A kind of intelligent domestic system and its control method based on voice control
Technical field
The invention belongs to smart home fields, and in particular to a kind of intelligent domestic system and its control based on voice control
Method.
Background technology
Smart home (English:Smart home, home automation) it is using house as platform, utilize comprehensive wiring
Technology, network communication technology, security precautions technology, automatic control technology, audio frequency and video technology are by the related facility collection of home life
Into building the management systems of efficient housing facilities and family's schedule affairs, promote house security, convenience, comfort, skill
Art, and realize the living environment of environmental protection and energy saving.
Smart home is the embodiment of the Thingsization under the influence of internet.Smart home will be in family by technology of Internet of things
Various equipment (such as audio & video equipment, lighting system, curtain control, airconditioning control, safety-protection system, Digital Theater System, audio-visual clothes
Business device, shadow cabinet system, network home appliance etc.) it connects together, home wiring control, Lighting control, remote control using telephone, indoor and outdoor are provided
The multiple functions such as remote control, burglar alarm, environmental monitoring, HVAC control, infrared forwarding and programmable Timer control and means.With
Common household is compared, and smart home not only has traditional inhabitation function, has both building, network communication, information household appliances, equipment certainly
Dynamicization provides comprehensive information exchange function, and even various energy expenditures save fund.
Language is the most important the vehicle of communication of the mankind, and most natural interactive mode.As a kind of man-machine interaction mode,
The purpose of speech recognition is exactly the language that machine is allowed " to understand " mankind.By the research of decades, speech recognition technology is
It is applied among the life of ordinary people.As living standard constantly improves, the concept of home dwelling demand also occurs for people
Variation.
Smart home using advanced computer, embedded system and network communications technology, provides safety and comfort, pleasant
High-grade family life.But traditional intelligence household system control at present is typically by dedicated remote control or cell phone application
The characteristics of completion, remote controler/cell phone application has the function of button and is relatively fixed, thus limit the function of intelligent domestic system
Extension, and remote controller key is usually smaller, is not suitable for the old and the children and uses, and easily generates faulty operation, therefore by language
Sound identification technology is introduced into Intelligent housing, by voice command, can be realized to home wiring control, instead of manually and/or
Remote control control.Therefore, strengthen the research in terms of smart home and speech recognition, constantly improve people’s lives, have positive
Meaning.
The content of the invention
Goal of the invention:The present invention has made improvements in view of the above-mentioned problems of the prior art, i.e. first mesh of the invention
Be to disclose a kind of intelligent domestic system based on voice control be provided, pass through sound identification module and collect people in house
Special instruction is converted into after one's voice in speech, identification and is sent to middle control host, by various intelligence in middle control host computer control house
The switch and response action of household electrical appliances.Second object of the present invention is to disclose a kind of control of the smart home based on voice control
Method processed.
Technical solution:A kind of intelligent domestic system based on voice control, including
At least one speech recognition equipment for collecting human speech, is then handled processed human speech
And corresponding control instruction is changed into, then send the control instruction;
At least one speech recognition equipment pedestal, speech recognition equipment pedestal pass through RS232 serial ports with speech recognition equipment
Or GPIO interface is electrically connected, and for powering for speech recognition equipment, is additionally operable to receive the control that speech recognition equipment is sent
Instruction, and pass through POE interfaces and send the control instruction;
Network device is connected at least one speech recognition equipment pedestal by POE cables, for receiving speech recognition dress
All control instructions that bottom set seat is sent, and all control instructions to receiving are grouped according to the period, then according to
Time order and function order sends control instruction to the POE network switch, wherein:
If receiving multigroup instruction in the same period, choose the most strong control instruction of signal strength and parsed, concurrently
The control instruction after parsing is sent to the POE network switch;
The POE network switch is connected with network device by POE cables, for providing an internal lan environment, and
The control instruction after the parsing that network device is sent is received, and the control instruction after parsing is sent;
Middle control host is connected with the POE network switch by POE cables or wireless Wi-Fi, is handed over for receiving POE networks
Change planes the control instruction sent, and secondary parsing carried out to the control instruction, then according to the control instruction after secondary parsing to
Intelligent appliance sends corresponding actions instruction;
Feedback device is connected with middle control host, for controlling the working condition of host in feedback;
Intelligent appliance is connected with middle control host by RS232 buses or RS485 buses or KNX buses, for being controlled in reception
The action command that host is sent, and operated accordingly according to the action command.
