System for performing sound-based sensing of an object in a sensing area
By using sound-based sensing systems in an office environment and using a network of sound generators and detectors, the problem of invasiveness and accuracy of sitting posture detection in traditional technology is solved, real-time monitoring of incorrect sitting postures and Early warning improves the accuracy and user experience of health monitoring.
Patent Information
- Application Number
- CN202380032895.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2022-04-14
- Filing Date
- 2023-03-29
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art is difficult to effectively detect and prevent health problems due to long-term incorrect sitting posture, especially in office environments where traditional camera or pressure sensor-based solutions have problems of invasiveness and insufficient accuracy.
Using a sound-based sensing system, the non-unit of the object is realized by distributing network equipment in the sensing area, using signal propagation between the sound generator and the sound detector, and combining RF signal communication and multipath transmission channel analysis. Invasive, accurate sitting posture detection and can be used for monitoring falls and abnormal postures.
Real-time detection and early warning of incorrect sitting postures is achieved, the accuracy and user experience of health monitoring are improved, and the interference to humans and the environment is reduced. It is suitable for scenes such as office desks, especially for the elderly and those who may have health risks. crowd.
Smart Images

Figure CN119012966A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to systems, methods and computer programs for performing sound-based sensing of objects in a sensing area. Background Art
[0002] Since many people work in front of computers and sit at desks, a person's sitting posture is very important in terms of improving their health. Non-ergonomic sitting posture can lead to decreased health, especially in terms of back pain. It should be noted that improper sitting posture at home or in the office may not lead to motivational injuries, but if there is long-term exposure to improper posture, the damage may be apparent months or years later.
[0003] Figure 1 A representation of a healthy sitting posture at a desk is shown. A person 30 sits on a chair 40 at a desk 40, typing on a keyboard 42 and watching a display 41. In addition, a lamp 34 is placed on the desk 40. For an ergonomic sitting posture, the waist angle 4 should be ≤20°, the neck angle 3 should be ≤20°, the visual angle 2 should be between 15 and 30°, and the viewing distance 1 should be between 50 and 80 cm. Prevention of incorrect sitting posture is very important for improving the health of people who often sit at a desk.
[0004] Several camera-based or pressure sensor-based solutions for detecting incorrect sitting posture are known.
[0005] Li et al. published "Sitsen: Passive sitting posture sensing based on wireless devices" in the International Journal of Distributed Sensor Networks on July 7, 2021, demonstrating a contactless wireless-based sitting posture detection using RFID tags.
[0006] US2012 / 116252A1 discloses detection of body orientation and / or posture. At least one wave sensor may be configured to output waves and collect measurement data based on reflections of the output waves. At least one processor may be configured to receive measurement data from at least one wave sensor and evaluate the received measurement data to determine the posture of the monitored object. Based at least in part on the determined posture, one or more suitable control actions may be implemented. Summary of the invention
[0007] It is an object of the present invention to provide a non-invasive sensing capability enabling detection of an incorrect sitting posture of a subject.
[0008] In a first aspect of the present invention, a system for performing sound-based sensing of an object in a sensing area is proposed. Sensing is performed by a network of network devices that perform sound-based sensing in the sensing area. At least one network device includes a sound generator and at least one network device including a sound detector. The network devices are distributed in the sensing area and communicate with each other based on RF signals. The sound generator generates a sound signal and the sound detector detects the sound signal after the sound signal propagates through at least a portion of the sensing area. The sound sensor generates a sensing signal indicating the detected sound. The sound generator and the sound sensor are arranged at or in different network devices. The sitting posture detector detects the sitting posture of the object based on the sensing signal. Therefore, active detection that allows objects in the sensing area to be actively sensed can be provided.
[0009] The network of network devices can also be understood as a sensing network, which includes at least two network devices, in particular at least one network device suitable for generating sound and at least two network devices suitable for detecting sound. Preferably, the network includes more than three network devices, wherein the number of network devices in the network can be adjusted based on the sensing area in which sensing should occur. For example, the larger the space, the more network devices can be provided in the network, and / or the more complex the shape of the space, the more network devices can be provided in the network. Preferably, all network devices include a sound generator and are therefore suitable for generating sound, and a sound detector and are therefore suitable for detecting sound. However, the network may also include one or more network devices dedicated to generating sound, and therefore only include a sound generator and two or more network devices dedicated to detecting sound, and therefore only include a sound detector.
[0010] The sitting posture detector can be adapted to control at least one sound generator, in particular to control a network device comprising a sound generator, to generate a predetermined sound. The predetermined sound can generally refer to any sound that can be provided by the sound generator and includes predetermined characteristics, such as a predetermined length, a predetermined spectrum, a predetermined amplitude, etc.
[0011] In one embodiment, the sitting posture detector can be configured to estimate the hip and head positions of an object (particularly a person) sitting at a table or desk. Using information about the hip and / or head positions, the sitting posture detector can perform better estimation of the sitting posture of the person.
[0012] In one embodiment, the sitting detector may include a controller to control the operation of at least one network device. Thus, for example, the operation of a sound generator and / or a sound detector in one of the network devices may be controlled by the sitting detector.
[0013] In one embodiment, the controller of the sitting detector may control the operation of the sound generator to determine the sound waves and / or frequencies of the sound signals generated in at least one first and / or second audio channel. Thus, the sitting detector may control the operation of the sound generator. Optionally, the controller may adjust the operation of the sound generator if necessary.
