WO2020087515A1 - 主动关怀的实现方法、系统及装置 - Google Patents

主动关怀的实现方法、系统及装置 Download PDF

Info

Publication number
WO2020087515A1
WO2020087515A1 PCT/CN2018/113750 CN2018113750W WO2020087515A1 WO 2020087515 A1 WO2020087515 A1 WO 2020087515A1 CN 2018113750 W CN2018113750 W CN 2018113750W WO 2020087515 A1 WO2020087515 A1 WO 2020087515A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
human body
sensing
information
sensor
Prior art date
Application number
PCT/CN2018/113750
Other languages
English (en)
French (fr)
Inventor
李修球
Original Assignee
李修球
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 李修球 filed Critical 李修球
Priority to PCT/CN2018/113750 priority Critical patent/WO2020087515A1/zh
Publication of WO2020087515A1 publication Critical patent/WO2020087515A1/zh

Links

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B15/00—Systems controlled by a computer
    • G05B15/02—Systems controlled by a computer electric

Definitions

  • the present invention relates to the field of artificial intelligence, and in particular, to a method, system, and device for implementing active care.
  • Smart home is based on the house, using integrated wiring technology, network communication technology, security protection technology, automatic control technology, audio and video technology to integrate home life related facilities, build an efficient residential facility and family schedule management system to improve Home safety, convenience, comfort, artistry, and environmental protection and energy saving living environment.
  • smart devices in smart home systems integrate multi-network, multi-system, and multi-sensor technologies to enable their products to provide users with security, intelligent interaction, and environmental monitoring.
  • the control methods of smart devices are also diverse, including: panel control, APP software control, remote control, voice control, gesture control, sensor linkage control, but these controls The methods all have control defects.
  • the panel control method the number of panels (visual intercom panel, air conditioning panel, alarm keyboard, etc.) that needs to be installed in the home is large, unsightly, and the construction is difficult; for the APP software control method, Users need to click on the hierarchical directory of the APP software to control the home equipment, which is inconvenient to operate, especially for the elderly and children groups; for the remote control method, the remote control placement is not fixed, it is easy to find when needed; for voice / Gesture control is limited by the recognition distance and recognition rate of voice / gesture, and it is difficult to ensure accuracy; for the sensor-linked control method, a human body sensor is usually installed at a specific location in the room, but a human body sensor can only sense one area and feel The measurement range is very limited, and a single sensor cannot judge the user ’s To wait, for example, when someone senses that someone is passing by, the lights are linked and turned on, but this does not have the intelligent analysis feature, that is, I do n’t know when to turn off the
  • the technical problem to be solved by the present invention is that the above-mentioned defects existing in the prior art are complicated operation, low accuracy, and poor user experience.
  • the technical solution adopted by the present invention to solve its technical problem is to construct an implementation method of active care, including the following steps:
  • sensing feature information includes the trigger time, trigger sequence and sensing area of the human sensor
  • the user demand model is established according to the following steps:
  • training sample data When training in the corresponding working mode, obtain training sample data, where the training sample data includes sensing data of at least two adjacent human body sensors triggered within a preset period of time;
  • the plurality of human sensors include infrared array sensors; and,
  • the triggered at least two adjacent human body sensors include the infrared array sensor, and the sensing feature information further includes user gesture information.
  • a camera is provided on the front of the device, and the method further includes:
  • the user characteristic information includes: type, expression, posture, identity, dress, image, quantity, and space characteristics of the house;
  • the invention also constructs an active care realization device, which is communicatively connected with multiple home devices.
  • the realization device includes a main body and a frame around the main body, and further includes:
  • the first acquisition module is used to acquire the sensing data of each human body sensor in real time
  • the first processing module is configured to process the sensing data of the at least two adjacent human body sensors within a preset time period to detect the triggering of the at least two adjacent human body sensors, so as to obtain sensing characteristic information,
  • the sensing feature information includes the triggering time, triggering sequence and sensing area of the human body sensor;
  • the first determining module is used to obtain the current working mode and obtain the user demand model corresponding to the current working mode established in advance, and input the sensing feature information into the user demand model, and according to the The output of the user demand model determines the user's current demand information;
  • the execution module is used to control the corresponding home device and / or output the corresponding first prompt information according to the current demand information of the user.
  • it also includes:
  • the modeling module is used to obtain training sample data when training in the corresponding working mode, and process the training sample data to obtain sensing feature information, and, using a preset training method, the sensing feature
  • the information and the user demand information are trained to obtain a user demand model corresponding to the corresponding working mode, wherein the training sample data includes sensing data of at least two adjacent human body sensors triggered within a preset period of time.
  • the plurality of human body sensors include an infrared array sensor; moreover, at least two adjacent human body sensors triggered include the infrared array sensor, and the sensing feature information further includes user posture information.
  • it also includes a camera, a second acquisition module and a second processing module, and,
  • the second obtaining module is used to obtain the image data of the camera
  • the second processing module is configured to process the image data to obtain user characteristic information, where the user characteristic information includes: type, expression, posture, identity, dress, image, quantity, and space characteristics of the house;
  • the execution module is further configured to output corresponding second prompt information according to the user characteristic information.
  • the plurality of human sensors include:
  • a first human sensor provided on the lower frame of the frame
  • Second human body sensors respectively provided on the left and right side frames of the frame, and the second human body sensors can be rotated along the vertical axis of the corresponding side frame to adjust the sensing direction.
  • the invention also constructs an active care implementation system, which includes a background server and a plurality of home appliances, and is characterized in that it further includes the above-mentioned active care implementation device.
  • Embodiment 1 is a flowchart of Embodiment 1 of a method for implementing active care according to the present invention
  • FIG. 2 is a schematic diagram of the sensing area of each human body sensor of the device for implementing active care according to the present invention
  • Embodiment 1 is a schematic diagram of Embodiment 1 of a sensing area of an implementation device installed in a residence of the present invention
  • 3B is a schematic diagram of the second embodiment of the sensing area of the implementation device installed in a residence of the present invention
  • 3C is a schematic diagram of Embodiment 3 of a sensing area of an implementation device installed in a residence of the present invention
  • Embodiment 4 is a logical structure diagram of Embodiment 1 of an apparatus for implementing active care according to the present invention
  • FIG. 5 is a schematic diagram of a three-dimensional structure of an apparatus for implementing active care according to the present invention.
  • FIG. 6 is a partial schematic diagram of the upward looking direction of the device for realizing active care in FIG. 5 and a schematic diagram of the sensing angle of the bottom sensor;
  • FIG. 7 is a partial schematic side view of the device for implementing active care in FIG. 5;
  • FIG. 8 is a schematic side view of the side sensor in FIG. 7;
  • FIG. 9 is a partial schematic view in a plan view of the device for realizing active care in FIG. 5 and a schematic view of a sensing angle of a side sensor.
  • FIG. 1 is a flowchart of Embodiment 1 of a method for implementing active care according to the present invention.
  • the method of this embodiment can be applied to devices such as smart TVs, smart indoor intercom extension devices, smart robot devices, smart gateways, smart routers, and cell door phones. , And the method includes the following steps:
  • the human body sensors are, for example, infrared, ultrasonic, microwave, pyroelectric element sensors, and may also be array sensors, laser sensors, radar sensors, or the like.
  • a total of 10 human sensors are provided on the frame of the device, and the number of 10 human sensors is not limited according to the sensing angle and accuracy of the human sensor.
  • the bottom of the left frame is set Two human body sensors that sense the space around the side frame have a detection area of 1, 2, and a human body sensor that senses the space below the left frame has a detection area at the bottom of the left frame 3; at the bottom of the right frame, there are two human sensors for sensing the space around the right frame, and their detection areas are 4, 5 respectively.
  • the human body sensor for sensing has a detection area of 6; two human body sensors for sensing the space around the front are provided on the front, the detection areas are 7, 8 respectively, and two pairs of lower frames are also provided at the bottom of the lower frame
  • the detection area of the human body sensor for sensing in the space below is 9, 10 respectively.
  • the human body sensors provided in the left frame and the right frame are angle-adjustable human sensors, that is, they can rotate along the vertical axis of the side frame of the device to adjust the sensing direction.
  • the spatial structure of the site (residential / hotel / meeting room / exhibition hall) should be combined so that the formed human body sensing area of 180 degrees to 360 degrees covers as many users as possible. Or the area that can reflect the most demand, for example, in an application as shown in FIG.
  • the intelligent indoor intercom extension device is installed at the entrance door, and its left detection area 1, 2, 3 covers the kitchen door, restaurant The aisle between the living room and the toilet / master bedroom; the right detection areas 4, 5, and 6 cover the door of bedroom one and the entrance door; and the front detection areas 7, 8, 9, and 10 cover the door of living room and bedroom two.
  • the smart TV is installed on the TV wall of the master bedroom, and its left detection areas 1, 2, 3 cover the toilet door and the main bedroom door; its right detection areas 4, 5, 6 cover the main bedroom window; its front detection area 7, 8, 9, 10 cover the main bed.
  • the intelligent indoor intercom extension device is installed at the entrance door, and its left detection areas 1, 2, 3 cover the entrance door; its right detection area 4, 5, 6 Cover the living room; its front detection areas 7, 8, 9, 10 cover the kitchen door and dining room.
  • the smart TV is installed on the TV wall of the master bedroom, and the distribution of its detection area is similar to FIG. 3A. In an application shown in FIG.
  • the smart indoor intercom extension device is installed at the entrance door, the smart TV is installed on the TV wall of the master bedroom, and the detection areas and figures of the human sensors of the smart indoor intercom extension device
  • 3B the detection area and figures of the human sensors of the smart indoor intercom extension device
  • the two devices are a smart indoor intercom extension device installed at the entrance and a smart TV installed in the master bedroom, respectively.
  • the interaction of the sensing data for example, when the user comes out of the master bedroom and enters the living room, the passing human sensors are triggered in sequence, so that the user can recognize the complete walking route from the master bedroom to the living room.
  • some human sensors can also be installed in the bedroom where the smart TV is not installed, and these human sensors can be connected to the smart indoor intercom extension device or the smart TV, so that, for example, when the user When the bedroom comes out into the living room or the master bedroom, the passing human sensors are triggered in sequence, so that the user can recognize the complete walking route from the bedroom to the living room or the master bedroom.
  • sensing feature information includes the trigger time, trigger sequence and sensing area of the human sensor
  • the home device includes, for example, smart lights, smart curtains, and the like.
  • the correspondence between the demand information and the control instructions / first prompt information of the home device can be preset and stored by the user, and can also be acquired by the background server through big data learning and transmitted to the corresponding device.
  • the demand information is the user When returning home from work, the control of the corresponding home device is: control the entrance light, living room light, kitchen light, etc .; control the air conditioner to turn on at 26 degrees.
  • the corresponding first prompt message is: Welcome the host to go home, etc.
  • the demand information is that when the user wakes up at night, the corresponding home device control is: controlling the toilet light to turn on; otherwise, the demand information is when the bathroom is finished after going to bed, the corresponding home device control is: controlling the toilet light to turn off.
  • a human body sensing area of 180 degrees to 360 degrees is formed, in which the human body sensor in any area is triggered, enters the awake state, and starts timing, when the user is in a custom
  • the adjacent second zone is triggered within a preset time, it means that the user has moved from the first zone to the second zone. It also shows that the user is indeed in this zone.
  • the user ’s current Demand information, and then output prompt information such as voice, video, audio, and images, or control home devices. If the second zone is not triggered within the preset time, it means that the user is only active in the first zone, or there is a false alarm in this zone.
  • the area or intersection area can be judged to be aware of the user according to the spatial structure of the apartment, and the single sensor can be confirmed through machine learning and big data analysis There is no false alarm in the area, and the N time can also be defined as zero.
  • the human sensor triggers to report the current status to the system. For example, after the arming is directed to the detection area of the door without signal interference, the single sensor area detects someone illegally invaded.
  • the system reports an illegal intrusion; if the horizontal sensor is triggered at the same time, the left / right / front proximity sensor at the bottom of the device is also triggered, indicating that the user is within 2 meters of the device in the horizontal trigger area; if the user only triggers Left / center / right level sensor, indicating that the user is within 2 meters of the device; when arming at night, only the sensors at the level of the entrance door and the lower part are armed, and other sensors can be used to realize the user's night-wake linkage Control function, plus machine learning and logic judgment, the user starts from the bedroom side sensing, connects to trigger multiple sensors and then touch When arming the area, the system still judges the normal night-up mode instead of triggering the alarm; when the holiday or the elderly are at home, the user may wander repeatedly because of thinking problems / boring / housework / playing, etc. Frequency, trajectory, time characteristics, user type, expression, behavior, etc. are machine learning and intelligent analysis judgment, and actively output voice, video, audio, image or control
  • the user demand model is established according to the following steps:
  • training sample data When training in the corresponding working mode, obtain training sample data, where the training sample data includes sensing data of at least two adjacent human body sensors triggered within a preset period of time;
  • the user demand model needs to be trained.
  • training first obtain a large amount of training sample data, then extract the sensing feature information, and then train the sensing feature information and the user's actual demand information according to the preset training method, so as to obtain the user demand model.
  • a high-probability event is: if the user is obtained when the user passes through the porch to open the door at about seven o'clock on the working day, the two human sensors at the porch are triggered successively, and then go to work; and usually around seven o'clock in the evening.