Further, speech recognition equipment includes sound pick-up, analog signal amplifier, AD conversion unit, processor, deposits
Reservoir and network connection device;
The output terminal of sound pick-up is connected with the input terminal of analog signal amplifier;
The output terminal of analog signal amplifier is connected with the input terminal of AD conversion unit;
The output terminal of AD conversion unit is connected with the input terminal of processor;
Processor is connected respectively with memory, network connection device.
Further, speech recognition equipment pedestal includes input interface, processor, memory, network connection device and POE
Interface;
The output terminal of input interface is connected with the input terminal of processor;
The output terminal of processor is connected with the input terminal of POE interfaces;
Processor is connected respectively with memory, network connection device, wherein:
Input interface is RS232 serial ports or GPIO interface.
Further, feedback device includes display, sound equipment, buzzer, LED light
Further, intelligent appliance is washed including intelligent lamp, smart lock, intelligent air condition, intelligent refrigerator, intelligent sound, intelligence
Clothing machine, intelligent electric fan, intelligent door, intelligent access control system, intelligent curtain.
Further, intelligent appliance includes status register, which is used to record the current work of intelligent appliance
Make state.
A kind of control method of the smart home based on voice control, using the above-mentioned smart home based on voice control
System comprises the following steps:
(1) human speech signal is obtained, and the human speech signal is converted into control instruction, then passes through GPIO interface
Or the control instruction is sent to speech recognition equipment pedestal by RS232 serial ports;
(2) control instruction received is transmitted to network device by speech recognition equipment pedestal, and network device is to receiving
All control instructions are grouped according to the period, then send control by the POE network switch according to time order and function order
It instructs to middle control host, wherein:
If receiving multigroup instruction in the same period, the control instruction for choosing voice maximum intensity is handed over by POE networks
It changes planes and is sent to middle control host;
(3) control instruction that aster Resolution receives is controlled in and judges whether the control instruction is effective, if the control instruction is
Illegal command enters step (4);If the control instruction is effective instruction, enter step (5);
(4) code that control instruction can not perform is back to speech recognition equipment and feedback device, then feedback device
The information that outwards issue present instruction can not perform, wherein:
Feedback device includes display, sound equipment, buzzer, LED light;
(5) status register of intelligent appliance that control instruction needs operate is read, judges whether the control instruction can be with
It performs, if can not perform, enters step (6), if can perform, enter step (7);
(6) code that control instruction can not perform is back to speech recognition equipment and feedback device, then feedback device
The information that outwards issue present instruction can not perform;
(7) by the corresponding interface intelligent appliance is controlled to perform relevant operation, (8) is entered step after the completion of operation, wherein:
The corresponding interface is RS232 bus interface or RS485 bus interface or KNX bus interface;
(8) the completed code of control instruction is back to sound identification module and feedback device, then feedback device to
The information that outer issue present instruction is finished.
Further, also need to calibrate n speech recognition equipment to be used before step (1), wherein, n is just
Integer concretely comprises the following steps:
S1 gathers n speech recognition equipment the sound of the normal intensity sent under the conditions of same source of sound same distance one by one
Sound signal, and the intensity of sound information that each sound identification module collects at this time is write down, it is denoted as { a1、a2、...、an};
The average value ρ for all intensities of sound that S2 calculation procedures S1 is collected, calculation formula are:
S3 calculates the correction coefficient F of each speech recognition equipmentx, calculation formula is:
Wherein:X is positive integer and 1≤x≤n.
Further, step (2) the voice intensity is actually to be collected into the intensity of sound and compensation coefficient FxBusiness.