[0014] In one embodiment, the first audio channel may be in an audible frequency range and the second audio channel may be in an inaudible frequency range. Optionally, the audible frequency range may be below 18 kHz. Optionally, the inaudible frequency range may be >18 kHz. Thus, it is also possible to generate sound signals that are not audible to humans and therefore have no negative impact on humans. The frequency range may also be adjusted if a pet is present. Optionally, the user of the system may provide information to the controller whether a pet may be present or whether it may occur or request other restrictions.
[0015] In one embodiment, at least one sound generator generates a sound signal in a first audio channel. At least one second generator generates a second sound signal in a second audio channel. At least one sound sensor is configured to detect the sound signal in the first audio channel and generate a sensing signal indicating the detected sound signal. The sound sensor may also be configured to detect the sound signal in the second audio channel and generate a sensing signal indicating the detected sound signal. Alternatively, a second sound sensor may be provided for the second audio channel. Thus, sound-based detection may be performed based on sound signals having different frequencies to increase the accuracy and robustness of sound-based object sensing.
[0016] In one embodiment, the controller of the sitting posture detector can control the sound generator to generate an audible sound signal or an inaudible sound signal in the first audio channel, and can control the sound generator to generate an inaudible sound signal or an audible sound signal in the second audio channel. The controller can optionally start the generation of the inaudible sound signal or the audible sound signal only when the sitting posture detector has detected the correct head position of the object. Therefore, it is possible to reduce the amount of sound signals emitted in the sensing area. If the detected head position is not the correct head position, the detection of the hip position can be avoided. Therefore, the audio pollution level in the system can be reduced.
[0017] In one embodiment, the sitting detector includes a fall detector to determine whether an object or person has fallen from a chair based on the sensing signal from the sound sensor. Thus, the system can be extended to fall detection. This can be particularly advantageous if the person sitting at the desk is an elderly person, or if the person working at the desk tends to fall asleep.
[0018] In one embodiment, the sitting posture detector can determine an abnormality about the sitting posture compared to the previously determined sitting posture based on the sensing signal from the at least one sound sensor, and issue a warning if an abnormality is determined. This is also advantageous for detecting or monitoring, for example, the elderly to warn about any sitting posture that may indicate disease or illness.
[0019] In another aspect of the present invention, a method for performing sound-based sensing on an object in a sensing area based on a network of network devices is provided, the network devices being configured to perform sound-based sensing in the sensing area. The network devices are distributed in the sensing area and are configured to communicate with each other based on RF signals. The sound signal is generated by a sound generator in one of the network devices. The sound signal from the sound generator is detected after propagating through at least a portion of the sensing area, and a sensing signal indicating the detected sound is generated by a sound sensor in one of the network devices. The sound generator and the sound sensor are arranged at different network devices. The sitting posture of an object in the sensing area is detected based on the sensing signal from the sound sensor.
[0020] In one embodiment, the user's absolute sitting posture or relative sitting posture can be detected. The relative sitting posture can be used to determine abnormalities in the sitting posture of, for example, the elderly. A slumped posture may indicate a diagnostic feature of depression.
[0021] According to one embodiment, the first head position is detected using inaudible sound sensing (which has only a short sensing range). This can detect the head position directionally. If the correct head position is detected, audible sound sensing (which has a sensing range of up to 5m) can be activated.
[0022] Alternatively, audible sensing is first performed to determine the torso position (hip position) using a long sensing range (audible audio signal). Only when the torso position is normal is inaudible sensing performed to detect the head position.
[0023] According to one embodiment, the smart device can be used to forward information to the user on how to position the smart device to improve audio sensing. For example, the first network device can be implemented as a lighting device with a sound generator, and the microphone of the smart device such as a laptop or smartphone can be used as a sound sensor in the system.
[0024] The sitting posture detector may include a controller that can control the operation of the network device. In particular, the controller can control the operation of the sound generator (e.g., the sound waveform, frequency of the sound signal). The sitting posture detector may also include a memory for storing and tracking the position of the head and / or hips detected over time. The sitting posture detector may also include a fall detector that can detect a person falling from a chair based on a detected audio signal from the sound sensor. The sitting posture detector may also include an alarm unit for alarming a person if an incorrect sitting posture is detected or if a fall occurs.
[0025] Typically, the sound generator and the sound detector can be considered as a sound propagation pair, and the sound generated by the sound generator propagates from the sound generator to the sound detector in a multipath transmission channel between the sound propagation pair. Preferably, the sound generator and the sound sensor are arranged at or in different network devices. For example, the audio multipath transmission channel can be three-dimensional and shaped so that it is narrower at the point where the sound is generated, wider during propagation through the space, and narrower again at the point detected by the sound detector. However, the exact shape of the audio multipath transmission channel is determined by the environment, in particular the sensing area through which the sound propagates. Typically, due to the different reflections of the generated sound on one or more surfaces, the audio multipath transmission channel can be considered to include multiple audio paths. For example, one of the multiple audio paths can refer to the sound reflected by the table before detection, one can refer to the direct path between generation and detection, and one can refer to the reflection from the wall before detection. Since the sound is generated by at least one sound generator and then detected by multiple sound detectors, different (i.e., multiple) channels can be used during sensing, and these channels include multiple audio paths. Therefore, the sound propagation after generation refers to multi-channel propagation, wherein each detector detects one of the multi-channels as detected sound. Based on the detected sound, the sitting detector is then adapted to control the detector to generate a sensing signal indicative of the detected sound. The sensing signal may particularly refer to an electrical signal indicative of the detected sound.