  • the user ’s demand can be determined to be in the morning; if the user is at 7 p.m. When you enter the door and pass the entrance, you can determine that the user's demand is to get off work at night; if the user enters the trigger for more than a predetermined time and the user is detected from the kitchen Entering the restaurant, it can be determined that the user's needs are for dinner; if the user wanders repeatedly, it can be judged that the user is thinking / boring / fitness; if only passes the proximity sensor multiple times, it can be judged as a child or an animal; if in the same If the area stays for a long time, you may rest / ill / entertainment.
  • the user demand model when the user demand model is obtained, the user demand model can be further verified using the training sample data. If the verification is passed, it can be applied. If the verification fails, the training sample data needs to be further collected to continue training until the acquisition To the verified user demand model.
  • the plurality of human body sensors include an infrared array sensor.
  • the infrared array sensor is disposed on the front of the device.
  • the at least two adjacent human body sensors triggered include the infrared array sensor.
  • the extracted sensing feature information After processing the sensing data of the infrared array sensor, the extracted sensing feature information also includes the user's posture information. For example, referring to FIG. 3A to FIG.
  • the posture information of the user obtained through the infrared array sensor includes: getting up, lying down, turning over, etc .; when the smart indoor intercom extension device is installed in the living room
  • the user's posture information that can be obtained through the infrared array sensor includes: falling, sitting down, walking, lying down, moving direction, speed, frequency, body temperature, etc.
  • a camera is provided on the front of the device, and the method of the present invention further includes:
  • the user characteristic information includes: type, expression, posture, identity, dress, image, quantity, and space characteristics of the house;
  • this embodiment when there are multiple members in the family, including the elderly, children, and middle-aged persons, for example, it is still impossible to distinguish which member triggered the human sensor based on the sensing data of the human sensor, so that the user cannot be accurately known
  • This embodiment can accurately analyze the real needs of the user by combining the sensing data of the human sensor and the image data of the camera. For example, it is also detected that the user goes out at seven in the morning.
  • a voice prompt prompting the elderly can be issued; if the user is determined to be a child, the user's current needs are determined to be school, and then a voice prompting the child can be issued Prompt; if it is determined that the user is a middle-aged person, it is determined that the current demand of the user is to go to work, and then a voice prompt for the middle-aged person may be issued.
  • the training sample data may also add image feature information.
  • the input information of the user demand model includes image feature information in addition to the sensing feature information, so that the user demand information output by the user demand model is more accurate and smarter.
  • Embodiment 4 is a logical structure diagram of Embodiment 1 of an implementation device for implementing active care according to the present invention.
  • the implementation device is in communication connection with multiple home devices (such as smart lights, air conditioners, and smart curtains provided in bedrooms, toilets, living rooms, and dining rooms).
  • the implementation device includes a main body (not shown) and a frame (not shown) around the main body, and further includes: a plurality of human sensors (not shown), a first acquisition module 10, a first processing module 20, The first determination module 30 and the execution module 40.
  • a plurality of human sensors are installed on the frame to form a plurality of human sensors to form a human body sensing area greater than 180 degrees on the periphery of the implementation device;
  • the first acquisition module 10 is used to acquire the real-time Sensing data;
  • the first processing module 20 is used to process the sensing data of the at least two adjacent human body sensors within a preset period to detect the triggering of the at least two adjacent human body sensors to obtain a sense Feature information, the sensing feature information includes the trigger time, trigger sequence and sensing area of the human sensor;
  • the first determination module 30 is used to obtain the current working mode and obtain the user corresponding to the pre-established current working mode Demand model, and input the sensing feature information into the user demand model, and determine the current demand information of the user according to the output result of the user demand model;
  • the execution module 40 is used to control according to the current demand information of the user The corresponding home device and / or corresponding first prompt information is output.
  • the implementation device of the present invention further includes a modeling module for acquiring training sample data when training in the corresponding working mode, and processing the training sample data to obtain sensing feature information, and Using a preset training method to train the sensing feature information and user demand information to obtain a user demand model corresponding to the corresponding working mode, wherein the training sample data includes at least two triggers that are triggered within a preset time period Data from two adjacent human sensors.
  • the plurality of human body sensors include an infrared array sensor; moreover, at least two adjacent human body sensors triggered include the infrared array sensor, and the sensing feature information further includes user posture information.
  • the implementation device of the present invention further includes a camera, and further includes a second acquisition module and a second processing module, wherein the second acquisition module is used to acquire the image data of the camera; the second processing module is used to The image data is processed to obtain user characteristic information, and the user characteristic information includes: type, expression, posture, identity, dress, image, number, and space characteristics of the house; the execution module is further used to output according to the user characteristic information The corresponding second prompt message.
  • the multiple human sensors include:
  • a first human sensor disposed on the lower frame of the frame preferably, the first human sensor is disposed on the bottom of the lower frame;
  • Second human body sensors respectively provided on the left and right side frames of the frame, and the second human body sensors can be rotated along the vertical axis of the corresponding side frame to adjust the sensing direction.
  • the second human sensor is provided at the bottom of the corresponding side frame.
  • the implementation device of this embodiment includes a main body 1 and a frame 2 located around the main body 1, and a blank area 3 with a certain width is reserved around the frame 2 to the main body 1. It is convenient for the normal transmission and reception of wireless signals, the heat dissipation of internal components to the outside, the integration of interactive components with users (such as touch buttons), and the installation of other sensors (such as cameras, infrared array sensors, microphone arrays, etc.) ).
  • the frame 2 adopts a hollow design, and forms an entire space for the installation of internal components together with the area other than the combined part of the main body 1 and the concave limit bracket area.
  • the plane bracket 4 is mainly connected by a detachable connecting shaft, so only the wall plane bracket 4 needs to be installed, and the connecting shaft can be inserted to form a firm connection between the bracket and the device, which is simple and easy to install.
  • the lower side and the left side of the concave limit area of the middle part of the back plane of the device are designed with an interface channel 41 for external connection.
  • the interface channel 41 includes a device interface or outlet slot, which is an internal component of the device connected to an external cable or interface After the device is installed close to the wall, all cables are concealed in the concave restricted area on the back of the device.
  • the space of the concave restricted area plus the wall hole space can fully accommodate all cables or interfaces that the device needs to connect to the outside.
  • the rotatable design is mainly a circular protruding axis 5 protruding in the middle of the concave limit area.
  • the circular protruding axis 5 mainly provides a rotary connection and also functions as a fixing device end bracket.
  • the protruding plane can be screwed to fix the shaft cover or welded to fix the shaft cover.
  • the protruding space can be enough to accommodate the screw protruding part of the fixing screw.
  • the shaft 5 is provided with a back plate rotatably connected to the shaft 5. Using the back plate as a plane rotation axis can reduce the thickness of the plane axis by 1/3, and also simplify the structure realization of the plane axis.
  • human body sensors 6 are provided at the lower two corners of the frame 2 respectively, and each human body sensing device 6 senses the space under the lower frame 21 of the frame 2 corresponding to the corner and the space around the side frame 22 of the frame 2 .
  • the human sensors 6 at the lower two corners combined with the human body sensor on the front side can realize human body sensing in multiple spaces such as the front side, the lower side of the lower frame 21 and the outer side of the side frame 22.
  • Each human body sensor 6 includes a bottom sensor 61 and a side sensor 62.
  • the bottom sensor 61 is provided at the bottom of the corresponding corner to sense the space under the lower frame 21 corresponding to the corner
  • the side sensor 62 is provided at the side of the corresponding corner.
  • the lower two corners of the frame 2 may be rounded.
  • the bottom sensor 61 includes two first human sensors 611, and the sensing direction of one of the first human sensors 611 is inclined toward the front side of the main body 1, and the sensing direction of the other first human sensor 611 is toward The outside of the side frame 22 of the frame 2 corresponding to the corner is inclined.
  • the surface of the lower frame 21 is provided with a A protruding structure 211 installed with a human body sensor 611 can make one of the two first human body sensors 611 incline below the downward direction of the frame 21 in the front direction of the plane intelligent robot, and the other towards the frame corresponding to the corner
  • the outside of the side frame 22 of 2 is inclined, and it will not be affected by blocking or the like after the inclination.
  • the frame 2 is a hollow trapezoidal structure design.
  • the device frame 2 is designed with a snap structure on the short side and the long side of the trapezoid, and the card is installed on this trapezoidal slope
  • the button-type sleeve 7 allows the user to change the color, pattern and style of the sleeve 7 according to the interior decoration style. It can also enhance the overall artistic effect of the device and reserve more space for installation of internal components for the device.
  • the side sensor 62 includes two second human sensors 621 that are horizontally arranged in the sensing direction, and the sensing directions of the two second human sensors 621 are disposed outside the side frame 22 where the two are located.
  • the sensing direction of the second human sensor 621 is set at an angle to increase the sensing range.
  • One side of the second human body sensor 621 in the sensing range is parallel to the back of the main body 1, and the other second human body sensor 621 swings toward the front side of the main body 1 to increase the sensing range.
  • the two second human sensors 621 are arranged side by side up and down.
  • the two second human sensors 621 can rotate along the vertical axis to adjust the sensing direction.
  • only one second human sensor 621 can rotate.
  • the outer side of the side frame 22 of the bezel 2 is provided with a semi-circular boss 221 that protrudes outward, and the axis of the arc surface of the boss 221 is vertically arranged.
  • the two second human body sensors 621 are rotatably disposed in the boss 221, and two convex sensing ports 222 are provided on the boss 221 for the two second body sensors 621 to sense outward
  • the arc-shaped sensing port 222 is horizontally arranged, and swings along the arc-shaped sensing port 222 when the second human body sensor 621 rotates to realize the adjustment of the sensing range.
  • the boss 221 is provided with two rotatably rotating knobs 223, and the two rotating knobs 223 are respectively connected to the two second human sensors 621 to drive the second human sensors 621 to rotate and adjust the The sensing direction of the second human sensor 621.
  • the detection angle of the second human sensor 621 can be adjusted in the horizontal direction.
  • the plane intelligent robot When the plane intelligent robot is installed near the right-angle wall, plus the sensor in the frontal space area, it can form a horizontal human detection range of 180-270 degrees. Both are near the right-angled wall, and can also form a horizontal 360-degree detection range, so it is suitable for the detection perception of any type of spatial structure.
  • the structural design of the semi-circular boss 221 saves the internal space of the planar intelligent robot, and also ensures the need for sensor angle adjustment, ensuring that the detection range of the sensor reaches the maximum value.
  • the present invention also constructs an implementation system for implementing active care, including a back-end server, multiple home appliances, and the implementation device shown in the foregoing embodiment.
  • the device senses that the user passes through the aisle at the entrance door at seven in the morning (for example, from the 2 # sensing area to the 1 # sensing area), or the device obtains the voice information of the user when going out. , I went to work, help me look good at home ", you can get the user's current needs to leave home, and then automatically reply" Master is safe!
  • a voice prompt "Please open my eyes (camera) / ear (microphone) perception before going out, I can help you look at your home", when the user is close to the device to open
  • Hidden sensors for example, when the camera is turned on, the device can obtain the user's current image / expression / clothing / body temperature and other statuses through video recognition, and actively voice care for the user according to the recognition result: “Master, you seem to have no reason to beard when you go out today" or "Master, don't you need to wear formal attire today?" Or “Master, I feel your temperature is a bit high, is it uncomfortable there?" Or “Master, I don't think you are in good spirits, you slept too late last night OK, please pay more attention to the body ", etc., the user can choose to reply, and then close the door and leave, the device automatically records the user's personal status when replying to the door, so as to care for the user as needed when returning home, the system automatically
  • the system automatically turns on the home mode, the device automatically turns on the living room lights, curtains and air conditioning, art screen display and other equipment, and is passing Video recognition acquires user identity / expression / body temperature and other characteristics, and actively cares for users with different semantic voices for different users or different expressions or different body temperatures or different time periods, such as: sensing the little master ’s return home in the afternoon after school time, device Active children's voice greets "Little Master, are you happy to go to school today?
  • the device when the device senses that the user opens the entrance door, it automatically greets the user, not only does it not disarm, but also senses that the user directly walks to other areas in the room, the system automatically alarms within a certain time, and automatically uploads the on-site visual and audio conditions to the property center or owner Mobile phone or service platform.
  • the device in the bedroom senses that the user uncovers the quilt, sits or stands up
  • the device senses that the user wakes up, automatically turns on the background music, opens the curtains, opens the artwork or early morning landscape painting display, and actively cares for the user "Master, good morning, yesterday evening
  • the indicators of the indoor air quality process have not exceeded the standard, should you sleep well?
  • the bathroom lights and exhaust fan are automatically turned on, which is convenient for the user to wash his face and clean his teeth before going to the toilet.
  • the bathroom lights and exhaust fan are automatically turned off.
  • the demand is to sleep at night, and output "good night", and then automatically turn on the bedroom lights, close the curtains, and adjust the air conditioner to the appropriate temperature.
  • the living room lights are automatically turned off, and the display screen of the device is automatically turned off to enter arming.
  • the device senses that the user uploads and covers the quilt during the usual sleeping time, or obtains the user's voice message "Little Ball, I'm Sleeping", and can also determine that the user's current demand is night Go to bed and output the voice message "Master, good night, now the bedroom temperature is 24 degrees, the humidity is 50%, you need to cover the air conditioner to sleep", and then automatically turn off the bedroom lights, close the curtains, and adjust the air conditioner to the appropriate temperature.
