：Data are grouped by interval of n, if data are y, then every group of separation is y0、yn、y2n..., the data on separation are straight
Connect and be transferred to server, and for yknTo y(k+1)nData y between (k=0,1 ..., (N/n) -1)i, then two kinds of feelings are divided to
Condition：
If 1. | yi-ykn|<Δ, Δ is the threshold value of setting, then it is assumed that data are in yiPlace's stabilization, abandons transmission, and server end exists
It is not received by then giving tacit consent to during the data sometime put and uses yknTo fill up；
If 2. | yi-ykn|>Δ, then be sent to server by this difference, by server by utilizing yknValue and the difference are calculated
Go out currency.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611060655.3A CN106790395A (en) | 2016-11-28 | 2016-11-28 | Towards the method for the filtering of numeric type sensing data and transmission of intelligent perception application |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611060655.3A CN106790395A (en) | 2016-11-28 | 2016-11-28 | Towards the method for the filtering of numeric type sensing data and transmission of intelligent perception application |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106790395A true CN106790395A (en) | 2017-05-31 |
Family
ID=58913255
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611060655.3A Pending CN106790395A (en) | 2016-11-28 | 2016-11-28 | Towards the method for the filtering of numeric type sensing data and transmission of intelligent perception application |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106790395A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111177883A (en) * | 2019-12-06 | 2020-05-19 | 中电投电力工程有限公司 | Historical climate data-based wind power plant operation and maintenance model |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101267446A (en) * | 2007-12-29 | 2008-09-17 | 中国科学院计算技术研究所 | Time domain data amalgamation method for wireless sensor network |
| CN102316496A (en) * | 2011-09-07 | 2012-01-11 | 上海交通大学 | Data merging method based on Kalman filtering in wireless sensor network |
| CN103176506A (en) * | 2011-12-21 | 2013-06-26 | 富士通株式会社 | Mobile terminal device, medium and control method |
| CN103401804A (en) * | 2013-06-06 | 2013-11-20 | 中国人民解放军理工大学 | Control system and method for node data caching and forwarding of wireless sensor network |
| CN104023356A (en) * | 2014-06-26 | 2014-09-03 | 南京农业大学 | Facilitate environmental control-oriented wireless sensor network data transmission method |
| CN105120516A (en) * | 2015-07-15 | 2015-12-02 | 华南理工大学 | Position information acquisition frame based on crowd sensing environment |
-
2016
- 2016-11-28 CN CN201611060655.3A patent/CN106790395A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101267446A (en) * | 2007-12-29 | 2008-09-17 | 中国科学院计算技术研究所 | Time domain data amalgamation method for wireless sensor network |
| CN102316496A (en) * | 2011-09-07 | 2012-01-11 | 上海交通大学 | Data merging method based on Kalman filtering in wireless sensor network |
| CN103176506A (en) * | 2011-12-21 | 2013-06-26 | 富士通株式会社 | Mobile terminal device, medium and control method |
| US20130165181A1 (en) * | 2011-12-21 | 2013-06-27 | Fujitsu Limited | Mobile terminal device, medium and control method |
| CN103401804A (en) * | 2013-06-06 | 2013-11-20 | 中国人民解放军理工大学 | Control system and method for node data caching and forwarding of wireless sensor network |
| CN104023356A (en) * | 2014-06-26 | 2014-09-03 | 南京农业大学 | Facilitate environmental control-oriented wireless sensor network data transmission method |
| CN105120516A (en) * | 2015-07-15 | 2015-12-02 | 华南理工大学 | Position information acquisition frame based on crowd sensing environment |
Non-Patent Citations (2)
| Title |
|---|
| MINGFU LI: "《Data Filtering and Distortion-Prevention for Power Saving in Wireless Sensor Networks》", 《IEEE》 * |
| 裴益轩: "《滑动平均法的基本原理及应用》", 《火炮发射与控制学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111177883A (en) * | 2019-12-06 | 2020-05-19 | 中电投电力工程有限公司 | Historical climate data-based wind power plant operation and maintenance model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| He et al. | Deep learning based energy efficiency optimization for distributed cooperative spectrum sensing | |
| GB2547816B (en) | Actually-measured marine environment data assimilation method based on sequence recursive filtering three-dimensional variation | |
| CN107786959B (en) | Compressed data collection method in wireless sensor network based on adaptive measuring | |
| CN112668128A (en) | Method and device for selecting terminal equipment nodes in federated learning system | |
| CN106294644B (en) | A kind of magnanimity time series data collection and treatment device and method based on big data technology | |
| Al-Qurabat et al. | Two level data aggregation protocol for prolonging lifetime of periodic sensor networks | |
| CN111367657A (en) | Computing resource collaborative cooperation method based on deep reinforcement learning | |
| CN104281664B (en) | Distributed figure computing system data segmentation method and system | |
| CN111984697B (en) | Cloud computing-based calorimeter metering system and method | |
| CN108120521A (en) | Coiling hot point of transformer temperature predicting method and terminal device | |
| CN108830417A (en) | A kind of residential energy consumption prediction technique and system based on ARMA and regression analysis | |
| CN110121171A (en) | Trust prediction technique based on exponential smoothing and gray model | |
| Zhang et al. | Adaptive directed evolved NSGA2 based node placement optimization for wireless sensor networks | |
| CN106790395A (en) | Towards the method for the filtering of numeric type sensing data and transmission of intelligent perception application | |
| CN106375944A (en) | Data acquisition system based on cloud computing | |
| CN116744289B (en) | Intelligent position privacy protection method for 3D space mobile crowd sensing application | |
| CN108683985A (en) | A kind of WIFI location fingerprints point prescreening method and storage medium | |
| CN110019563A (en) | A kind of portrait modeling method and device based on multidimensional data | |
| CN110610261B (en) | Water body dissolved oxygen prediction method based on neural network | |
| Wei et al. | A multi-objective algorithm for joint energy replenishment and data collection in wireless rechargeable sensor networks | |
| CN109151760B (en) | Distributed state filtering method based on square root volume measurement weighting consistency | |
| CN104794359B (en) | A kind of variable multi-step Q learning adaptive approach of iteration step length | |
| CN116894173A (en) | Water quality tracing method, device, equipment and storage medium | |
| CN113645702B (en) | Internet of things system supporting block chain and optimized by strategy gradient technology | |
| Yan et al. | Service caching for meteorological emergency decision-making in cloud-edge computing |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170531 |