[0026] The sensing signal utilized by the object determination unit indicates the sound detected within a predetermined time period, such as a few seconds. Therefore, the sensing signal generally does not refer to a value, but to a continuous measurement of the detected sound within a predetermined time period. For example, the time period can be determined by the length of the predetermined sound generated by the sound generator, and the measurement of the detected sound can start at the same time as the sound generator starts generating the predetermined sound. In particular, in a small sensing area, the propagation time of the sound between the sound generator and the sound detector can be negligible.
[0027] In one embodiment, the state and / or position is determined based on i) the signal strength of multiple detected sensing signals and / or based on ii) channel state information derived from multiple detected sensing signals and predetermined generated sounds. With respect to the first option, the state and / or position of at least one object in the sensing area can be based on the signal strength of multiple detected sensing signals. Since the signal strength of the sensing signal depends on the amplitude of the sound reaching the sound detector directly or indirectly, a single intensity indicates different paths that the sound can propagate from the sound generator to the sound detector. Since these paths are highly dependent on the environment between the sound generator and the sound detector, and therefore also on the location of the object in the environment, the signal strength indicates these locations. In addition, since the signal strength of multiple detected sensing signals is utilized, where the sensing signals are generated by detection of sound detectors arranged in different orientations, information about the network device environment from multiple different sound paths is provided by the sensing signals.
[0028] In a second option, the position of at least one object in the sensing area can be based on channel state information derived from multiple detected sensing signals and determined generated sounds. The channel state information indicates the properties of the path that the sound has taken from the sound generator to the sound detector, and thereby describes how the sound has propagated from the sound generator to the sound detector. Therefore, the channel state information also indicates the interaction of the sound with the object along the propagation path. Therefore, the channel state information provides very accurate information about the network environment with which the sound has interacted (for example, the sound has been reflected, scattered or absorbed from it). Since the predetermined generated sounds are known, and due to the network characteristics of the network, the channel state information can be derived from the sensing signals and the predetermined generated sounds.
[0029] In one embodiment, the sitting detector is adjusted so that the sound generator generates a predetermined sound as a directional sound, wherein the directional sound is directed to at least one person. The advantage of generating directional sound is that the influence of other objects that should not be detected can be minimized. In addition, the influence of general environment such as walls, ceilings or floors on the detected sound can also be minimized. If the direct line of sight from the sound generator to the person is blocked, the directional sound can also be directed to a flat surface in the room so that the reflection of the flat surface reaches the object. In this embodiment, it is preferred that the flat surface does not change its state and position frequently, so that the change of the sensing signal only indicates the change of the object, and does not indicate the change of the flat surface also listed in the signal path. In order to generate directional sound, any known method can be used. For example, the sound generator can be suitable for including a speaker array with multiple sound generator speakers, which allows the sound generated by the speaker array to be guided based on the interference of the sound generated by each individual speaker.
[0030] In one embodiment, the sitting posture detector is adjusted so that the sound generator generates the predetermined sound as an omnidirectional sound. Generating the predetermined sound as an omnidirectional sound has the advantage that the state and / or position of the entire environment of the network can be taken into account.
[0031] In one embodiment, each sound detector comprises an array of sound detectors, so that each of the plurality of sensing signals indicates a direction from which the sound detected therefrom has reached the detection array, wherein the sitting detector is adapted to determine the state of the person further based on the directional information provided by each sensing signal. In particular, the array of sound detectors allows a more accurate determination of the path that the sound has propagated from the sound generator to the sound detector, and in particular to distinguish between these different paths. This allows a more accurate determination of the state and / or position of the object. In particular, this can be advantageous when using directional information to determine the state of a person in space.
[0032] In one embodiment, each network device comprises a sound detector and a sound generator. The sitting detector is adapted to control the sound generator of the network device to generate a predetermined sound and control the sound detectors of all other network devices to detect the generated sound, so that for each sound generated by a different sound generator, a plurality of detected sensing signals are generated, wherein the position of the person is determined based on each of the plurality of audio sensing signals.
[0033] In one embodiment, the sitting detector is adapted to control the sound generator of the network device to subsequently generate a predetermined sound, and to control the sound detectors of all other network devices to detect the subsequently generated sound. In particular, the sitting detector may be adapted to control the first network device (i.e. the sound generator of the first network device) to generate a predetermined sound, and to control all other network devices (i.e. the detectors of all other network devices) to detect the generated sound of the first network device, so as to generate a sensing signal corresponding to the first generated predetermined sound. Then, the sitting detector is adapted to control the second network device to generate a predetermined sound, and to control all other network devices to detect the predetermined sound to generate a sensing signal corresponding to the second generated predetermined sound, and to do so until all network devices have generated the predetermined sound at least once. Then, the sitting detector is adapted to determine the position of at least one person in the space based on all the sensing signals, wherein also in this case, for example, the method for determining the state and / or position of an object based on each of a plurality of audio sensing signals already described above may also be used. The time sequences of different predetermined sounds generated by different sound generators may be similar to each other or may be different from each other.