  • the device When the device is installed in the bedroom, when the device senses that the user uncovers the quilt, sits or stands up, the system automatically turns on the bedroom dark light. When the user opens the bedroom door, it automatically turns on the dark light in the living room or corridor. When the device senses the user from the bedroom side to the toilet side through the sensor, the lights of the relevant corridor and toilet light up. When the device senses that the user reverses from the toilet to the bedroom side through the sensor, the lights of the related corridor and toilet are automatically turned off. After the user goes to bed, the dark lights in the bedroom are automatically turned off.
  • the device controls the light switch by recognizing the user's needs, which is energy-saving, hygienic and privacy-protective.
  • the user will automatically sense and open the night mode when the bedroom door is opened, and the dark light in the living room will be turned on.
  • the related corridor and toilet lights are on. Up.
  • the lights of the related corridor and toilet are automatically turned off.
  • the device can define a single sensor, dual sensors, multi-sensors and the trigger logic of the sensor to sense the human body. If an illegal intrusion is detected, the system will automatically Voice alarm and push alarm information to property and smart home service platform.
  • Night Arming Status Arming at night mainly arm the sensors in the detection area of the entrance door and important doors and windows, and define the logical trigger from the user ’s bedroom side sensor as the first trigger to the night mode, and the whole process from the bedroom to the toilet
  • the sensing sequence logic is defined as the light-on mode, and when the reverse sensing logic is triggered from the toilet to the bedroom, it is defined as the light-off mode, and the dim light from the toilet to the bedroom is gradually turned off.
  • the night-up mode if the user needs to pass the armed area sensor during the toilet trip, the device still judges it as the night-up mode, but if the sensor on the entrance door side is the first sensing user, the system considers it an illegal intrusion.
  • the device When the device senses that only the elderly are at home indoors, for example, if the elderly stays in an area for a long time, such as in the living room sofa, the device actively cares for "elderly, you have been on the sofa continuously when the device judges that it exceeds the custom time threshold. After sitting for 3 hours, you are not in good health and not suitable for long periods of immobility "or” Elderly, you have been reading for more than 3 hours, let's relax with some music "or” Elderly, there is a square dance music you like Do n’t interact together? ”If the voice is not answered by the old man, the device automatically turns on the camera and microphone, and the video analyzes the user ’s current state.
  • the emergency music is automatically turned on, which reminds the old man if he does not reply within 2 minutes. , That is, an alarm signal is formed.
  • the elderly have hearing problems, they can also flash the lights to remind the user that the system will automatically push the current emergency state of the elderly to the property or relatives or service platform after more than 2 minutes. If the elderly stays in a room that the device cannot perceive for a long time, instead of the normal rest time, when the elderly's children return home, they will promptly inform the user "Master, your father / mother N times in a certain room Staying for a long time is not good for your health.
  • the device plays emergency music in the living room. For elderly people who are physically walking or hearing impaired, the device automatically dials the elderly phone or flashes the indoor lights to remind the elderly to walk around. If the voice / phone / music / flashing lights are not answered by the elderly, the system will automatically send the elderly to the current emergency state Push to property or relatives or service platform.
  • the elderly user needs to take medicine at about 11 o'clock every day, when the device actively cares for the elderly at 11 o'clock in the morning, "Master, your gastrointestinal tract is not good. The gastrointestinal medicine needs to be taken 1 hour before meals.
  • Emergency button trigger When the emergency button is triggered, the camera is automatically turned on to identify the current state of the elderly on the spot, or the elderly is sensing the area on the side of the device, and immediately analyzes whether the sensor perceives that the user has been in the alarm area for the last N minutes. If it indicates that the elderly may be sick A state of emergency is caused. If it does not mean that the elderly may be falling, immediately report to the property or relatives or service platform, and report the relevant data of the perception user and the health data of the elderly.
  • the elderly are idle: the device senses that the elderly are wandering N times (frequency / speed / direction / distance / distance / region / elderly habits / time node etc. to determine whether it is wandering), and can actively care for the elderly "master, are you encountering Unhappy? "Or” Master, you wander around at home with limited space, you can go outside and take your positioning table so that I can know where you are at any time "or” Master, there is a treasure calligraphy work recently You can check it out, who created it in N years, and the background at that time was ... "or” Master, you have n’t heard the storytelling last time, would you like to listen to it? ".
  • the initiative to care for and interact with the elderly according to the current situation of the elderly, let the elderly do what they want to do during leisure, and relieve the elderly from being bored at home, lack of care and communication.
  • the device senses that there are users or multiple users or the last user walking from the dining room or bedroom to the living room, and the time is 8 o'clock at the beginning of a continuous round, and then actively cares for the "master, the TV station watched last night. Now it ’s started, do you want to watch it together? ”If the user replies“ Okay ”, it will automatically turn on the audio and video mode, automatically turn off the lights in the dining room, toilet, and bedroom, and dim the lights in the living room. When the sound of children playing in the living room is too loud, the device actively cares for the user "Children, please do not make loud noises at night and do not affect others to rest”.
  • the device When someone temporarily goes to the toilet, the device senses the user's perception logic from the living room to the toilet, and automatically turns on the corridor and toilet lights. When the user comes out of the toilet and goes to the living room to watch TV, the device senses the user's perception logic from the toilet to the living room, automatically Turn off lights in toilets and corridors.
  • the device senses that the user is triggering the 9 # and 10 # sensors, it automatically switches the content of the art display, and can set the appropriate voice volume to introduce the origin, author, characteristics, historical story and other content of the artwork
  • the device When the device senses that the user is not interested in leaving the area near the display, it will automatically turn off the voice content without interfering with the user's life.
  • the device senses that there are high-temperature items and people gathered in the restaurant during the dining time period, it will automatically open the dining mode, open Restaurant lights, light music for dining sounds, turn off lights in other rooms (perceive the user has left);
  • the device perceives that the user opens the entrance door from inside the house, and perceives that the guest (face recognition) enters the room, he actively cares for the guest "Dear guest, welcome to come to our house as a guest", and automatically opens the guest mode, For example, the displayed art picture or switch to the welcoming pine picture, after a custom N minutes, it will automatically switch to the precious and famous painting that the user likes.
  • the device senses that the guest passes the logical sensing range of 2 meters in front of the front of the device (that is, the 9 # and 10 # sensors are triggered in the positive and negative directions), it automatically switches the user's other precious famous painting to attract the attention of the guest and increase the communication topic between the host and the guest.
  • each different conference room installs different versions of smart TVs or flat robots (applications outside the home are collectively referred to as flat robots) and other devices according to the size of the space.
  • the user enters the reservation time through the mobile APP / PC / platform / local , Participants, duration, conference room number, conference theme, remote area and other related information.
  • 1 hour before the meeting a user arrives in the reserved meeting room.
  • the device determines the least two of the logical relationships of the meeting room to be determined through machine learning / user input / big data analysis / video analysis according to the spatial structure of each meeting room. More than one sensor logically senses to determine whether the user enters the meeting room.
  • the device When it is determined that the user enters the meeting room, it automatically starts the display screen and displays company-related advertisements or introductions, and actively cares for the user "colleagues, hello! This meeting room has 1 hour later Important meetings, please pay attention to the time "; 10 minutes before the meeting, the device automatically activates the air conditioning in the meeting room, and adjusts the temperature and humidity to an appropriate value. When the device senses that the user enters the meeting room, it actively cares for the user.” Welcome colleagues to participate in a certain meeting.
  • the device actively reminds all the colleagues of the meeting that "the meeting time is up, everyone is ready for the meeting", and displays the meeting report PPT or user computer investment
  • the host announced that "today's meeting is here, break up, ball, we are gone”
  • the device senses that the last user leaves the meeting room, automatically turns off the meeting room lights, and at the same time, the display is in standby.
  • Display mode different booths are installed with different sizes of smart TVs or flat robots (applications outside the home are collectively referred to as flat robots) and other devices according to the size of the space.
  • the user enters the appointment visit time through the mobile APP / PC / platform / local, Relevant information such as the visiting unit, the visiting staff, the interpreter, the accompanying staff, the duration of the tour, the flow chart, and the purpose of the tour.
  • the device uses the close-range human sensor / Bluetooth / fingerprint / face recognition to explain the personnel or the customer ’s face that has been entered to the display screen, and actively cares for "the leader, welcome you to the company In the exhibition hall visit, I mainly show the content of XX.
  • the content system that has not been visited remains silent by default until the interpreter arrives to avoid the sound of each device from interfering with the normal visit process ;
  • the device senses that the last visitor has passed the logical sensing range of 2 meters (for example, the trigger logic of the sensor combination such as 3 # / 6 # / 9 # / 10 #), taking active care Visitors "Relevant leaders of XX company, the main contents of today's visit are these, you are welcome to come to guide next time" Hall welcome visitors open the door, when the sensing means away from the last visitor sensing area, such as lighting or air conditioning automatically shut down.
  • Office mode each different office, corridor, front desk, open office area, etc., according to the requirements of space size / structure characteristics / content display / control management / environmental monitoring, etc., install different versions of smart TVs or flat robots (collectively referred to as applications outside the home) It is a plane robot) and other devices.
  • the device senses the logical sensing range of 2 meters (for example, the trigger logic of the sensor combination such as 3 # / 6 # / 9 # / 10 #), combined with video analysis review and quantization, it actively outputs voice and video. Audio, image or control, etc., to help solve the user's possible needs.
  • the device at the company senses and recognizes the user, automatically disarms and turns on the lights, curtains, and air conditioning of the card position and corridor where the user is located, and actively cares for the user.
  • Mr. Chen came to the company so early !
  • the device automatically turns on the background music to remind the user that it is close to work time, please prepare for work.
  • the device automatically plays background music to remind users that they have left work at noon, and turn off office area lights and curtains within 10 minutes after work, so that users can return to the company for lunch after lunch.
  • the device automatically turns on the background music to remind the user that it is close to work time, please prepare for work; in the afternoon after work, the device automatically turns off the lights in the area when it detects that there is no one in the area, until All personnel leave, turn off all office lights, curtains and air conditioners, and securely arm.
  • the device perceives (or recognizes the identity of the video) when the overtime employee passes by, and actively cares for the user "colleague (after the video is identified, the user's title is changed to" Chairman Chen "). Are you still working overtime?
  • Hotel mode Each different room is equipped with different sizes of smart TVs and other devices according to the size of the space.
  • the device senses that the guest opens the door, automatically opens the room lights, curtains (closed at night), air conditioning, etc., and actively cares for the guests.
  • XX guests welcome to the XX hotel, I am the butler of your hotel, if you have any needs, you can tell me directly. " When the device senses that the user is located within a range of nearly 2 meters of the TV, the TV is automatically turned on.
  • the device When the guest said to the device "Butler, I will get up at 7 o'clock tomorrow, going to a meeting, and now go to bed”, the device automatically replied “Okay, I am going to wake you up at 7 o'clock”, and the device senses that the guest is lying down Go to sleep, automatically start the sleep mode, turn off the lights and curtains, adjust the air conditioner to the appropriate temperature, and arm the hotel to the door / window side of the hotel for arming. If the door and window side sensor first senses the person, it will automatically alarm; if it senses the guest from the device The sensor corresponding to the bed side is triggered first, and the infrared array sensor is added to identify the user's status and actions to determine the user's needs as the night.
  • the device senses that the user lifts the quilt and sits up, it automatically turns on the night-up mode and turns on the dimmed lights of the customer and the toilet.
  • the night-up mode is automatically turned off; when 6:50 in the morning
  • the device senses that the guest wakes up, automatically plays the wake-up music, and actively cares for the user "Mr. Chen, you wake up at 7 o'clock, are you not sleeping well?" Or "Mr. Chen, it's 6 o'clock now 50 minutes, but you have a meeting in the morning, you can get up now "or" Mr.
  • Smart rural mode Rural users are generally independent houses, and there are more left-behind elderly and children.
  • a flat robot is installed in the living room of the entrance door, or a smart TV and other devices are installed in the user's master bedroom.
  • Device In addition, farmers are most concerned about the care and support of the elderly, and the situation of agricultural crops. Therefore, the device has a built-in LoRa long-distance wireless module, which can receive various sensor data in the field within a few KM nearby, so that farmers can control the field.
  • Real-time data monitoring and management for example: get up in the morning, the device senses that the user uncovers the quilt and gets up, and actively cares for the user "Master, good morning!
  • the technical solution of the present invention breaks the conventional home control methods (such as APP, panel, passive voice wake-up control, etc.), based on the trigger time of multiple human sensors, area confirmation trigger events, multi-trigger logic relationship, video analysis, indoor space structure, Node time, environmental characteristics, specific scenes, user gestures, emergency help, facial expressions, health data and other events, perform machine learning and big data analysis on user behavior, actions, movement speed, frequency, and trajectory to intelligently judge user needs and accurately give Users provide special services such as active care, interaction, inquiry, warning, and control, so that users can truly feel the warmth and wisdom of the home, which can increase user stickiness, solve user's emotional and attribution problems, and allow users to enjoy the wisdom system to bring users The affordable, fast, convenient, smart and comfortable user experience allows users to re-recognize smart systems, like smart systems, and think of more consumer smart systems, allowing users to truly feel the warmth and wisdom of their homes.