Further, the acquisition human speech signal in step (1) and the human speech signal is converted into control instruction,
Specifically include following steps:
(11) sound pick-up is collected into voice signal, subsequently into step (12);
(12) voice signal that step (11) is collected into is converted into voice digital signal by AD conversion unit;
(13) mode of voice digital signal parsing is selected, (14) are then entered step according to high in the clouds analysis mode, if adopting
With the mode of local parsing, then enter step (15);
(14) voice digital signal that step (12) obtains is uploaded to high in the clouds, step (12) is obtained by Cloud Server
Voice digital signal parsed, the semanteme for parsing voice digital signal is then back to local, will be entered step
(16);
(15) local by Fourier transformation or neural network method to the voice digital signal that step (12) obtains into
Row parses, and after the semanteme for parsing voice digital signal, enters step (16);
(16) the semantic of the voice digital signal parsed is compared with default keyword, if the semanteme parsed
Keyword, then speech recognition equipment can send control instruction corresponding with keyword, if the semanteme parsed is not keyword,
Then speech recognition equipment does not do any reaction.
Further, the process that control instruction is sent to speech recognition equipment pedestal in step (1) is:Speech recognition fills
Bottom set seat sends the handshake with speech recognition equipment, and speech recognition equipment pedestal receives speech recognition equipment handshake and builds
Vertical link accesses whether the inquiry of speech recognition equipment delivery outlet has control instruction more every set time speech recognition equipment pedestal
Newly, the control instruction received packing is transferred to speech recognition equipment pedestal if there is updating then speech recognition equipment.
Further, the process that control instruction is sent to speech recognition equipment pedestal in step (1) is:Speech recognition fills
Bottom set seat sends the handshake with speech recognition equipment, and speech recognition equipment pedestal receives speech recognition equipment handshake and builds
Vertical link, the delivery outlet of speech recognition equipment once have instruction to update, and actively transmit control instruction and are filled to speech recognition
Bottom set seat.
Further, the control instruction sent in step (2) by speech recognition equipment pedestal is successively by frame originating point information, voice
Identification device device number, instruction type, voice signal strength information, host address IP and postamble composition, wherein:
Voice signal intensity of sound information is the voice signal strength information and the voice collected in speech recognition equipment
The business of identification device compensation coefficient.
Further, network device can monitor the finger that all speech recognition equipment pedestals send in real time in step (2)
Information is made, the information that each speech recognition equipment pedestal is sent can judge the true and false of information using voting mechanism, really
What guarantor received is correct information;In section at the same time, if network device only receives a control instruction, which is referred to
Order is sent to middle control host according to ICP/IP protocol via the POE network switch;Within the same period, if network device receives
The control instruction that multiple speech recognition equipment pedestals send over, the control instruction of network device selection intensity of sound maximum are sent
To middle control host.
Advantageous effect:A kind of intelligent domestic system and its control method based on voice control have the advantages that:
1st, without fixed button, the personnel that can understand all age brackets spoken is suitble to use;
2nd, switch and action of the personnel by smart home in voice control house in house can be realized, including background sound
Pleasure is turned on and off, house light is turned on and off.
Description of the drawings
Fig. 1 is a kind of schematic diagram of the intelligent domestic system based on voice control disclosed by the invention;
Fig. 2 a are a kind of workflow schematic diagram of speech recognition equipment;
Fig. 2 b are another workflow schematic diagram of speech recognition equipment;
Fig. 3 is a kind of flow chart of the control method of the smart home based on voice control disclosed by the invention;
Fig. 4 a are a kind of workflow schematic diagram of speech recognition equipment pedestal;
Fig. 4 b are another workflow schematic diagram of speech recognition equipment pedestal;
The composition schematic diagram that Fig. 5 is instructed in order to control.
Fig. 6 is the structural schematic block diagram of speech recognition equipment;
Fig. 7 is the structural schematic block diagram of speech recognition equipment pedestal.
Specific embodiment:
The specific embodiment of the present invention is described in detail below.