[0034] In another preferred embodiment, the sitting posture detector is suitable for controlling the sound generator to generate different predetermined sounds at the same time, and controlling the sound detectors of all other network devices to detect the different generated sounds, so that a sensing signal for each different predetermined sound is generated by the sound detector. The different predetermined sounds preferably refer to sounds listed in different frequency ranges. For example, the sitting posture detector can be suitable for controlling the first sound generator to generate a first sound with a first frequency in the audible frequency range, and simultaneously controlling the second sound generator to generate a second sound with a second frequency in the audible frequency range. If the first and second frequencies are selected to be listed in sufficiently different frequency ranges, the two detected sounds can be separated by the sound detector to generate different sensing signals for different predetermined sounds. Therefore, sensing signals involving different combinations of sound generators and sound detectors can be sensed simultaneously. In addition, the above embodiments can also be combined.
[0035] In such an embodiment, the sitting posture detector is adapted to control the sound generator of the network device to subsequently generate a different predetermined sound, and to control the sound detectors of all other network devices to detect the subsequently generated different sound. In particular, such a combination may be advantageous in cases where different combinations of sound detectors and sound generators require different frequencies to determine the state and / or position of an object.
[0036] According to one embodiment, the network device can be implemented as a lamp (ceiling lamp, table lamp), a smart device (smartphone, smart watch, tablet computer, smart speaker), or an electronic device with network function (such as a laptop computer).
[0037] In one embodiment, at least one network device includes a lighting unit for implementing a lighting function. However, in other embodiments, the network device may also include other functions, such as an entertainment function, a monitoring function, and the like.
[0038] In another aspect of the present invention, a computer program product for performing sound-based sensing of a person in a sensing area is provided, wherein the computer program product comprises program code means for causing the above system to perform the above method.
[0039] It shall be understood that the system, the method, the computer program and the network comprising the system have similar and / or identical preferred embodiments, in particular as defined in the dependent claims.
[0040] It shall be understood that a preferred embodiment of the invention can also be any combination of the dependent claims or the above-described embodiments with the respective independent claim.
[0041] It shall be understood that the above aspects and in particular the system of claim 1, the method of claim 14 and the computer program product of claim 15 have similar and / or identical preferred embodiments, in particular as defined in the dependent claims.
[0042] These and other aspects or embodiments of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] In the following figures:
[0044] Figure 1 shows the performance of healthy sitting posture at the desk,
[0045] Figure 2 A block diagram of a system for performing sound-based sensing of a person in a sensing area is shown,
[0046] Figure 3A , 3B and 3C each show a sound-based sensing representation of a person at a desk, and
[0047] Figure 4 A flow chart of a method for sound-based sensing of a person in a sensing area is shown. DETAILED DESCRIPTION
[0048] Figure 2 A block diagram of a system for performing sound-based sensing of a person in a sensing area is shown. The sensing system includes a network 10, which preferably includes several network devices 100. The network devices 100 can communicate with each other based on RF signals. The network device 100 includes an RF transceiver 110 or an RF transmitter and an RF receiver for RF communication. The network device 100 may include a sound generator 120 and / or a sound detector 130. The sound generator 120 can be implemented as a speaker capable of generating directional or omnidirectional sound. The sound sensor 130 can be implemented as a microphone or a microphone array. Optionally, the sound sensors 130 of several network devices 100 can be used as a microphone array. Here, the detected sound can be processed and the location information of various network devices can be used during the sound processing.
[0049] The network device 100 may optionally include a lighting unit 140. Furthermore, the network device may include a controller 150.
[0050] The network device 100 may be arranged at different locations in the sensing area 20. In the sensing area, a person 30 (e.g., a subject) may, for example, sit on a chair 40 at a desk 40. The person 30 may have a sitting posture characterized by a head position 31 and a hip position 32.
[0051] At least one sound generator 120 of one of the network devices 100 generates a sound that propagates through the sensing area 20 and is affected by the object 40 and the person 30. The reflected sound (i.e., audio multipath transmission) can be picked up by at least one sound sensor 130. Based on the received (reflected) sound, the sitting posture detector 200 can determine the sitting posture of the person 30.
[0052] According to one example, multiple audio sensing channels may be used to determine a 3D position estimate of a person 30 in the sensing area 20. In particular, a head position 31 and a hip position 32 are detected.
[0053] The sitting posture detector 200 includes a controller 210, which can control the operation of the network device 100. In particular, the controller 210 can control the operation of the sound generator 120 (e.g., the sound waveform, frequency of the sound signal). The sitting posture detector 200 may also include a memory 220 for storing and tracking the positions of the head 31 and / or the buttocks 32 detected over time. The sitting posture detector 200 may also include a fall detector 230, which can detect a person falling from the chair 40 based on the detected audio signal from the sound sensor 130. The sitting posture detector 200 may also include an alarm unit 240 for issuing an alarm to the person if an incorrect sitting posture is detected or if a fall occurs.
[0054] The memory 220 may be used to track historical data of the user's head 31 and / or hip positions 32. The historical data may also be used by the sitting posture detector 200 to estimate the current head and hip positions.
[0055] According to one embodiment, the sitting detector 200 can collect the output of the sound sensor 130 over time to track the movement and body posture over time. The sitting detector 200 can use this information to perform long-term analysis of sitting posture or changes in sitting posture. In particular, the sitting detector can perform fall detection (e.g., determine whether a person has fallen from a chair).
[0056] Preferably, at least two network devices 100 are used to generate sound and detect sound to be able to improve the accuracy of audio-based sensing of the person 30. Optionally, one network device 100 can be placed on a table 40 adjacent to the user 30. The network device 100 can be implemented as or can include a lighting unit and used as a desk lamp. Therefore, the network devices 100 arranged at different positions in the sensing area 20 and the network device 100 implemented as a desk lamp can be used to generate (audible or inaudible) sound and detect reflected sound to enable audio-based sensing of the position of the person 30.