Abstract

本发明涉及了一种主动关怀的实现方法、系统及装置，该实现方法包括：实时获取各个人体传感器的感测数据；若预设时段内检测到至少两个相邻的人体传感器触发，则获取感测特征信息；将感测特征信息输入到用户需求模型中，并根据用户需求模型的输出结果确定用户的当前需求信息；根据用户的当前需求信息控制相应的家居装置和/或输出相应的第一提示信息。

Description

主动关怀的实现方法、系统及装置 技术领域
本发明涉及人工智能领域，尤其涉及一种主动关怀的实现方法、系统及装置。
 背景技术
智能家居是以住宅为平台，利用综合布线技术、网络通信技术、安全防范技术、自动控制技术、音视频技术将家居生活有关的设施集成，构建高效的住宅设施与家庭日程事务的管理系统，提升家居安全性、便利性、舒适性、艺术性，并实现环保节能的居住环境。
目前，智能家居系统中的智能装置（例如，智能电视、智能室内对讲分机装置）通过集成多网络、多系统、多传感器技术，使其产品自身能为用户提供安全保障、智慧互动、环境监测、远程医疗/教育、家居控制等多项功能。伴随着智能装置的核心硬件越来越强大，智能装置的控制方式也多种多样，例如包括：面板控制、APP软件控制、遥控器控制、语音控制、手势控制、传感器联动控制，但是，这些控制方式都具有控制缺陷，例如，对于面板控制方式，家中所需安装的面板（可视对讲面板、空调面板、报警键盘等）数量多，不美观，施工难度也大；对于APP软件控制方式，需要用户通过点击APP软件的分层目录去控制家居设备，操作不便，尤其是对于老人和小孩群体；对于遥控器控制方式，遥控器放置位置不固定，在需要时容易找不到；对于语音/手势控制，受限于语音/手势的识别距离及识别率，很难保证准确率；对于传感器联动控制方式，通常在室内特定位置安装人体传感器，但是一个人体传感器只能感测一个区域，且感测范围很有限，并且单一传感器无法判断用户的需求等，例如，当感测到有人经过的时候联动开启灯光，但这并不具备智能分析特性，即不知何时再关闭灯光，如果只是延时关闭的话，在延时的时间内用户可能并不希望关闭灯光，因此用户体验较差。
 技术问题
本发明要解决的技术问题在于，现有技术存在的上述操作复杂、准确率低、用户体验差缺陷。
 技术解决方案
本发明解决其技术问题所采用的技术方案是：构造一种主动关怀的实现方法，包括以下步骤：
S10.实时获取各个人体传感器的感测数据，其中，装置的外侧设置多个人体传感器，以在装置的外围形成大于180度的人体感测区域；
S20.若预设时段内检测到至少两个相邻的人体传感器触发，则对所述至少两个相邻的人体传感器的感测数据进行处理，以获取感测特征信息，所述感测特征信息包括人体传感器的触发时间、触发顺序及感测区域；
S30.获取当前工作模式，并获取预先建立的所述当前工作模式所对应的用户需求模型，而且，将所述感测特征信息输入到所述用户需求模型中，并根据所述用户需求模型的输出结果确定用户的当前需求信息；
S40.根据用户的当前需求信息控制相应的家居装置和/或输出相应的第一提示信息。
优选地，所述用户需求模型根据下列步骤建立：
在相应工作模式下训练时，获取训练样本数据，其中，所述训练样本数据包括预设时段内触发的至少两个相邻的人体传感器的感测数据；
对所述训练样本数据进行处理，以获取感测特征信息；
使用预设的训练方法，对所述感测特征信息和用户需求信息进行训练，以获得相应工作模式所对应的用户需求模型。
优选地，所述多个人体传感器包括红外阵列传感器；而且，
在所述步骤S20中，所触发的至少两个相邻的人体传感器包括所述红外阵列传感器，且所述感测特征信息还包括用户姿态信息。
优选地，装置的正面设置有摄像头，而且，所述方法还包括：
S50.获取所述摄像头的图像数据；
S60.对所述图像数据进行处理，以获取用户特性信息，所述用户特性信息包括：类型、表情、姿态、身份、着装、形象、数量、房屋空间特征；
S70.根据所述用户特性信息输出相应的第二提示信息。
本发明还构造一种主动关怀的实现装置，与多个家居装置通讯连接，所述实现装置包括主体及位于主体周围的边框，还包括：
安装在所述边框上的多个人体传感器，以在所述实现装置的外围形成大于180度的人体感测区域；
第一获取模块，用于实时获取各个人体传感器的感测数据；
第一处理模块，用于在预设时段内检测到至少两个相邻的人体传感器触发时，对所述至少两个相邻的人体传感器的感测数据进行处理，以获取感测特征信息，所述感测特征信息包括人体传感器的触发时间、触发顺序及感测区域；
第一确定模块，用于获取当前工作模式，并获取预先建立的所述当前工作模式所对应的用户需求模型，而且，将所述感测特征信息输入到所述用户需求模型中，并根据所述用户需求模型的输出结果确定用户的当前需求信息；
执行模块，用于根据用户的当前需求信息控制相应的家居装置和/或输出相应的第一提示信息。
优选地，还包括：
建模模块，用于在相应工作模式下训练时，获取训练样本数据，并对所述训练样本数据进行处理以获取感测特征信息，而且，使用预设的训练方法，对所述感测特征信息和用户需求信息进行训练，以获得相应工作模式所对应的用户需求模型，其中，所述训练样本数据包括预设时段内触发的至少两个相邻的人体传感器的感测数据。
优选地，所述多个人体传感器包括红外阵列传感器；而且，所触发的至少两个相邻的人体传感器包括所述红外阵列传感器，且所述感测特征信息还包括用户姿态信息。
优选地，还包括摄像头、第二获取模块和第二处理模块，而且，
所述第二获取模块，用于获取所述摄像头的图像数据；
所述第二处理模块，用于对所述图像数据进行处理，以获取用户特性信息，所述用户特性信息包括：类型、表情、姿态、身份、着装、形象、数量、房屋空间特征；
所述执行模块，还用于根据所述用户特性信息输出相应的第二提示信息。
优选地，所述多个人体传感器包括：
设置在所述边框的下框上的第一人体传感器；
分别设置在边框的左、右两侧框上的第二人体传感器，且所述第二人体传感器可沿相应侧框的竖直轴线转动，以调节感测方向。
本发明还构造一种主动关怀的实现系统，包括后台服务器、多个家居装置，其特征在于，还包括以上所述的主动关怀的实现装置。
 有益效果
实施本发明的技术方案，通过在装置上设置多个人体传感器，在应用时，根据多个人体传感器的感测数据，并结合预先建立的用户需求模型，可智慧判断出用户的真实需求，进而准确地给用户提供主动关怀服务，使用户得到方便、智能、舒适的用户体验。
 附图说明
图1是本发明主动关怀的实现方法实施例一的流程图；
图2是本发明主动关怀的实现装置的各人体传感器的感测区域的示意图；
图3A是本发明安装在住宅的实现装置的感测区域实施例一的示意图；
图3B是本发明安装在住宅的实现装置的感测区域实施例二的示意图；
图3C是本发明安装在住宅的实现装置的感测区域实施例三的示意图；
图4是本发明主动关怀的实现装置实施例一的逻辑结构图；
图5是本发明主动关怀的实现装置的立体结构示意图；
图6是图5中的主动关怀的实现装置的仰视方向局部示意图及底部传感器的感测角度示意图；
图7是图5中的主动关怀的实现装置的侧视方向局部示意图；
图8是图7中的侧面传感器的侧向示意图；
图9是图5中的主动关怀的实现装置的俯视方向局部示意图及侧面传感器的感测角度示意图。
 本发明的实施方式
图1是本发明主动关怀的实现方法实施例一的流程图，该实施例的方法可应用在智能电视、智能室内对讲分机装置、智能机器人装置、智能网关、智能路由器、单元门口机等装置中，且该方法包括以下步骤：
S10.实时获取各个人体传感器的感测数据，其中，装置的外侧设置多个人体传感器，以在装置的外围形成大于180度的人体感测区域；
在该步骤中，人体传感器例如为：红外线、超声波、微波、热释电元件传感器，也可为阵列传感器或激光传感器或雷达传感器等。在图2所示的一个例子中，在装置的边框上共设置了10个人体传感器，且根据人体传感器感测角度与准确度，不局限10个人体传感器数量，其中，左侧框底部设置了两个对侧框外围的空间进行感测的人体传感器，其探测区域分别为1、2，左侧框底部还设置了一个对左侧框下侧的空间进行感测的人体传感器，其探测区域为3；右侧框底部设置了两个对右侧框外围的空间进行感测的人体传感器，其探测区域分别为4、5，右侧框底部还设置了一个对右侧框下侧的空间进行感测的人体传感器，其探测区域为6；正面设置了两个对正面外围的空间进行感测的人体传感器，其探测区域分别为7、8，下框底部还设置了两个对下框下侧的空间进行感测的人体传感器，其探测区域分别为9、10。通过这十个人体传感器的设置，可以装置为中心形成大于180度的人体感测区域。
优选地，设置在左侧框、右侧框的人体传感器为角度可调人体传感器，即，其可沿装置的侧框的竖直轴线转动，以调节感测方向。另外，应理解，在安装该装置时，应结合现场（住宅/酒店/会议室/展厅）的空间结构，使所形成的180度至360度的人体感测区域尽可能多地覆盖用户经常出入或能体现最多需求的区域，例如，在如图3A所示的一个应用中，智能室内对讲分机装置安装在入户门处，且其左侧探测区域1、2、3覆盖厨房门、餐厅、客厅与厕所/主卧的过道；其右侧探测区域4、5、6覆盖卧室一的门、入户门；其正面探测区域7、8、9、10覆盖客厅、卧室二的门。智能电视安装在主卧的电视墙，且其左侧探测区域1、2、3覆盖厕所门、主卧门；其右侧探测区域4、5、6覆盖主卧窗户；其正面探测区域7、8、9、10覆盖主卧床。在如图3B所示的一个应用中，智能室内对讲分机装置安装在入户门处，且其左侧探测区域1、2、3覆盖入户门；其右侧探测区域4、5、6覆盖客厅；其正面探测区域7、8、9、10覆盖厨房门、餐厅。智能电视安装在主卧的电视墙上，且其探测区域的分布与图3A类似。在如图3C所示的一个应用中，智能室内对讲分机装置安装在入户门处，智能电视安装在主卧的电视墙上，智能室内对讲分机装置的各人体传感器的探测区域与图3B所不同的仅是：通过调节智能室内对讲分机装置的左侧人体传感器的特侧角度，是其探测区域1、2、3覆盖入户门、厨房门、餐厅；智能室内对讲分机装置的右侧探测区域4、5、6覆盖客厅的一部分；智能室内对讲分机装置的正面探测区域7、8、9、10覆盖客厅的另一部分及过道，其它与图3B类似。进一步地，当住宅中安装至少两个装置时，如上所示，两个装置分别为安装在玄关处的智能室内对讲分机装置及安装在主卧的智能电视，还可实现这两个装置的感测数据的交互，例如，用户从主卧出来进入客厅，所经过的人体传感器依次触发，从而可识别出用户从主卧到客厅完整的行走路线。当然，在其它的一些实施例中，也可在没有安装智能电视的卧室内安装一些人体传感器，并将这些人体传感器与智能室内对讲分机装置或智能电视进行连接，这样，例如，当用户从该卧室出来进入客厅或主卧，所经过的人体传感器依次触发，从而可识别出用户从该卧室到客厅或主卧完整的行走路线。
S20.若预设时段内检测到至少两个相邻的人体传感器触发，则对所述至少两个相邻的人体传感器的感测数据进行处理，以获取感测特征信息，所述感测特征信息包括人体传感器的触发时间、触发顺序及感测区域；
S30.获取当前工作模式，并获取预先建立的所述当前工作模式所对应的用户需求模型，而且，将所述感测特征信息输入到所述用户需求模型中，并根据所述用户需求模型的输出结果确定用户的当前需求信息；
S40.根据用户的当前需求信息控制相应的家居装置和/或输出相应的第一提示信息。