Specific embodiment 1
As shown in Figure 1, a kind of intelligent domestic system based on voice control, including
At least one speech recognition equipment for collecting human speech, is then handled processed human speech
And corresponding control instruction is changed into, then send the control instruction;
At least one speech recognition equipment pedestal, speech recognition equipment pedestal pass through RS232 serial ports with speech recognition equipment
It is electrically connected, for powering for speech recognition equipment, is additionally operable to receive the control instruction that speech recognition equipment is sent, and passes through
POE interfaces send the control instruction;
Network device is connected at least one speech recognition equipment pedestal by POE cables, for receiving speech recognition dress
All control instructions that bottom set seat is sent, and all control instructions to receiving are grouped according to the period, then according to
Time order and function order sends control instruction to the POE network switch, wherein:
If receiving multigroup instruction in the same period, choose the most strong control instruction of signal strength and parsed, concurrently
The control instruction after parsing is sent to the POE network switch;
The POE network switch is connected with network device by POE cables, for providing an internal lan environment, and
The control instruction after the parsing that network device is sent is received, and the control instruction after parsing is sent;
Middle control host is connected with the POE network switch by POE cables, for receiving the control that the POE network switch is sent
System instruction, and secondary parsing is carried out to the control instruction, then sent according to the control instruction after secondary parsing to intelligent appliance
Corresponding actions instruct;
Feedback device is connected with middle control host, for controlling the working condition of host in feedback;
Intelligent appliance is connected with middle control host by RS232 buses, for controlling the action command that host is sent in reception,
And it is operated accordingly according to the action command.
Further, as shown in fig. 6, speech recognition equipment includes sound pick-up, analog signal amplifier, analog-to-digital conversion list
Member, processor, memory and network connection device;
The output terminal of sound pick-up is connected with the input terminal of analog signal amplifier;
The output terminal of analog signal amplifier is connected with the input terminal of AD conversion unit;
The output terminal of AD conversion unit is connected with the input terminal of processor;
Processor is connected respectively with memory, network connection device.
Further, as shown in fig. 7, speech recognition equipment pedestal includes input interface, processor, memory, network company
Connection device and POE interfaces;
The output terminal of input interface is connected with the input terminal of processor;
The output terminal of processor is connected with the input terminal of POE interfaces;
Processor is connected respectively with memory, network connection device, wherein:
Input interface is RS232 serial ports.
Further, feedback device includes display, sound equipment, buzzer, LED light
Further, intelligent appliance is washed including intelligent lamp, smart lock, intelligent air condition, intelligent refrigerator, intelligent sound, intelligence
Clothing machine, intelligent electric fan, intelligent door, intelligent access control system, intelligent curtain.
Further, intelligent appliance includes status register, which is used to record the current work of intelligent appliance
Make state.
Specific embodiment 2
It is roughly the same with specific embodiment 1, it differs only in:
Speech recognition equipment pedestal is electrically connected with speech recognition equipment by GPIO interface;
Middle control host is connected with the POE network switch by wireless Wi-Fi;
Intelligent appliance is connected with middle control host by RS485 buses,
Specific embodiment 3
It is roughly the same with specific embodiment 1, it differs only in:
Intelligent appliance is connected with middle control host by KNX buses;
The input interface of speech recognition equipment pedestal is GPIO interface.
Specific embodiment 4
As shown in figure 3, a kind of control method of the smart home based on voice control, using above-mentioned based on voice control
Intelligent domestic system, comprise the following steps:
(1) human speech signal is obtained, and the human speech signal is converted into control instruction, then passes through GPIO interface
The control instruction is sent to speech recognition equipment pedestal;
(2) control instruction received is transmitted to network device by speech recognition equipment pedestal, and network device is to receiving
All control instructions are grouped according to the period, then send control by the POE network switch according to time order and function order
It instructs to middle control host, wherein:
If receiving multigroup instruction in the same period, the control instruction for choosing voice maximum intensity is handed over by POE networks
It changes planes and is sent to middle control host;
(3) control instruction that aster Resolution receives is controlled in and judges whether the control instruction is effective, if the control instruction is
Illegal command enters step (4);If the control instruction is effective instruction, enter step (5);
(4) code that control instruction can not perform is back to speech recognition equipment and feedback device, then feedback device
The information that outwards issue present instruction can not perform, wherein:
Feedback device includes display, sound equipment, buzzer, LED light;
(5) status register of intelligent appliance that control instruction needs operate is read, judges whether the control instruction can be with
It performs, if can not perform, enters step (6), if can perform, enter step (7);
(6) code that control instruction can not perform is back to speech recognition equipment and feedback device, then feedback device
The information that outwards issue present instruction can not perform;
(7) by the corresponding interface intelligent appliance is controlled to perform relevant operation, (8) is entered step after the completion of operation, wherein:
The corresponding interface is RS232 bus interface;
(8) the completed code of control instruction is back to sound identification module and feedback device, then feedback device to
The information that outer issue present instruction is finished.