[0057] According to one embodiment, the network device 100 can be implemented as an intelligent device on which an application is run. Here, such an intelligent device includes a microphone and a speaker, and can be used to generate sound and / or detect the sound emitted. The result can be used by the sitting posture detector 200 to determine the sitting posture of the person 30. If one of the network devices 100 is implemented as an intelligent device like a smart phone, the internal sensor of the smart phone that generates orientation data can be used to guide the user to change the position and / or orientation of the smart phone so that sound generation or sound reception can be improved. If the smart phone has a touch screen, the touch screen can be used as a user interface. If the intelligent device is implemented as a laptop, user instructions can be displayed on the laptop to change the position and / or orientation of the laptop to improve sound generation or sound reception to improve the accuracy of sitting posture detection. In one example, the user instructions can be displayed on one of the network devices 100. In this example, at least one network device 100 has a suitable screen to display the user instructions.
[0058] According to one embodiment, if the network device 100 is implemented as a desk lamp, the user may receive audible and / or visual information on how to position the network device to improve sitting posture detection accuracy.
[0059] According to one embodiment, the network device 100 can be implemented as a ceiling lamp. A first audible sensing channel can be established between two ceiling lamps to estimate the shape and position of the desk 40 in the sensing area 20. When the first audio sensing channel has located the desk surface and the chair position, a second inaudible audio sensing channel formed between the ceiling lamp 100 and the network device 100 implemented as a desk lamp can be used. Based on the information from the second audio sensing channel, the 3D position of the person's head 31 and the hip position 32 (the pressure point between the body and the chair surface) can be estimated and tracked. Based on the head position 31 and the pressure point position 32 (hip position), the sitting posture detector 200 can estimate whether the person's position is related to a healthy or incorrect sitting posture.
[0060] According to one embodiment, a first audible sound emitted by a sound generator 120 of one of the network devices 100 (e.g. implemented as a ceiling light) can be used to detect the head position 31. The accuracy of such sensing can be >1m. A second inaudible sound can be emitted by one of the sound generators 120 (network device implemented as a desk lamp). This second audio channel can enable accuracy for head position estimation within cm range. The hip position (pressure point) can be determined based on the combination of the signals from the first and second audio sensing channels.
[0061] The frequencies used for the first and second audio channels must take into account possible interference from third party equipment.
[0062] According to one embodiment, the system is also able to detect several individuals sitting in the sensing area and determine their head and hip positions to determine whether they are sitting properly.
[0063] In order to minimize any audible interference or ultrasonic sound pressure to the user's ears or animals in the sensing area, the first and second sensing channels can be activated subsequently. For example, the second audible audio sensing channel is activated only when the first audible audio sensing channel has determined that the person's head posture is correct. Then, the second inaudible audio sensing channel is started to determine the user's hip position. If the information from the first audio sensing channel indicates that the head posture is incorrect, no further audio sensing is performed, because a bad head posture will inevitably lead to a bad sitting posture. Optionally, the sitting posture detector 200 can activate a bad sitting posture alarm to the user. This can be performed as an audio signal via the computer on which the user is working or via the user's smartphone. Alternatively, the alarm can also be an optical alarm, such as turning on and off a desk lamp or ceiling lamp. The alarm may also be a combination of the above alarms.
[0064] The network device according to the embodiment can be regarded as any device suitable for forming a network with other network devices. In particular, the network device includes a network device communicator suitable for receiving and transmitting wired or wireless signals (such as radio frequency signals, infrared signals, electrical signals, etc.).
[0065] The network between the network devices can then be formed by communication between the network devices following known network communication protocols (such as WiFi, ZigBee, Bluetooth, etc.). Preferably, the network device refers to an intelligent device, i.e. a device that includes a communication unit for receiving and transmitting network communication signals but which implements the functions of the corresponding traditional device in other ways. In particular, such an intelligent device may be a smart home or office device, in which case the corresponding traditional functions will be the functions of a traditional home or office device. Preferably, the traditional functions refer to lighting functions, and the network device refers to a network lighting device that is further adapted to include a sound generator and / or a sound detector. However, the network device may also refer to, for example, a smart plug, a smart switch, etc.
[0066] The system may be part of a network, for example, may be part of one or more network devices. In particular, the system may be provided as hardware and / or software as part of one of the network devices, or distributed across a plurality of network devices that communicate with each other to form a system. However, the system may also be provided as a standalone system, for example in a device that is not part of a network of network devices but communicates directly or indirectly with at least one network device, for example to control the network device. For example, the system may be provided as part of a handheld computing device, such as a smartphone, a tablet computer, a laptop computer, etc. However, the system may also be located in a cloud formed by one or more servers, wherein in this case the system may communicate with the network, in particular the network devices, via one or more devices connected to the cloud, such as a router.