在该步骤中，家居装置例如包括智能灯、智能窗帘等。需求信息与家居装置的控制指令/第一提示信息之间的对应关系，可由用户预先设置并存储，也可由后台服务器通过大数据学习获取到后传送至相应的装置中，例如，需求信息为用户下班回家时，相对应的家居装置的控制为：控制玄关灯、客厅灯、厨房灯等开启；控制空调开启在26度。相对应的第一提示信息为：欢迎主人回家，等。需求信息为用户夜晚起夜时，相对应的家居装置的控制为：控制卫生间灯开启；反之，需求信息为上完卫生间继续睡觉时，相对应的家居装置的控制为：控制卫生间灯关闭。
在该实施例中，通过在装置的外侧设置多个传感器，形成180度至360度的人体感测区域，其中任意区域的人体传感器被触发，进入唤醒状态，开始计时，当用户在自定义的预设时间内触发相邻第二区时，说明用户从第一区已经移动至第二区域，还说明用户确实在这个区域，根据当前的工作模式及学习获取的用户需求模型，获取用户当前的需求信息，进而输出语音、视音频、图像等提示信息或对家居装置进行控制等。如果在预设时间内没有触发第二区，则说明用户只在第一区活动，或该区存在误报。
另外，结合图2，如果第一区触发后，0时间触发第二区时，根据户型空间结构，可以判断该区域或交叉区域确定感知到用户，通过机器学习与大数据分析，确认单传感区没有误报，也可以定义N时间为零，人体传感器触发即向系统报告当前状态，如布防后的对准没有信号干扰的入户门探测区域，单传感区域探测有人非法入侵，即向系统报告有非法入侵发生；如果水平传感在被触发的同时，装置下部左/右/前近距离传感器也被触发，说明用户在水平触发区域位置的距离装置2米以内范围；如果用户只触发左/中/右侧水平传感器，说明用户在距离装置2米以外的范围；夜晚布防时，只对入户门一侧的水平与下部的传感器进行布防，其它传感器则可用于实现用户起夜联动控制功能，并且加上机器学习与逻辑判断，用户从卧室侧传感起，连接触发多个传感器再触发布防区域时，系统仍然判断为正常起夜模式，而非触发报警；节假日或老人在家时，用户可能会因思考问题/无聊/做家务/玩耍等缘故反复徘徊，装置会根据用户的移动速度、频次、轨迹、时间特征、用户类型、表情、行为等进行机器学习与智慧分析判断，主动输出语音、视音频、图像或控制等，帮助解决用户可能存在的需求。
进一步地，用户需求模型根据下列步骤建立：
在相应工作模式下训练时，获取训练样本数据，其中，所述训练样本数据包括预设时段内触发的至少两个相邻的人体传感器的感测数据；
对所述训练样本数据进行处理，以获取感测特征信息；
使用预设的训练方法，对所述感测特征信息和用户需求信息进行训练，以获得相应工作模式所对应的用户需求模型。
在该实施例中，当装置安装好后，或，用户对人体传感器的探测角度重新做了调整后，需对用户需求模型做训练。在训练时，首先获取大量的训练样本数据，然后提取感测特征信息，再按照预设训练方法对感测特征信息及用户的实际需求信息进行训练，从而可获取用户需求模型。例如，在自学习阶段，大概率的事件为：若获取到用户在工作日的早七点左右经过玄关开门时先后触发玄关处的两个人体传感器，然后去上班；且通常在晚七点左右进门，然后经过玄关时先后触发玄关处的两个人体传感器（与早上出门时触发顺序反向）；还有，通常在晚上七点左右进入厨房并在厨房待一定时间（例如半小时），然后在感测到用户从厨房进入餐厅时便可开始吃饭；还有，在休闲时间，用户以相近速度、相同区域、间隔相同时间内反复徘徊；还有，用户多次只经过装置近距离传感器；还有，用户在同一区域长时间停留，等等，则可在用户需求模型中确定：若用户在早七点经过玄关并开门，则可确定用户的需求为早上上班；若用户在晚七点进门且经过玄关，则可确定用户的需求为晚上下班；若用户进入触发超预定时间且感测到用户从厨房进入餐厅，则可确定用户的需求为吃晚餐；若用户反复徘徊，则可判断用户在思考问题/无聊/健身；若多次只经过近距离传感器，则可判断为小孩或动物；若在同一区域长时间停留，则可能休息/生病/娱乐。
另外，当获取到用户需求模型后，还可进一步使用训练样本数据对用户需求模型进行验证，如果验证通过则可应用，如果验证通不过则需进一步采集训练样本数据，以继续做训练，直至获取到验证通过的用户需求模型。
进一步地，多个人体传感器包括一红外阵列传感器，优选地，将红外阵列传感器设置在装置的正面。而且，在步骤S20中，所触发的至少两个相邻的人体传感器包括该红外阵列传感器，当对该红外阵列传感器的感测数据进行处理后，可提取出的感测特征信息还包括用户姿态信息。例如，结合图3A至图3C，当智能电视安装在主卧时，通过该红外阵列传感器可获取的用户姿态信息包括：起床、躺下、翻身等；当智能室内对讲分机装置安装在客厅时，通过该红外阵列传感器可获取的用户姿态信息包括：跌倒、坐下、走动、躺下、移动方向、速度、频次、体温等。
进一步地，装置的正面设置有摄像头，而且，本发明的方法还包括：
S50.获取所述摄像头的图像数据；
S60.对所述图像数据进行处理，以获取用户特性信息，所述用户特性信息包括：类型、表情、姿态、身份、着装、形象、数量、房屋空间特征；
S70.根据所述用户特性信息输出相应的第二提示信息。
在该实施例中，当家庭中有多个成员时，例如包括老人、小孩、中年人，由于仅根据人体传感器的感测数据还是无法区分是哪个成员触发了人体传感器，进而无法准确获知用户的真实需求，所以该实施例通过结合人体传感器的感测数据及摄像头的图像数据，可准确分析出用户的真实需求，例如，同样是探测到用户早上七点出门，通过图像分析，若确定出该用户为老人，则确定该用户的当前需求为外出锻炼，进而可发出提示老人的语音提示；若确定出该用户为小孩，则确定该用户的当前需求为上学，进而可发出提示小孩的语音提示；若确定出该用户为中年人，则确定该用户的当前需求为上班，进而可发出提示中年人的语音提示。
进一步地，在建立用户需求模型时，训练样本数据也可加入图像特征信息。在用户需求模型训练好后，该用户需求模型的输入信息除了包含感测特征信息外，还包含图像特征信息，这样，该用户需求模型所输出的用户需求信息更准确、更智慧。
图4是本发明实现主动关怀的实现装置实施例一的逻辑结构图，该实现装置与多个家居装置（例如设置在卧室、卫生间、客厅、餐厅的智能灯、空调、智能窗帘等）通讯连接。而且，该实现装置包括主体（未示出）及位于主体周围的边框（未示出），还包括有：多个人体传感器（未示出）、第一获取模块10、第一处理模块20、第一确定模块30及执行模块40。其中，多个人体传感器安装在所述边框上的多个人体传感器，，以在所述实现装置的外围形成大于180度的人体感测区域；第一获取模块10用于实时获取各个人体传感器的感测数据；第一处理模块20用于在预设时段内检测到至少两个相邻的人体传感器触发时，对所述至少两个相邻的人体传感器的感测数据进行处理，以获取感测特征信息，所述感测特征信息包括人体传感器的触发时间、触发顺序及感测区域；第一确定模块30用于获取当前工作模式，并获取预先建立的所述当前工作模式所对应的用户需求模型，而且，将所述感测特征信息输入到所述用户需求模型中，并根据所述用户需求模型的输出结果确定用户的当前需求信息；执行模块40用于根据用户的当前需求信息控制相应的家居装置和/或输出相应的第一提示信息。
进一步地，本发明的实现装置还包括建模模块，该建模模块用于在相应工作模式下训练时，获取训练样本数据，并对所述训练样本数据进行处理以获取感测特征信息，而且，使用预设的训练方法，对所述感测特征信息和用户需求信息进行训练，以获得相应工作模式所对应的用户需求模型，其中，所述训练样本数据包括预设时段内触发的至少两个相邻的人体传感器的感测数据。
进一步地，多个人体传感器包括一红外阵列传感器；而且，所触发的至少两个相邻的人体传感器包括所述红外阵列传感器，且所述感测特征信息还包括用户姿态信息。
进一步地，本发明的实现装置还包括摄像头，而且，还包括有第二获取模块和第二处理模块，其中，第二获取模块用于获取所述摄像头的图像数据；第二处理模块用于对所述图像数据进行处理，以获取用户特性信息，所述用户特性信息包括：类型、表情、姿态、身份、着装、形象、数量、房屋空间特征；执行模块还用于根据所述用户特性信息输出相应的第二提示信息。
进一步地，多个人体传感器包括：
设置在所述边框的下框上的第一人体传感器，优选地，第一人体传感器设置在下框底部；
分别设置在边框的左、右两侧框上的第二人体传感器，且所述第二人体传感器可沿相应侧框的竖直轴线转动，以调节感测方向。优选地，第二人体传感器设置在相应侧框的底部。
在一个具体实施例中，结合图5-图9，该实施例的实现装置包括主体1以及位于主体1周围的边框2，边框2到主体1之间的四周预留一定宽度的空白区域3，便于无线信号正常发射与接收数据，还便于内部元器件向外部散热，也便于集成与用户的互动元器件（如触控按键），还便于安装其它传感器（如摄像头、红外阵列传感器、麦克风阵列等）。边框2采用空心设计，和主体1与凹限支架区结合部分以外的区域一起，形成内部元部件安装的整个空间。平面支架4主要通过可拆卸连接轴连接，因此只需要安装好墙面平面支架4，插上连接轴即可形成支架与本装置的牢固连接，简单易安装。实现装置的背部平面中间部位的凹限区下侧面与左侧面分别设计有与外部连接的接口通道41，接口通道41包括装置接口或出线槽孔，是装置内部部件与外部线缆或接口连接的通道，这样装置紧贴墙壁安装后，所有线缆全部隐蔽在装置背部凹限区内，凹限区空间加上墙壁洞孔空间，完全可以容纳装置需与外部连接的所有线缆或接口，即美观，也便安装、维护。可旋转设计主要是凹限区中部突出的圆形突出轴心5，当用户需要旋转主体1增加用户体验时，圆形突出轴心5主要提供旋转连接，还起到固定装置端支架的作用，突出平面即可螺丝固定轴盖板，也可以焊接固定轴盖板，突出空间可以足够容纳固定螺丝的螺杆突出部分。通常，轴心5上套设有与轴心5转动连接的背板，利用背板作为平面转动轴，可减少平面轴的1/3的厚度，同时也简化平面轴的结构实现。
另外，边框2的下侧两角落分别设有人体传感器6，每一人体感应装置6对所在角落对应的边框2的下框21下侧的空间以及边框2的侧框22外围的空间进行感测。下侧两角落的人传感器6结合正面的人体传感器，可以实现正面、下框21下侧、侧框22外侧等多个空间的人体感测。每一人体传感器6包括底部传感器61、侧面传感器62，底部传感器61在相应的角落底部设置，对所在角落对应的下框21下侧的空间感测，侧面传感器62在相应角落侧面设置，对所在角落对应的侧框22外围的空间感测。在其它实施例中，边框2的下部两角可为圆角结构。
进一步地，结合图6，底部传感器61包括两个第一人体传感器611，且其中一个第一人体传感器611的感应方向向主体1的前侧倾斜，另一个第一人体传感器611的感应方向向所在角落对应的边框2的侧框22外侧倾斜。另外，因实现装置紧贴墙安装，而一般人体传感器的探测角度为60度，为了让第一人体传感器611的辐射范围不受干扰，尽量最大范围的感测，下框21表面设有供第一人体传感器611安装的凸出结构211，可以让两个第一人体传感器611中的一个向下框21的下方靠平面智慧机器人的正面方向倾斜，另一个向能向所在角落对应的所述边框2的侧框22外侧倾斜，且倾斜后不会受到遮挡等影响。
