CN104968121A - Automatic learning light control method and apparatus - Google Patents
Automatic learning light control method and apparatus Download PDFInfo
- Publication number
- CN104968121A CN104968121A CN201510417344.7A CN201510417344A CN104968121A CN 104968121 A CN104968121 A CN 104968121A CN 201510417344 A CN201510417344 A CN 201510417344A CN 104968121 A CN104968121 A CN 104968121A
- Authority
- CN
- China
- Prior art keywords
- data
- neural net
- weight data
- light
- light node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
- Y02B20/00—Energy efficient lighting technologies, e.g. halogen lamps or gas discharge lamps
- Y02B20/40—Control techniques providing energy savings, e.g. smart controller or presence detection
Abstract
The invention discloses an automatic learning light control method and an apparatus. The method comprises the following steps of S1, through a setting sensor, acquiring current sensing data influencing a certain light node, taking the current sensing data and current configuration data of the light node as input data and inputting to a nerve network; S2, according to current weight data in the nerve network, calculating and acquiring an output variance value, determining whether the output variance value is in a setting threshold scope, executing a step S4 if the output variance value is in the setting threshold scope, and executing a step S3 if the output variance value is not in the setting threshold scope; S3, according to a preset weight adjusting formula, adjusting the current weight data, taking the adjusted weight data as the current weight data at next iteration, returning to the step S1 and carrying out iteration; S4, taking expected configuration data of the light node as a control selection request and pushing to a client of a user, wherein the expected configuration data is calculated and acquired according to the current weight data in the nerve network. Through a nerve network learning algorithm, automatic control of the light node is realized.
Description
Technical field
The present invention relates to a kind of lamp light control method and device of automatic learning.
Background technology
There is a lot of lamp light control system on the market, as HUE and the yeel ight etc. of PHILIPS.The former uses ZigBee+WiFi, and mobile terminal accesses this system by WIFI, and ZigBee realizes MANET and controls, and the latter uses BLE low-power consumption bluetooth to realize controlled in wireless.Above-mentioned technology can change light configuration, but all just allows user can realize the controlled in wireless of Local or Remote without exception, does not solve the problem of Based Intelligent Control.
Summary of the invention
In order to overcome the deficiencies in the prior art, the object of the present invention is to provide a kind of lamp light control method and device of automatic learning, being realized the automatic control of Light node by Learning Algorithm.
For solving the problem, the technical solution adopted in the present invention is as follows:
Scheme one:
A lamp light control method for automatic learning, comprises the following steps:
S1: obtain the current sensor data affecting certain Light node by the transducer of setting, inputs to neural net using the current configuration data of current sensor data and this Light node as input data;
S2: calculate output variance value according to the present weight data in neural net, judges this output variance value whether in the threshold range of setting, if so, performs step S4, if not, performs step S3;
S3: the weight adjusting formula according to presetting adjusts present weight data, as present weight data during next iteration, and returns step S1 and carries out iteration;
S4: select request to push in the client of user as control the desired configuration data of this Light node calculated according to the present weight data in neural net.
Preferably, further comprising the steps of after step s4:
S5: accordingly adjustment is configured to this Light node according to selection result, and according to selection result and the weight adjusting formula preset, present weight data are adjusted, as present weight data during next iteration, and return step S1 and carry out iteration.
Preferably, in step s 4 which, if be provided with the default s election condition for this Light node in the client of user, then directly perform this acquiescence and select.
Preferably, described neural net is BP neural net.
Scheme two:
A Light Control Unit for automatic learning, comprises with lower module:
Data input module: obtain the current sensor data affecting certain Light node for the transducer by setting, inputs to neural net using the current configuration data of current sensor data and this Light node as input data;
Exporting judge module: for calculating output variance value according to the present weight data in neural net, judging this output variance value whether in the threshold range of setting, if so, carry out desired configuration pushing module, if not, performs the first iteration module;
First iteration module: for adjusting present weight data according to the weight adjusting formula preset, as present weight data during next iteration, and return data input module carries out iteration;
Desired configuration pushing module: for selecting request to push in the client of user as control the desired configuration data of this Light node calculated according to the present weight data in neural net.
Preferably, also comprise with lower module after desired configuration pushing module:
Secondary iteration module: for configuring adjustment according to selection result accordingly to this Light node, and according to selection result and default weight adjusting formula, present weight data are adjusted, as present weight data during next iteration, and return data input module carries out iteration.
Preferably, in desired configuration pushing module, if be provided with the default s election condition for this Light node in the client of user, then directly perform this acquiescence and select.
Preferably, described neural net is BP neural net.