[0067] Figure 3A , 3B 3C each show a sound-based sensing performance of a person at a desk. The network 10 may include three network devices 100. Two of the network devices 100 may be implemented as ceiling lights, and the third network device 100 may be implemented as a desk lamp or smart device placed on a desk near the person 30. In the sensing area 20, the person 30 is sitting at the desk 40, and the third network device 100 (desk lamp) is placed on top of the desk 40. The network device 100 may be implemented or may include a lighting unit. The network device 100 according to FIG. 3 may correspond to the network device 100 according to FIG. Figure 1 Therefore, the network device 300 may include a sound generator 120 and / or a sound detector 130. The frequency of the sound generated by the sound generator 120 in the ceiling network device 100 may be different from the frequency of the sound generated by the sound generator 200 in the network device 100 implemented as a table lamp. In particular, the frequency of the sound generated by the sound generator 120 of the table lamp 100 is preferably within the inaudible sound range. Based on the reflected sound picked up by the sound sensor 130 in the table lamp network device 100, the position of the person's head 31 can be detected. If the distance between the network device and the user's head 31 is less than 1m, sensing based on inaudible sound can be performed.
[0068] Therefore, the (desk lamp) network device 100 emits a first sound at a first frequency range corresponding to the inaudible frequency range. Based on the detected (reflected) inaudible audio signal, the sitting posture detector 200 estimates the head position 31 and the chair position based on the multipath characteristics of the inaudible sensing signal.
[0069] The ceiling light network device 100 is arranged at a distance from the person 300, which is usually greater than one meter. Therefore, the distance is too large for inaudible sound sensing. Therefore, the ceiling network device 100 is emitting a second audible sensing signal. The reflected audio sensing signal can be received by the second ceiling network device 100 or by the ceiling network device that emitted the second audible sensing signal. Based on the received audible sensing signal, the sitting posture detector 200 estimates the shape and position of the desk surface and the orientation of the desk relative to the architectural structure of the room and other objects in the sensing area.
[0070] exist Figure 3A In FIG, the head position 31 of the person 30 is determined by the (desk lamp) network device 100. Figure 3B In the embodiment, further location detection is performed by the ceiling network device 100.
[0071] According to one embodiment, a first audio channel based on inaudible sound and at least one second audio channel based on audible sound are used to detect the sitting posture of the person 30. The first audio channel can be provided by a network device 100 (such as a desk lamp) placed near the person (<1m). The second audio channel can be implemented by a network device 100 arranged at a greater distance. Such a network device can be, for example, a ceiling network device (ceiling light unit). The network device for the second audio channel can also be a light unit arranged on the wall of the sensing area.
[0072] According to a further embodiment, the first audio channel with inaudible sound may be implemented by a smart device or a laptop computer arranged in the vicinity of the person 30 whose sitting posture is to be detected and evaluated.
[0073] According to one embodiment, the sound sensor in the first and / or second audio channel may be the same device as the transmitter. Alternatively, the sound generator (transmitter) may be arranged in a network device different from the transmitter that is emitting the inaudible audio signal in the first audio channel. In addition, optionally, the transmitter that is emitting the audible audio signal in or for the second audio channel may be arranged in the same network device as the detector (microphone) for the audio signal. Alternatively, the sound detector may be arranged in a network device different from the network device in which the sound generator for the second audio channel is arranged. As an example, the sound sensor may be implemented as a microphone array, in which the sound detectors of several network devices are used to detect the audible sound of the second audio channel. This is advantageous because it will allow more accurate sensing of the head position 31 and hip position 32 of the person 30.
[0074] In one embodiment, the sound generators from different network devices can be used in combination to generate and emit audible sound in the second audio channel. For different network devices, the frequency of the sound signal generated by the corresponding sound generator may be different.
[0075] According to a further aspect of the invention, in a first step, the approximate position of a person or object 30 in the sensing area is detected, and then a more precise position detection can be performed, wherein the frequency at which the sound generator generates the sound signal can vary over time, or different frequencies can be generated for different network devices (the frequency can depend on the position of the network device 100 in the sensing area 20). In one example, the approximate and / or precise orientation / position of the object or person 30 can be determined by RF-based sensing. The position of the network device 100 can be relative to the position of the object 30. This can result in a more precise determination of the user's head and / or hip position to allow estimation of the user's sitting posture.
[0076] By using the first audio channel (non-audible audio signal), it is possible to track the head and hip positions of a person relative to the recliner furniture at the cm level. Based on the 3D head position 31 and hip position 32, the sitting posture detector 200 can determine a fitness score associated with various positions of the head and hips of the person 30. If an incorrect sitting posture is detected, an (audible and / or visual) alarm or haptic can be output.
[0077] According to one embodiment, a network 10 having a plurality of network devices 100 implemented as lighting or light emitter devices may be provided. The network device 100 includes a microphone (sound sensor 130) and / or a speaker (sound generator 120). Therefore, the sound generator 120 and the sound sensor 130 may be distributed in the sensing area 20.
[0078] The sitting posture detector 200 may be implemented as a dedicated device or may be integrated in a wireless hub or a central control unit. Alternatively, the sitting posture detection 200 may also be performed by one of the network devices 100. Alternatively, the sitting posture detection may be performed by a remote device such as a server in the cloud. Furthermore, the sitting posture detection may be performed by an IoT edge device. Therefore, as long as the function is performed based on the audio signal detected by the sound sensor of the network device, the location where the sitting posture detection is performed is irrelevant.
[0079] In addition, sitting detection can also be performed by a smart device that can be located in the sensing area and can communicate with the network 10 and the network device. As an example, an application may be running on such a smart device that performs sitting detection. Alternatively, sitting detection can be performed by a smart speaker or a smart watch that can be arranged in the sensing area.