另外，在图6中，边框2是空心梯形结构设计，为将装置打造成精美艺术器，设计将装置边框2梯形短边侧与长边侧设计带卡扣结构，在此梯形斜面上安装卡扣式套边7，用户即可以根据室内装饰风格自行更换套边7的颜色、图案及风格样式等，也可以提升装置整体艺术感的效果，还给装置预留更大的内部部件安装空间。
进一步地，结合图7-9，侧面传感器62包括感测方向水平设置的两个第二人体传感器621，两个第二人体传感器621的感测方向向所在侧框22的外侧设置，且两个第二人体传感器621的感测方向呈夹角设置，增加感测范围。其中一个第二人体传感器621的感测范围的一边与主体1的背面平行，另一第二人体传感器621则向主体1的正面一侧摆动，增加感测范围。
两个第二人体传感器621上下并排设置，优选地，两个第二人体传感器621可沿竖直轴线转动，以调节感测方向。在其他实施例中，也可只有一个第二人体传感器621可以转动。
边框2的侧框22外侧设有向外凸起的半圆形凸台221，所述凸台221的弧面的轴线竖直设置。两个所述第二人体传感器621可转动地设置在所述凸台221内，所述凸台221上设有分别供两个第二人体传感器621向外感测的两个弧形感测口222，所述弧形感测口222水平设置，在所述第二人体传感器621转动时沿所述弧形感测口222摆动，实现感测范围的调节。优选地，凸台221上设有两个可转动地转动旋钮223，两个所述转动旋钮223与两个第二人体传感器621分别连接，以分别带动所述第二人体传感器621转动调节所述第二人体传感器621的感测方向。
第二人体传感器621探测角度水平方向可调，当平面智慧机器人安装在临直角墙边时，加上正面空间区域的传感器，可以形成水平180-270度的人体探测范围，如果平面智慧机器人两侧均临直角墙边，还可以形成水平360度的探测范围，因此适合任意户型空间结构的探测感知。半圆形凸台221结构设计，节约了平面智慧机器人内部空间，也保证传感器角度调整需要，保证了传感器的探测范围达到最达值。
最后，本发明还构造一种实现对主动关怀的实现系统，包括后台服务器、多个家居装置、以及上述实施例所示的实现装置。
下面结合一些具体应用来说明本发明的实现装置（例如名字为小球，且以下简称装置）是如何实现主动关怀的：
用户离家：装置感知到用户早上七点从入户门侧的过道经过（例如从2#感测区域至1#感测区域），或装置获取到用户出门时的语音信息“小球小球，我上班去了，帮我看好家”，便可获取用户的当前需求为离家，进而自动回复“主人一路平安！”。若此时装置感知用户的部分涉隐传感器关闭，还会发出语音提示“出门前请您打开我的眼睛（摄像头）/耳朵（麦克风）感知，我好帮您看好家”，当用户靠近装置打开涉隐传感器，例如打开了摄像头，装置便可通过视频识别获取用户当前形象/表情/衣着/体温等状态，根据识别的结果主动语音关怀用户：“主人，您今天出门好像没有理胡须哟”或“主人，今天工作日您不需要穿正装吗？”或“主人，我感知您体温有点高，是不是那里不舒服？”或“主人，我看您精神不太好，您昨晚太夜睡觉好，请多注意身体”等，用户可选择回复，然后关门离开，装置自动记录用户回复出门时的个人状态，以便回家时根据需要关怀用户，用户离家后一定时间内，系统自动进入白天布防状态。而且，在布防状态下，室内特定单个人体传感器或任意相邻多个传感器在一定时间内先后或同时被触发时，系统都可判断为非法入侵报警。
用户回家：装置感知到用户下午七点从入户门侧的过道经过（例如从1#感测区域至2#感测区域），或装置获取到用户回家时的语音信息“小球小球，我下班了”，便可获取用户的当前需求为下班回家，进而自动回复“主人，欢迎主人回家，请撤防”。当用户在一定时间内通过指纹识别或密码或视频识别或声纹识别等解除安全布防状态，系统自动开启回家模式，装置自动开启客厅灯光、窗帘及空调、艺术屏展示等设备，并在通过视频识别获取到用户身份/表情/体温等特征，针对不同的用户或不同表情或不同体温或不同的时间段，以不同的语义语音主动关怀用户，如：下午放学时间感知小主人回家，装置主动童声问候“小主人，今天上学开心吗？今天上了什么课程？”，小主人回复“开心，今天上了英文、数据、体育课”，装置会回复“那我们一起温习一下今天学习的英文单词吧！”，这样小朋友一回来就可以跟装置一起开始学习了；下班时间主人回家，装置主动问候“主人，看您样子很开心，要不要来点音乐”或“主人，早上出门时您的体温有点高，现在是不是好一点了？”或“主人，小主人今天放学我与他温习了50个英文单词，这些单词他都能拼读了”或“主人，男主人晚上有饭局，不在家吃饭”等。另外，当装置感知用户打开入户门，自动语音问候，感知用户不但不撤防，而且感知用户直接走向室内其它区域，一定时间内系统自动报警，并将现场视音情况自动上传至物业中心或主人手机或服务平台。
用户早上起床：例如，早上用户一般定时7点起床，但今天临时有事6点40就起床，而用户的妻子或老公却要8点起床，卧室内的装置感知用户揭开被子，坐或站起来时，自动卧室开启暗灯光；如果用户单身或同时起床，则装置感知用户起床，自动开启背景音乐，打开窗帘，开启艺术品或清晨风景画展示，并主动关怀用户“主人，早上好，昨天晚上室内空气质量全过程各项指标没有超标，您应该睡得还好吧？”或“主人，早上好，昨晚您去了几次厕所，是不是肠胃不好，要多注意身体才是”或“主人，早上好，昨晚3点室内CO2超标，但卧室没有新风机，所以我没能帮您调节室内空气质量，您需要打开窗户睡觉或安装新风机以便我能主动调节室内空气质量”等。如果卧室没有安装装置，当客厅的装置确定出用户起床打开卧室门，便自动开启背景音乐，打开窗帘，开启艺术品或清晨风景画展示，并根据不同区域传感器对应探测的不同卧室的不同用户，针对不同用户或不同时间段，以不同的语义和语音主动关怀用户“主人，早上好，昨晚您去了几次厕所，是不是肠胃不好，要多注意身体才是”或“主人，早上好，昨晚您太晚睡觉，现在又这么早起床，要多注意身体”或“主人，早上好，昨晚客厅空间甲醛指标超标，我已为您开启新风或窗户调节空气（没有新风机或电动窗时，提醒用户安装）”或“老人家，早上好，今天室外气温12度，您要多注意保暖”或“小朋友，早上好，今天是第十天按时早起，值得表扬，继续保持”。当装置感知用户从卧室走向洗手间，自动开启洗手间灯光及排风机，方便用户洗脸擦牙上厕所。当装置感知用户洗刷完毕出来后，洗手间灯光及排风机自动关闭。
用户晚上睡觉：如果卧室没有安装装置，客厅的装置感测到用户在平时的睡觉时间进入卧室，或者，获取到用户的语音消息“小球小球，我睡觉了”，便可确定用户的当前需求为晚上睡觉，并输出“晚安”，然后，自动开启卧室灯光、关闭窗帘、空调调整到合适温度。同时，一定时间内，客厅灯光自动关闭，自动关闭装置的展示屏，进入布防。如果卧室已安装装置，该装置感测到用户在平时的睡觉时间上传、盖被子，或者，获取到用户的语音消息“小球小球，我睡觉了”，也可确定用户的当前需求为晚上睡觉，并输出语音消息“主人，晚安，现在卧室温度24度，湿度50%，您需要盖空调被睡觉”，然后自动关闭卧室灯光、关闭窗帘、空调调整到合适温度。
用户晚上起夜：如果卧室已安装装置，当该装置感知到用户揭开被子，坐或站起来时，系统自动开启卧室暗灯光，当用户打开卧室门时，自动联动开启客厅或走廊暗灯光。当装置通过传感器感知用户从卧室侧到厕所侧时，相关走廊和厕所灯光亮起。当装置通过传感器感知用户反向从厕所至卧室侧时，相关走廊和厕所灯光自动关闭。用户上床睡觉后，自动关闭卧室暗灯光。装置通过识别用户的需求来控制灯的开关，既节能也卫生又能保护隐私，还免除语音控制时干扰亲人休息，尤其是晚上睡觉时，也免除手动开关的麻烦等。如果卧室没有安装装置，根据户型结构特点，用户打开卧室门，便自动感知并开启起夜模式，客厅暗灯光开启，当装置通过传感器感知用户从卧室侧到厕所侧时，相关走廊和厕所灯光亮起。当装置通过传感器感知用户反向从厕所至卧室侧时，相关走廊和厕所灯光自动关闭。
白天布防状态：根据户型结构、可能存在的信号干扰、门窗位置等信息特点，装置可定义单个传感器、双传感器、多传感器以及传感器的触发逻辑来感知人体，若感知存在非法入侵时，系统自动现场语音报警及推送报警信息至物业和智慧家庭服务平台。
晚上布防状态：晚上布防主要对探测区域内的入户门、重要门窗侧传感器进行布防，并定义从用户卧室侧传感器为第一触发的逻辑感知为起夜模式，并且从卧室至厕所的全过程传感顺序逻辑定义为开灯模式，从厕所至卧室反向传感逻辑触发时定义为关灯模式，从厕所到卧室的暗光逐步关闭。在起夜模式下，如用户去厕所过程中需经过已布防的区域传感器，装置仍判断为起夜模式，但如果入户门侧传感器为第一感知用户，则系统认为为非法入侵。
老人在家：装置感知到室内只有老人在家时，例如，如老人长时间地待在一个区域，如在客厅沙发，装置判断超过自定义时间阀值时，主动关怀“老人家，您已经在沙发上连续坐了3小时了，您身体不好，不适合长时间不动”或“老人家，你看书已经超过3小时了，来点音乐一起放松一下吧”或“老人家，有一首您喜欢的广场舞曲要不要一起互动一下？”，如语音没有得到老人的回复，装置自动开启摄像机、麦克风，视频分析用户当前状态，如果老人状态出现异常，便自动开启紧急状态音乐，即提醒老人如果2分钟内不回复，即形成报警信号，如果老人听力有问题，也可闪烁灯光来提醒用户，超过2分钟时，系统自动将老人当前紧急状态推送给物业或亲属或服务平台。如老人在装置不能感知的房间长时间持续停留，而非正常休息时间，便在老人的子女回家时，第一时间语音告知用户“主人，您父亲/母亲N次在自己某某房室连续长时间停留，这样不利于健康，您可以在房间安装一个我的扩展版机器人，这样我就可以为您照顾老人了”或“装置在客厅播放紧急音乐，老人家听到音乐知道装置在提醒他活动身体而出来走动”或听力障碍的老人，装置自动拔打老人电话或闪烁室内灯光来提醒老人出来走动，如语音/电话/音乐/闪灯均没有得到老人的回复，系统自动将老人当前紧急状态推送给物业或亲属或服务平台。另外，若老人用户每天11点左右需要服药，则当装置在上午11点主动关怀老人“主人，您肠胃不好，某某肠胃药需要饭前1小时服用，请您现场服药”。假如N分钟（如十分钟）之内，如果装置感知不到用户靠近药柜取药、吃药，则可再次主动关怀老人“主人，刚才肠胃药吃过了吗？”。
紧急按钮触发：当紧急按钮触发时，自动开启摄像头现场识别老人当前状态，或老人在装置侧方感知区域，立即分析传感器感知用户最近N分钟是否都在报警区域，如果是说明场老人可能是疾病引起紧急状态，如果不是说明老人可能是跌倒，则立即向物业或亲属或服务平台报警，并上报感知用户的相关数据及老人的健康数据。
老人空闲：装置感测到老人徘徊N次（频次/速度/方向/远近距离/区域/老人习惯/时间节点等综合来判断是为徘徊），便可主动关怀老人“主人，您是不是遇到不开心事了？”或“主人，您在家徘徊走动空间有限，您可以去外面走动一下，带上您的定位表，好让我随时知道你在那里”或“主人，最近有一幅珍品书法作品，您可以过来看一下，它是谁于N年创造，当时的背景是……”或“主人，您上次没有听完的评书，要不要接着听一下？”。通过根据老人当前状况，主动关怀并与老人交流互动，让老人在休闲的时候做自己想做的事，缓解老人在家独单无聊，缺少关爱与交流。