Compared to existing technology, beneficial effect of the present invention is: achieve and carry out training study by neural net to the light configuration scene that user commonly uses, when after learning success, the client expectation light configuration learning to obtain being pushed to user carries out selecting whether to perform or directly perform the configuration of this expectation light, to realize the intelligent automatic control to this Light node.
Accompanying drawing explanation
Fig. 1 is the flow chart of the lamp light control method of automatic learning of the present invention.
Embodiment
Below, by reference to the accompanying drawings and embodiment, the present invention is described further:
With reference to the lamp light control method that figure 1 is a kind of automatic learning of the present invention, comprise the following steps:
S1: obtain the current sensor data affecting certain Light node by the transducer of setting, inputs to neural net using the current configuration data of current sensor data and this Light node as input data;
S2: calculate output variance value according to the present weight data in neural net, judges this output variance value whether in the threshold range of setting, if so, performs step S4, if not, performs step S3;
S3: the weight adjusting formula according to presetting adjusts present weight data, as present weight data during next iteration, and returns step S1 and carries out iteration;
S4: select request to push in the client of user as control the desired configuration data of this Light node calculated according to the present weight data in neural net; If be provided with the default s election condition for this Light node in the client of user, then directly perform this acquiescence and select.
S5: accordingly adjustment is configured to this Light node according to selection result, and according to selection result and the weight adjusting formula preset, present weight data are adjusted, as present weight data during next iteration, and return step S1 and carry out iteration.
This programme preferably uses BP neural network algorithm as learning algorithm; its algorithm complex is lower; descent performance is good; and it is relatively simple; also other neural network algorithms can be selected to substitute BP neural network algorithm; BP neural network algorithm is the comparatively extensive comparatively ripe neural network algorithm of application, and concrete internal algorithm is not that this programme is required for protection, does not too much repeat in this programme.
Below in conjunction with concrete application scenarios example, said method step is described, such as, certain user once repeatedly at night about 7 can be away from home and go for a trot, usually all the light of certain Light node can be dimmed certain brightness when leaving, based on this scene, when people leaves, the pyroelectric sensor that this Light node is corresponding can detect low level, now record the low level signal of this pyroelectric sensor, this pyroelectric sensor transfers the low level time (namely user gos out the time left) to, and the current configuration of this Light node, current configuration comprises the present intensity of light and the numbering of this Light node.The data of above-mentioned record are inputed in the BP neural net of setting as input data, carries out calculating output variance value according to the present weight data in neural net, judge this output variance value whether in the threshold range of setting.When certain sky, this user gos out again within this time period, but have forgotten the lamplight brightness regulating this Light node, if the variance yields that the data now inputted obtain after BP neural computing is in threshold range, then represent Data Convergence, think that the light configuration data wherein calculated is the expectation light configuration of user, now then select request to push in the client of user as control the expectation light configuration data of this Light node calculated, control selection and comprise agreement and do not agree to; If not in threshold range, then represent data also no convergence, needs adjust present weight data according to the data of the current output of BP neural net and weight adjusting publicity, as present weight data during next iteration, wait for the data input of next iteration cycle, then the present weight data that Use Adjustment is crossed are trained, and also namely restart to perform step S1.Object through successive ignition training makes Data Convergence, even if output variance value is in threshold range, preferably, an iterations upper limit is set, when iterations arrives the upper variance yields exported in limited time still not in threshold range, think that data are difficult to convergence, then abandon the data received again about this scene and input.
In step s 4 which, using expectation light configuration data as after control selects request to push to the client of user, if be provided with the default s election condition for this Light node in the client of user, then directly perform this acquiescence to select, such as user is defaulted as and agrees to this adjustment, then the direct light configuration data according to pushing adjusts this Light node, selects without the need to user.
Application scenarios illustrated above is just wherein a kind of, other light application scenarios can also be applied to, generally speaking, as long as the scene regulating round light and frequently occur, inputed in neural net by the acquisition data relevant to this scene and carry out iterative learning, when obtaining output variance convergence, just think and complete automatic learning process, the light configuration that user expects then is thought in the light configuration now obtained, thus realizes the automatic control to light.
Present invention achieves, by neural net, training study is carried out to the light configuration scene that user commonly uses, when after learning success, the client expectation light configuration learning to obtain being pushed to user carries out selecting whether to perform or directly perform the configuration of this expectation light, to realize the intelligent automatic control to this Light node.
The invention also discloses a kind of Light Control Unit of automatic learning, comprise with lower module:
Data input module: obtain the current sensor data affecting certain Light node for the transducer by setting, inputs to neural net using the current configuration data of current sensor data and this Light node as input data;
Exporting judge module: for calculating output variance value according to the present weight data in neural net, judging this output variance value whether in the threshold range of setting, if so, carry out desired configuration pushing module, if not, performs the first iteration module;
First iteration module: for adjusting present weight data according to the weight adjusting formula preset, as present weight data during next iteration, and return data input module carries out iteration;
Desired configuration pushing module: for selecting request to push in the client of user as control the desired configuration data of this Light node calculated according to the present weight data in neural net.