Further, also need to calibrate n speech recognition equipment to be used before step (1), wherein, n is just
Integer concretely comprises the following steps:
S1 gathers n speech recognition equipment the sound of the normal intensity sent under the conditions of same source of sound same distance one by one
Sound signal, and the intensity of sound information that each sound identification module collects at this time is write down, it is denoted as { a1、a2、...、an};
The average value ρ for all intensities of sound that S2 calculation procedures S1 is collected, calculation formula are:
S3 calculates the correction coefficient F of each speech recognition equipmentx, calculation formula is:
Wherein:X is positive integer and 1≤x≤n.
Further, step (2) the voice intensity is actually to be collected into the intensity of sound and compensation coefficient FxBusiness.
Further, as shown in Fig. 2 a, Fig. 2 b, acquisition human speech signal in step (1) simultaneously believes the human speech
Number control instruction is converted into, specifically includes following steps:
(11) sound pick-up is collected into voice signal, subsequently into step (12);
(12) voice signal that step (11) is collected into is converted into voice digital signal by AD conversion unit;
(13) mode of voice digital signal parsing is selected, (14) are then entered step according to high in the clouds analysis mode, if adopting
With the mode of local parsing, then enter step (15);
(14) voice digital signal that step (12) obtains is uploaded to high in the clouds, step (12) is obtained by Cloud Server
Voice digital signal parsed, the semanteme for parsing voice digital signal is then back to local, will be entered step
(16);
(15) parsed in local by the voice digital signal that Fourier transform pairs step (12) obtains, parse language
After the semanteme of sound digital signal, enter step (16);
(16) the semantic of the voice digital signal parsed is compared with default keyword, if the semanteme parsed
Keyword, then speech recognition equipment can send control instruction corresponding with keyword, if the semanteme parsed is not keyword,
Then speech recognition equipment does not do any reaction.
Further, as shown in fig. 4 a, control instruction is sent to the process of speech recognition equipment pedestal in step (1)
For:Speech recognition equipment pedestal sends the handshake with speech recognition equipment, and speech recognition equipment pedestal receives speech recognition
Device handshake establishes the link, and accessing the inquiry of speech recognition equipment delivery outlet every set time speech recognition equipment pedestal is
It is no to have control instruction update, the control instruction received packing is transferred to speech recognition if there is updating then speech recognition equipment
Device pedestal.
Further, as shown in Figure 4 b, control instruction is sent to the process of speech recognition equipment pedestal in step (1)
For:Speech recognition equipment pedestal sends the handshake with speech recognition equipment, and speech recognition equipment pedestal receives speech recognition
Device handshake establishes the link, and the delivery outlet of speech recognition equipment once has instruction to update, and is actively packaged control instruction and sends out
Give speech recognition equipment pedestal.
Further, as shown in figure 5, the control instruction sent in step (2) by speech recognition equipment pedestal is successively by frame
Header, speech recognition equipment device number, instruction type, voice signal strength information, host address IP and postamble composition,
In:
Voice signal intensity of sound information is the voice signal strength information and the voice collected in speech recognition equipment
The business of identification device compensation coefficient.
Further, network device can monitor the finger that all speech recognition equipment pedestals send in real time in step (2)
Information is made, the information that each speech recognition equipment pedestal is sent can judge the true and false of information using voting mechanism, really
What guarantor received is correct information;In section at the same time, if network device only receives a control instruction, which is referred to
Order is sent to middle control host according to ICP/IP protocol via the POE network switch;Within the same period, if network device receives
The control instruction that multiple speech recognition equipment pedestals send over, the control instruction of network device selection intensity of sound maximum are sent
To middle control host.
Specific embodiment 5
It is roughly the same with specific embodiment 4, it differs only in:
The control instruction is sent to by speech recognition equipment pedestal by RS232 serial ports in step (1);
The corresponding interface is RS485 bus interface in step (7);
It is parsed in step (15) in local by the voice digital signal that neural network method obtains step (12).
Specific embodiment 6
It is roughly the same with specific embodiment 4, it differs only in:
The corresponding interface is KNX bus interface in step (7).
Embodiments of the present invention are elaborated above.But present invention is not limited to the embodiments described above,
In the knowledge that technical field those of ordinary skill possesses, it can also be done on the premise of present inventive concept is not departed from
Go out various change.