休闲娱乐：装置感知客厅有用户或多个用户或最后一个用户从餐厅或卧室走到客厅，同时时间为某连续局开始的8点钟，便主动关怀“主人，昨晚看的某某电视局现在已经开始了，要不要一起看呀？”，若用户回复“好呀”，便自动开启影音模式，自动关闭餐厅、厕所、卧室的灯光，调暗客厅灯光。当客厅小孩打闹声音太大时，装置主动关怀用户“小朋友，夜晚请不要大声喧哗，不要影响他人休息”。当有人临时起去厕所时，装置感知用户从客厅到厕所的感知逻辑，自动开启走廊和厕所灯光，当用户从厕所出来再去客厅看电视时，装置感知用户从厕所到客厅的感知逻辑，自动关闭厕所及走廊的灯光。在休闲时间段，装置感知用户正反向触发9#与10#传感器时，自动切换艺术展示内容，并可设置伴随合适的语音音量介绍该艺术品的来由、作者、特点、历史故事等内容，当装置感知用户无兴趣离开展示附近区域，则自动关闭语音内容，不给用户生活带来干扰。通过这样不经意、碎片化的时间中感染、影响与提升用户的艺术素养与知识，变被动为主动信息服务，充实用户业余生活。用户可能会因思考问题/无聊/做家务/玩耍等缘故反复徘徊，装置会根据用户的移动速度、频次、轨迹、时间特征、用户类型、表情、行为等进行机器学习与智慧分析判断，主动关怀用户“主人，您喜欢的某某歌星出了一张新的MTV，要不要一起欣赏一下？”或“主人，您已经在N分钟内，反复徘徊N次了，您是不是有什么不开心事，能不能跟我诉说一下？”或装置感知用户反复徘徊达到一定阀值时，自动开启摄像机现场视频分析用户当前表情/行为/动作等，主动关怀用户“主人，您已经在N分钟内，反复徘徊N次了，看您脸色很不好，您是不是那里不舒服？”。
用户就餐：假如用户平时的就餐时间为7点/12点/18点，在机器学习分后，如果装置感知到在就餐时间段内有高温物品及人员聚集在餐厅时，自动开启就餐模式，打开餐厅灯光，就餐轻音乐响起，关闭其它房间灯光（感知用户已离开）；
用户会客：当装置感知用户从屋内打开入户门，并感知到客人（人脸识别）进入室内，主动关怀客人“尊贵的某某客人，欢迎来我家做客”，同时自动开启会客模式，例如，展示的艺术图或切换到迎客松图面，自定义N分钟后，自动切换成用户喜欢的珍贵名画。当装置感知到客人经过装置正面前2米逻辑感知范围时(即正反向触发9#与10#传感器)，自动切换用户其它一张珍贵名画，吸引客人注意，增加主人与客人的沟通话题，同时也展示主人的兴趣爱好。当拜访限时时间段左右，且装置感知用户打开入户门，便自动关怀用户“贵客是不是要走了？欢迎下次再来”，同时，自动呼叫电梯到用户楼层接客人，单元大厅感知有客人从电梯出来，自动播放欢送音乐歌/曲。
用户会议：每个不同的会议室根据空间大小安装大小不同版本的智能电视或平面机器人(家庭以外应用统称为平面机器人)等装置，用户通过移动APP端/PC端/平台/本地输入预约会议时间、参会人员、时长、第几号会议室、会议主题、远程区域等相关信息。会议1小时前，有用户到已被预约会议室，装置根据每个会议室的空间结构，通过机器学习/用户输入/大数据分析/视频分析待确定的会议室感知的逻辑关系，确定最少二个以上传感器逻辑感知来判断用户是否进入会议室，当判断用户进入会议室时，自动启动显示屏并显示公司相关广告或介绍，且主动关怀用户“同事，你好！这个会议室1小时后有重要会议，请注意把握时间”；会议10分钟前，装置自动启动会议室空调，并将温湿度调到合适值，当装置感知用户进入会议室时，主动关怀用户“欢迎同事参加某某会议，请到我这里签到（例如，通过右下角指纹签到或视频识别签到）”；会议时间到，装置主动提醒所有参会同事“会议时间到，大家准备开会”，并显示会议汇报PPT或用户电脑投屏内容，主持人回复装置“好的，小球，请帮我把PPT切换到第2页”，同事们我们今天的会议……，主持人语音“小球，下一页或上一页”，装置自动帮用户PPT切换到相应展示页面。会议结束，主持人宣布“今天的会议到就到这，散会，小球，我们走了”，装置感知最后一个用户离开会议室，自动关闭会议室灯光，同时，显示屏处于待机状态。
展示模式：每个不同的展位根据空间大小安装大小不同版本的智能电视或平面机器人(家庭以外应用统称为平面机器人)等装置，用户通过移动APP端/PC端/平台/本地输入预约参观时间、参观单位、参观人员、讲解人员、陪同人员、参观时长、流程图、参观目的等相关信息。当参观客户到达展示厅时，装置通过近距离人体传感器/蓝牙/指纹/人脸识别感知讲解人员或已输入的客户人脸到那个展示屏，主动关怀“某某领导，欢迎您到某某公司展厅参观，我主要展示的是某某内容，如果你想了解更多可以向我提问，我帮您解答”或“欢迎参观某某板块内容，我公司在本板块主要……”，装置根据讲解人员的移动路径判断参观人员移动路径，已参观过的内容后续参观者可以随意提问互动，未参观过的内容系统默认保持静默状态，直到讲解人员到达，避免每个装置都发出声音干扰正常参观流程；当参观人群参观完最后一个显示屏离开展厅时，装置感知最后参观者经过2米逻辑感知范围时(例如，3#/6#/9#/10#等传感器组合的触发逻辑)，主动关怀参观者“某某公司相关领导，今天参观的主要内容就这些，欢迎您下次莅临指导”，装置自动打开展厅大门欢送参观者，当装置感知最后参观者离开感知区域后，自动关闭灯光或空调等。
办公模式：每个不同的办公室、走廊、前台、开放办公区域等根据空间大小/结构特点/内容展示/控制管理/环境监控等需求，安装大小不同版本的智能电视或平面机器人(家庭以外应用统称为平面机器人)等装置，装置感知经过2米逻辑感知范围时 (例如，3#/6#/9#/10#等传感器组合的触发逻辑)，结合视频分析复核与量化，主动输出语音、视音频、图像或控制等，帮助解决用户可能存在的需求。在早上上班时间，公司前台或大厅的装置感知与识别用户，自动撤防并开启该用户所在卡位及走廊的灯光、窗帘及空调，并主动关怀用户“早上好！陈总这么早就来公司了!”或上班前10分钟，装置自动开启背景音乐，提醒用户离上班时间已经很近，请大家做好上班准备。在中午休息时间、下班时间，装置自动播放背景音乐，提醒用户中午已经下班，并在下班后10分钟内关闭办公区灯光、窗帘，以便用户吃完中午饭后回公司午休。在下午上班前5分钟，装置自动开启背景音乐，提醒用户离上班时间已经很近，请大家做好上班准备；在下午下班时间，装置感知到某区域没有人便自动关闭该区域的灯光，直至所有人员离开，关闭所有办公室灯光、窗帘与空调，并安防布防。在加班时间，例如晚上8点，装置感知（或视频识别身份）加班人员经过时，主动关怀用户“同事（视频识别身份后改称用户职称‘陈总’），您还在加班呢?感谢您为公司所做的付出”或晚上10点感知（或视频识别身份）加班人员经过装置时，主动关怀用户“同事（视频识别身份后改称用户职称‘陈总’），您应该回去休息了，多注意身体，工作可以明天做”等，对不同加班时间段，定义不同的问候语，让员工感受公司的温暖，增强员工的归属感与责任感。
酒店模式：每个不同的客房根据空间大小安装大小不同版本的智能电视等装置，该装置感知到客人开门，自动打开客房灯光、窗帘（晚上为关闭）、空调等，并主动关怀客人“尊贵的某某客人，欢迎入住某某酒店，我是您酒店的管家，您有什么需求可以直接跟我讲”。当装置感知到用户位于电视近2米范围时，自动开启电视。当客人睡觉前跟装置讲“管家，我明天7点起床，要去开个会，现在睡觉了”，装置自动回复“好的，某总我准备7点叫您起床”，装置感知客人躺下睡觉，自动启动睡觉模式，关闭灯光、窗帘，将空调调至合适温度，对准酒店入户门/窗侧传感器进行布防状态，如果门窗侧传感器首先感知到人，自动报警；如果感知客人从装置对应床侧的传感器先触发，并加上红外阵列传感器识别用户状态及动作，确定用户的需求为起夜。另外，若装置感知用户揭开被子坐起来，自动开启起夜模式，开启客户与厕所的暗灯光，当用户反向触发传感逻辑回到床上后，起夜模式自动关闭；当早上6点50分，客人揭开被子坐起来，装置感知客人起床，自动播放起床音乐，并主动关怀用户“陈总，您定时7点起床，是不是您没有睡好？”或“陈总，现在时间6点50分，不过您早上有会要开，您现在可以起床了”或“陈总，昨天晚上睡觉还好吧，您上了3次厕所，是不是您肠胃不好，多注意身体！”或用户没有按时起床，此时已经8点30分，装置根据昨天客人12点睡觉，自动播放起床音乐，并主动关怀用户“陈总，我们酒店的早餐用餐时间是7点至9点，现在起床还能赶上最后的就餐时间”。在上午10点，装置若感知用户在客房反复徘徊达到一定阀值，主动关怀客人“陈总，您已经徘徊N次，是不是有什么不开心的事？”或“陈总，您已经徘徊N次，要不要来点轻音乐或电视放松一下？”，用户回复“好呀！”，装置自动开始播放音乐或视频，并可询问客人想看什么想听什么，并满足用户需求输出相关内容。在中午11点30分，装置若感知客人在客房或经过装置2米范围时，主动关怀客人“陈总，我们酒店最迟退房时间是12点，您需要续房吗？”或感知客人频繁在各传感区域移动，之后只在客房入户门方向停留，加上11点30分左右时间，装置判断客人可能准备退房，自动关怀客人“陈总，您是不是准备退房了”或用户直接语音告知装置“管家，我走了”，装置自动回复“一路平安，欢迎下次再次入住，我现在安排服务员查房”。
智慧农村模式：农村用户一般都是独立的房屋，留守老人与儿童比较多，一般在入户门客厅安装平面机器人或用户主卧室安装智能电视等装置。装置。另外，农民最多关注的是老人的关爱、陪护等，以及农产作物的情况，因此，装置内置LoRa远距离无线模块，可以接收附近几KM内的田里各种传感数据，以便农户对田间数据实时监控与管理，例如：早上起床，装置感知用户揭开被子起床，主动关怀用户“主人，早上好！您家7块水稻田，现成土壤表皮含水量只有N%，最迟明天必须补水”或“主人，今天要下大雨，您今天应该降低稻田水位，以免水多涝灾”。对于留守儿童：当儿童放学时间回到家，经过装置2米感知范围或声音识别或视频辅助识别时，主动关怀儿童“小明，回来了！今天在学校开心吗？”或“小明，回来了！今天学校上什么课程呢？”或“小明，回来了！爷爷/奶奶在外面做农活，你在家先做好功课吧！”或“小明，回来了！昨天我们一起学习的英文单词我们一起温习一下吧”等。早上装置定时叫醒儿童“小明，现在已经6点了，该起床了”，同时播放起床音乐，打开窗帘；如果感知儿童赖床超过N分钟，再次主动叫醒儿童“小明，你已经连续10天准时起床了，不能因为这次赖床，打破你的记录哟”；早上上学时间，儿童主动与装置道别或装置感知儿童出门，装置主动关怀儿童“小明，该出门了，不然要迟到了，一路要注意安全”。
本发明的技术方案打破了常规的家居控制方式（如APP、面板、被动语音唤醒控制等），根据多人体传感器的触发时间、区域确认触发事件、多触发逻辑关系、视频分析、室内空间结构、节点时间、环境特征、特定场景、用户姿态、紧急求助、表情、健康数据等事件，对用户行为、动作、移动速度、频次、轨迹进行机器学习与大数据分析，智慧判断用户需求，准确地给用户提供主动关怀、交互、询问、警告、控制等特色服务，让用户真正感受家的温暖、家的智慧，可增加用户黏性，解决用户的情感与归属问题，让用户享受智慧系统给用户带来的实惠、快捷、方便、智能、舒适的用户体验，让用户重新认识智慧系统、喜欢智慧系统、想更多的消费智慧系统，让用户真正感受家的温暖，家的智慧。
以上仅为本发明的优选实施例而已，并不用于限制本发明，对于本领域的技术人员来说，本发明可以有各种更改和变化。凡在本发明的精神和原则之内，所作的任何纂改、等同替换、改进等，均应包含在本发明的权利要求范围之内。