[0080] According to one embodiment, several network devices 100 are arranged in the sensing area 20. The first plurality of network devices 100 can be arranged adjacent to the person 30. The second plurality of network devices 100 can be arranged farther from the first plurality of network devices. In order to improve the accuracy of the sitting detection, the sitting detector 200 can select some of the first plurality of network devices for the first sensing channel (e.g., inaudible sound signal). In addition, the sitting detector 200 can select the frequency of the inaudible signal and the type of audible signal generated and transmitted. The sitting detector 200 can take into account the distance between the network device 100 (signal generator; sound sensor) and the person, and the effect of any interference in the sensing area on the sensing signal. In addition, any interference from another person adjacent to the first person can be detected. If necessary, the frequency of the sensing signal can be adjusted. In addition, optionally, the sensing frequency can be selected according to the distance between the network device and the chair or person. In particular, the longer the distance between the network device and the chair, the lower the frequency of the sensing signal. If the signal strength received at the sound detector is insufficient, another network device in the network can be selected to generate and / or detect the sensing signal. Furthermore, network devices (speakers; microphones) may be selected based on the CSI.
[0081] In addition, if it is determined that another person or object is moving in the sensing area, the sitting posture detector 200 can pause or interrupt the sitting posture detection. Once the interference has been eliminated, the sitting posture detection can be resumed.
[0082] In a further embodiment, sitting posture detection can be used to detect the health status of an elderly person, for example. The sitting posture detector 200 can be determined as described above. If a slumped posture is detected, this may indicate that the person has depression. In this case, the sitting posture detector can send an alarm to the person to stimulate the person to sit up straight. According to this embodiment, the sitting posture of the elderly is monitored based on audio sensing (as described above).
[0083] According to one embodiment, the sitting posture detector 200 can track the sitting posture of a person and can determine whether the person has recently started to lean to the left or right, or if a sudden drop in sitting posture stability is detected, which may indicate a recent muscle or spinal injury. A corresponding alarm can be issued.
[0084] According to a further embodiment, sitting posture detection can also be used for fall detection. In particular, a sitting posture detector can detect when a person sitting in a chair falls out of the chair. The audio sensing of the person can analyze the pressure points (hip position) and can detect, for example, an elderly person falling on a table surface or the ground.
[0085] According to further embodiments, information from sitting posture detection may be used during RF position or motion sensing. Radio frequency based sensing is a sensing mechanism involving a wireless transceiver (or transmitter / receiver) arranged to transmit and receive radio frequency (RF) signals. These RF signals, which may also be used for radio communications, when passing through a sensing volume, are affected by the presence / motion of people within the sensing volume, for example via reflection, absorption, scattering, etc. Radio frequency based sensing uses such deviations in the radio frequency signals to infer the presence or motion of people. Radio frequency based sensing also extends to other applications such as orientation detection, fall detection, gesture detection, vital sign detection, etc., which are also based on how the radio frequency signals are affected in the sensing volume.
[0086] The network devices 100 may be configured to communicate with each other based on RF signals. In one example, these RF signals may be used for radio frequency based sensing. As an alternative to using the network devices 100 for RF based sensing, other network devices may be used for RF based sensing. In one example, the presence of an object (30) in the sensing area 20 may be determined via RF based sensing.
[0087] In particular, knowledge of body posture can be used to improve the capture of respiration and heart rate associated with body movement. For example, the orientation of the network device 100 (or other device) for RF-based sensing can be selected based on the sensing results of audio sensing (e.g., the results from the sitting detector 200). In other words, the selection of the orientation of the network device 100 can be based on the sitting posture of the object 30. In one example, a network device 100 that is close to the chest / body of the object or has a direct line of sight can be selected for respiration and other heart rates associated with the movement of the object 30. Therefore, after performing audio-based sitting posture detection sensing, this information can be used for RF motion sensing. RF motion sensing can be performed by the network device 100. Here, any RF signal transmitted by the first network device 100 and received by the second network device 100 can be analyzed in view of the movement of the person. Therefore, a combination of audio and RF-based sensing is provided.
[0088] According to one embodiment, if it is detected based on sitting posture detection that a person has fallen out of a chair, the sensing arrangement may be reconfigured and those network devices that have a direct line of sight to the chest area of the person who may have fallen out of the chair may be selected to more accurately measure the person's vital signs. Vital sign measurements may be performed based on audio sensing or RF sensing. A slouching posture leaning toward a table may indicate a stroke, so checking the person's breathing rate and / or heart rate may be performed.
[0089] exist Figure 3AIn the embodiment, the network device 100 is placed on a desk and can be implemented as a desk lamp (the network device can also include a lamp unit 140). The sound generator 120 of the network device can emit audible sounds, for example, between 20 Hz and 15 KHz. The sound generator 120 can also be capable of generating inaudible sounds. Any network device 100 arranged at the ceiling can also include a lighting unit 140.
[0090] exist Figure 3B In the embodiment, the ceiling network device 100 may further include a speaker 120, which is capable of emitting and receiving audible sounds, for example, between 20 Hz and 15 KHz. The sound generator 120 may also be capable of generating inaudible sounds.
[0091] exist Figure 3C In an embodiment, the network device 100 can be placed on a table 400 and can include a sound generator 120 and a sound sensor 130. The sound generator generates and emits sounds with a frequency within the inaudible sound range, that is, for example, 18KHz-20KHz. The network device can also include a sound sensor 130 for receiving multipath signals after interacting with a person 30. The corresponding sensing signals can be used to detect a healthy sitting posture or sitting position or an incorrect sitting posture. Here, the 3D position (X, Y, Z) of the head 31 and the body pressure point (buttocks) 32 on the chair can be used as an indicator of whether the sitting posture is healthy. If based on Figure 1 The typical lumbar angle of 20° shown in Figure 1 (lumbar angle describes the degree of deviation of the body from the central axis) and assuming that the typical length of the human torso plus head is about 75 cm, the maximum deviation of the head position from the central axis is only allowed to be between 25 cm to form a healthy sitting posture. Therefore, the detection accuracy must be less than 25 cm.