Preferably, also comprise with lower module after desired configuration pushing module:
Secondary iteration module: for configuring adjustment according to selection result accordingly to this Light node, and according to selection result and default weight adjusting formula, present weight data are adjusted, as present weight data during next iteration, and return data input module carries out iteration.
Preferably, in desired configuration pushing module, if be provided with the default s election condition for this Light node in the client of user, then directly perform this acquiescence and select.
Preferably, described neural net is BP neural net.
To one skilled in the art, according to technical scheme described above and design, other various corresponding change and deformation can be made, and all these change and deformation all should belong within the protection range of the claims in the present invention.
Claims (8)
1. a lamp light control method for automatic learning, is characterized in that, comprises the following steps:
S1: obtain the current sensor data affecting certain Light node by the transducer of setting, inputs to neural net using the current configuration data of current sensor data and this Light node as input data;
S2: calculate output variance value according to the present weight data in neural net, judges this output variance value whether in the threshold range of setting, if so, performs step S4, if not, performs step S3;
S3: the weight adjusting formula according to presetting adjusts present weight data, as present weight data during next iteration, and returns step S1 and carries out iteration;
S4: select request to push in the client of user as control the desired configuration data of this Light node calculated according to the present weight data in neural net.
2. the lamp light control method of automatic learning according to claim 1, is characterized in that, further comprising the steps of after step s4:
S5: accordingly adjustment is configured to this Light node according to selection result, and according to selection result and the weight adjusting formula preset, present weight data are adjusted, as present weight data during next iteration, and return step S1 and carry out iteration.
3. the lamp light control method of automatic learning according to claim 1, is characterized in that, in step s 4 which, if be provided with the default s election condition for this Light node in the client of user, then directly perform this acquiescence and selects.
4. the lamp light control method of automatic learning according to claim 1, is characterized in that, described neural net is BP neural net.
5. a Light Control Unit for automatic learning, is characterized in that, comprises with lower module:
Data input module: obtain the current sensor data affecting certain Light node for the transducer by setting, inputs to neural net using the current configuration data of current sensor data and this Light node as input data;
Exporting judge module: for calculating output variance value according to the present weight data in neural net, judging this output variance value whether in the threshold range of setting, if so, carry out desired configuration pushing module, if not, performs the first iteration module;
First iteration module: for adjusting present weight data according to the weight adjusting formula preset, as present weight data during next iteration, and return data input module carries out iteration;
Desired configuration pushing module: for selecting request to push in the client of user as control the desired configuration data of this Light node calculated according to the present weight data in neural net.
6. the Light Control Unit of automatic learning according to claim 5, is characterized in that, also comprises with lower module after desired configuration pushing module:
Secondary iteration module: for configuring adjustment according to selection result accordingly to this Light node, and according to selection result and default weight adjusting formula, present weight data are adjusted, as present weight data during next iteration, and return data input module carries out iteration.
7. the Light Control Unit of automatic learning according to claim 5, is characterized in that, in desired configuration pushing module, if be provided with the default s election condition for this Light node in the client of user, then directly perform this acquiescence and selects.