Claims (10)

  1. 一种主动关怀的实现方法，其特征在于，包括以下步骤：
    S10.实时获取各个人体传感器的感测数据，其中，装置的外侧设置多个人体传感器，以在装置的外围形成大于180度的人体感测区域；
    S20.若预设时段内检测到至少两个相邻的人体传感器触发，则对所述至少两个相邻的人体传感器的感测数据进行处理，以获取感测特征信息，所述感测特征信息包括人体传感器的触发时间、触发顺序及感测区域；
    S30.获取当前工作模式，并获取预先建立的所述当前工作模式所对应的用户需求模型，而且，将所述感测特征信息输入到所述用户需求模型中，并根据所述用户需求模型的输出结果确定用户的当前需求信息；
    S40.根据用户的当前需求信息控制相应的家居装置和/或输出相应的第一提示信息。
  2. 根据权利要求1所述的主动关怀的实现方法，其特征在于，所述用户需求模型根据下列步骤建立：
    在相应工作模式下训练时，获取训练样本数据，其中，所述训练样本数据包括预设时段内触发的至少两个相邻的人体传感器的感测数据；
    对所述训练样本数据进行处理，以获取感测特征信息；
    使用预设的训练方法，对所述感测特征信息和用户需求信息进行训练，以获得相应工作模式所对应的用户需求模型。
  3. 根据权利要求1所述的主动关怀的实现方法，其特征在于，所述多个人体传感器包括红外阵列传感器；而且，
    在所述步骤S20中，所触发的至少两个相邻的人体传感器包括所述红外阵列传感器，且所述感测特征信息还包括用户姿态信息。
  4. 根据权利要求1所述的主动关怀的实现方法，其特征在于，装置的正面设置有摄像头，而且，所述方法还包括：
    S50.获取所述摄像头的图像数据；
    S60.对所述图像数据进行处理，以获取用户特性信息，所述用户特性信息包括：类型、表情、姿态、身份、着装、形象、数量、房屋空间特征；
    S70.根据所述用户特性信息输出相应的第二提示信息。
  5. 一种主动关怀的实现装置，与多个家居装置通讯连接，所述实现装置包括主体及位于主体周围的边框，其特征在于，还包括：
    安装在所述边框上的多个人体传感器，，以在所述实现装置的外围形成大于180度的人体感测区域；
    第一获取模块，用于实时获取各个人体传感器的感测数据；
    第一处理模块，用于在预设时段内检测到至少两个相邻的人体传感器触发时，对所述至少两个相邻的人体传感器的感测数据进行处理，以获取感测特征信息，所述感测特征信息包括人体传感器的触发时间、触发顺序及感测区域；
    第一确定模块，用于获取当前工作模式，并获取预先建立的所述当前工作模式所对应的用户需求模型，而且，将所述感测特征信息输入到所述用户需求模型中，并根据所述用户需求模型的输出结果确定用户的当前需求信息；
    执行模块，用于根据用户的当前需求信息控制相应的家居装置和/或输出相应的第一提示信息。
  6. 根据权利要求5所述的主动关怀的实现装置，其特征在于，还包括：
    建模模块，用于在相应工作模式下训练时，获取训练样本数据，并对所述训练样本数据进行处理以获取感测特征信息，而且，使用预设的训练方法，对所述感测特征信息和用户需求信息进行训练，以获得相应工作模式所对应的用户需求模型，其中，所述训练样本数据包括预设时段内触发的至少两个相邻的人体传感器的感测数据。
  7. 根据权利要求5所述的主动关怀的实现装置，其特征在于，所述多个人体传感器包括红外阵列传感器；而且，所触发的至少两个相邻的人体传感器包括所述红外阵列传感器，且所述感测特征信息还包括用户姿态信息。
  8. 根据权利要求5所述的主动关怀的实现装置，其特征在于，还包括摄像头、第二获取模块和第二处理模块，而且，
    所述第二获取模块，用于获取所述摄像头的图像数据；
    所述第二处理模块，用于对所述图像数据进行处理，以获取用户特性信息，所述用户特性信息包括：类型、表情、姿态、身份、着装、形象、数量、房屋空间特征；
    所述执行模块，还用于根据所述用户特性信息输出相应的第二提示信息。
  9. 根据权利要求5所述的主动关怀的实现装置，其特征在于，所述多个人体传感器包括：
    设置在所述边框的下框上的第一人体传感器；
    分别设置在边框的左、右两侧框上的第二人体传感器，且所述第二人体传感器可沿相应侧框的竖直轴线转动，以调节感测方向。
  10. 一种主动关怀的实现系统，包括后台服务器、多个家居装置，其特征在于，还包括权利要求5-9任一项所述的主动关怀的实现装置。
PCT/CN2018/113750 2018-11-02 2018-11-02 主动关怀的实现方法、系统及装置 WO2020087515A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/113750 WO2020087515A1 (zh) 2018-11-02 2018-11-02 主动关怀的实现方法、系统及装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/113750 WO2020087515A1 (zh) 2018-11-02 2018-11-02 主动关怀的实现方法、系统及装置

Publications (1)

Publication Number Publication Date
WO2020087515A1 true WO2020087515A1 (zh) 2020-05-07

Family

ID=70463100

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/113750 WO2020087515A1 (zh) 2018-11-02 2018-11-02 主动关怀的实现方法、系统及装置

Country Status (1)

Country Link
WO (1) WO2020087515A1 (zh)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN203941577U (zh) * 2014-07-08 2014-11-12 天津鹏安数迅科技有限公司 自助银行安全提示预警系统
CN104504623A (zh) * 2014-12-29 2015-04-08 深圳市宇恒互动科技开发有限公司 一种根据动作感知进行场景识别的方法、系统及装置
CN204990616U (zh) * 2015-08-14 2016-01-20 上海正苒智能科技有限公司 监护告警系统
US20160134773A1 (en) * 2014-11-11 2016-05-12 Ricoh Company, Ltd. Information processing apparatus, user detection method and computer program product
CN108542420A (zh) * 2018-02-23 2018-09-18 山东沃尔德生物技术有限公司 一种老年人看护系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN203941577U (zh) * 2014-07-08 2014-11-12 天津鹏安数迅科技有限公司 自助银行安全提示预警系统
US20160134773A1 (en) * 2014-11-11 2016-05-12 Ricoh Company, Ltd. Information processing apparatus, user detection method and computer program product
CN104504623A (zh) * 2014-12-29 2015-04-08 深圳市宇恒互动科技开发有限公司 一种根据动作感知进行场景识别的方法、系统及装置
CN204990616U (zh) * 2015-08-14 2016-01-20 上海正苒智能科技有限公司 监护告警系统
CN108542420A (zh) * 2018-02-23 2018-09-18 山东沃尔德生物技术有限公司 一种老年人看护系统

Similar Documents

Publication Publication Date Title
CN109634129B (zh) 主动关怀的实现方法、系统及装置
US11830333B2 (en) Systems, methods, and devices for activity monitoring via a home assistant
CN109445288B (zh) 一种智慧家庭普及应用的实现方法
FI122787B (fi) Etävalvontajärjestelmä
Jiang et al. Smart home research
US9672728B2 (en) Smart hazard detector drills
CN110275443A (zh) 主动的智能控制方法、系统及智能装置
CN110599747A (zh) 一种用户提醒方法、装置及智能门铃系统
CN110049404B (zh) 智能装置及其音量控制方法
Bregman et al. A universal implementation model for the smart home
CN111754672B (zh) 一种智能门铃识别系统
Alesii et al. Short range wireless solutions enabling ambient assisted living to support people affected by the Down syndrome
WO2020087515A1 (zh) 主动关怀的实现方法、系统及装置
US11869104B2 (en) Visitor-tailored property configuration
WO2020223967A1 (zh) 主动的智能控制方法、系统及智能装置
CN116931439A (zh) 一种构建无命令式主动智能的实现方法、装置和设备
US11775132B1 (en) System and method for the management and use of building systems, facilities, and amenities using internet of things devices and a metaverse representation
WO2020215204A1 (zh) 智能装置及其音量控制方法
WO2019232719A1 (zh) 一种艺术智能电视装置
Monnier Rethinking the Study-Bedroom: How to Better the Quality of Life in Student Housing and Its Environment Focusing on Mental Health
KR20230083445A (ko) 스마트 홈 제어 시스템
Qiu et al. Analysis for Interactive Design in the Elderly’s Residential Space
Yang et al. Analysis and Design of Interior Space for Elderly under the Background of Smart Elderly Care
WO2020150927A1 (zh) 智能装置、室内人体三维定位方法及智慧家庭实现方法
CN114764222A (zh) 计算机可读存储介质、智能面板及其消息推送方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18938820

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18938820

Country of ref document: EP

Kind code of ref document: A1