8. the Light Control Unit of automatic learning according to claim 5, is characterized in that, described neural net is BP neural net.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510417344.7A CN104968121B (en) | 2015-07-15 | 2015-07-15 | A kind of lamp light control method and device learnt automatically |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510417344.7A CN104968121B (en) | 2015-07-15 | 2015-07-15 | A kind of lamp light control method and device learnt automatically |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104968121A true CN104968121A (en) | 2015-10-07 |
| CN104968121B CN104968121B (en) | 2018-02-16 |
Family
ID=54222003
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510417344.7A Active CN104968121B (en) | 2015-07-15 | 2015-07-15 | A kind of lamp light control method and device learnt automatically |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104968121B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108012389A (en) * | 2017-10-27 | 2018-05-08 | 深圳和而泰智能控制股份有限公司 | Light adjusting method, terminal device and computer-readable recording medium |
| CN108712809A (en) * | 2018-05-18 | 2018-10-26 | 浙江工业大学 | A kind of luminous environment intelligent control method based on neural network |
| CN112040590A (en) * | 2020-09-09 | 2020-12-04 | 安徽世林照明股份有限公司 | LED ceiling lamp with induction and light control functions |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050200311A1 (en) * | 2004-02-19 | 2005-09-15 | Oz Optics Ltd. | Light source control system |
| US20120062736A1 (en) * | 2010-09-13 | 2012-03-15 | Xiong Huaixin | Hand and indicating-point positioning method and hand gesture determining method used in human-computer interaction system |
| CN102413605A (en) * | 2011-08-12 | 2012-04-11 | 苏州大学 | Intelligent street lamp energy-saving control system based on artificial neutral network |
| CN202488816U (en) * | 2012-02-07 | 2012-10-10 | 英飞特电子（杭州）有限公司 | Light-emitting diode (LED) light-adjusting control device |
-
2015
- 2015-07-15 CN CN201510417344.7A patent/CN104968121B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050200311A1 (en) * | 2004-02-19 | 2005-09-15 | Oz Optics Ltd. | Light source control system |
| US20120062736A1 (en) * | 2010-09-13 | 2012-03-15 | Xiong Huaixin | Hand and indicating-point positioning method and hand gesture determining method used in human-computer interaction system |
| CN102413605A (en) * | 2011-08-12 | 2012-04-11 | 苏州大学 | Intelligent street lamp energy-saving control system based on artificial neutral network |
| CN202488816U (en) * | 2012-02-07 | 2012-10-10 | 英飞特电子（杭州）有限公司 | Light-emitting diode (LED) light-adjusting control device |
Non-Patent Citations (1)
| Title |
|---|
| 赵晓华，石建军，李振龙，赵国勇: "基于Q—learning和BP神经元", 《公路交通科技》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108012389A (en) * | 2017-10-27 | 2018-05-08 | 深圳和而泰智能控制股份有限公司 | Light adjusting method, terminal device and computer-readable recording medium |
| CN108712809A (en) * | 2018-05-18 | 2018-10-26 | 浙江工业大学 | A kind of luminous environment intelligent control method based on neural network |
| CN108712809B (en) * | 2018-05-18 | 2019-12-03 | 浙江工业大学 | A kind of luminous environment intelligent control method neural network based |
| CN112040590A (en) * | 2020-09-09 | 2020-12-04 | 安徽世林照明股份有限公司 | LED ceiling lamp with induction and light control functions |
| CN112040590B (en) * | 2020-09-09 | 2023-03-07 | 安徽世林照明股份有限公司 | LED ceiling lamp with induction and light control functions |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104968121B (en) | 2018-02-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110308660B (en) | Intelligent equipment control method and device | |
| CN104168371A (en) | Method for adjusting backlight brightness of mobile terminal and mobile terminal | |
| CN110806708A (en) | Intelligent household control method and system and computer readable storage medium | |
| CN104482656A (en) | Smart water heater temperature adjustment method and device and water heater | |
| CN204733413U (en) | A kind of Intelligent LED lamp control device based on degree of depth study | |
| CN109028480B (en) | Constant temperature and humidity control system and method thereof | |
| CN104968121A (en) | Automatic learning light control method and apparatus | |
| CN107172790B (en) | A kind of office's light intelligent regulating system | |
| CN105912632A (en) | Device service recommending method and device | |
| CN106647295B (en) | Smart home system and cooperative operation method thereof | |
| CN111736481A (en) | Model training and intelligent home control method and device based on user behavior characteristics | |
| CN111179878A (en) | Brightness control method for display device of washing equipment | |
| CN111077925A (en) | Control method and device of illumination adjusting equipment, electronic equipment and storage medium | |
| CN110913538A (en) | Desk lamp control method, system and device, desk lamp and desk lamp controller | |
| CN106094555A (en) | A kind of Smart Home voice activated control and home equipment control method | |
| CN111381505A (en) | Intelligent household control system and control method thereof | |
| CN105698396A (en) | Water heater as well as control method and device thereof and mobile communication terminal | |
| CN107390503B (en) | Method and device for controlling alarm clock signal | |
| CN111288604B (en) | Automatic temperature adjusting method and system for air conditioner | |
| CN110320839B (en) | Method and device for controlling intelligent curtain, intelligent curtain and storage medium | |
| CN205320343U (en) | Adjustable lighting models's lighting device | |
| CN106534735A (en) | Intelligent TV volume adjustment method | |
| CN112128951A (en) | Control equipment, method and device | |
| CN110045619A (en) | Curtain intelligent control method and intelligence control system based on self study technology | |
| WO2023279709A1 (en) | Control method and apparatus for air conditioner, and air conditioner |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20180629 Address after: 518000 Feng Tang Road 606, Baoan District Fuyong street, Shenzhen, Guangdong. Patentee after: Classmate Intelligent Technology (Shenzhen) Co., Ltd. Address before: 518000 Guangdong Shenzhen Baoan District Fuyong street high tech Zone Feng Tang Avenue star science and Technology Park B 1 to four floors. Patentee before: ShenZhen Top Technology Co., Ltd. |
|
| TR01 | Transfer of patent right |