[0092] Figure 4 FIG. 1 is a flow chart showing a method for sensing a person in a sensing area based on sound. In step S1, initialization is started. For example, as shown in FIG. Figure 1 In the described system, several network devices 100 can be implemented as ceiling-based network devices. The network devices 100 can include sound generators 120 and sound sensors 130. Sound signals are emitted by at least one sound generator 120, and reflected audio signals are detected by at least one sound sensor 130. Based on the reflected sound signals, the position of the desk 40 in the sensing area 20 is determined. For example, this can be performed by analyzing the distance and arrival angle of the reflected signal from the desk surface through acoustic imaging.
[0093] Alternatively, the desk locations in the sensing area 20 may also be manually input, or the information of the desk locations may be obtained through RF-based sensing performed by the network device 100, for example.
[0094] Furthermore, in step S1, in addition to determining the position and orientation of the desk, the position and orientation of the chair or office chair where the person is sitting or about to sit is also determined. If a plurality of network devices are arranged in the sensing area 20, the sitting posture detector 200 can select some of these network devices 100. Preferably, at least one network device with a line of sight toward the chair is selected. This is advantageous because it can be avoided that other network devices that are not in the line of sight toward the chair may send audio signals, but these signals are blocked by the desk or desk surface. Preferably, several network devices 100 can be selected for sitting posture detection, which are arranged at different positions in the seating area 20. This can allow a significant improvement in sitting posture accuracy. Preferably, in particular, those network devices 100 are selected from the plurality of network devices 100 in the sensing area 20, which have a side view of the chair used by the user. With such a side view, audio sensing can have an unobstructed view of the pressure points on the chair (i.e., the hip position). Optionally, the orientation of the chair backrest is determined, and then the corresponding network device that is most ideally suitable for the audio sensing direction parallel to the backrest is selected.
[0095] In step S2, the presence of a person can be detected. If the presence of a person is not detected, the presence detection of the person is repeated. If a person is detected, the process can continue to step S3. Here, a network device 100 (e.g., in the form of a desk lamp or table lamp) on or at a desk can be used. The sitting posture detector 200 (or controller 210) controls the network device 100 on or at a desk to generate and emit a first audible audio signal on a first audio channel. The frequency of the waveform is preferably within the audible range. The sound sensor 130 of the network device 100 in or at the sensing area detects audio multipath signals that have interacted with a chair, a desk, and a person 30 sitting at the desk. Optionally, multiple sound sensors 130 can be activated to receive multipath signals. Therefore, a microphone array for detecting audio sensing signals can be implemented.
[0096] The sitting posture detector 200 may implement an audio sensing algorithm that determines the position of a person's head and / or detects the person's pressure points (ie, hip positions) on the surface of the chair.
[0097] In step S4, the accuracy of the first sensing signal is determined. If the accuracy is not enough, the network device 100 arranged in the ceiling or at the ceiling in the sensing area can be activated to emit a second sensing signal with an audible waveform. Therefore, it is determined whether the first sensing signal in the first audio sensing channel allows accurate estimation of the hip position. It should be noted that due to the presence of the desk surface, the detection accuracy based on the first sensing signal (with an audible waveform) may not be sufficient to reliably detect the pressure point (i.e., the hip position) of the person. In addition, the desk surface may also reflect some of the first sensing signals from the network device 100 placed in the desk or at the desk, so that the audio signal received by the sound sensor 130 in the network device (e.g., arranged at the ceiling) may be too weak to be used for accurate and reliable detection.
[0098] The sitting detector 200 can compare the position estimated based on the first audio sensing channel (the first audible waveform from the desk network device) with the position estimate from the second audio signal emitted by the (ceiling) network device. If the two position estimates are sufficiently consistent, then in step S6, the position estimate based on the first audio sensing signal can be selected. In this case, only the sensing signal from the (desk) network device 100 can be selected. Therefore, the (ceiling) network device does not have to emit a sound. This is advantageous because it will reduce the noise generated by the sitting detection.
[0099] According to one embodiment, in a first step, the (desktop) network device 100 emits an inaudible waveform to detect the position of the person 30 sitting at the desk. If the signal strength from this first sensing step is good enough, the sitting detector will estimate the user's sitting posture based on the signal. However, if the signal strength is insufficient, the sitting detector 200 can control the (ceiling) network device to emit a second sensing signal (audible signal). Alternatively, the sitting detector 200 can control the (desktop) network device 100 to emit an audible audio sensing signal.
[0100] According to one embodiment, the sitting detector 200 controls a plurality of network devices 100 arranged in the sensing area 20 (e.g., at the ceiling) to perform sitting detection. Preferably, at least one sound generator 120 in these network devices 100 generates a sound waveform in the audible range. The reflected sound is detected by at least one sound sensor 130 of at least one network device 100, and the sitting detector 200 determines, for example, the head position of the user. In addition, the reflected audio signal is analyzed to determine the positions of the chair, the table, and the person 30. Here, preferably, the audio waveform is in the audible frequency range because the network device 100 will be arranged at a distance of more than 2m from the person. The sitting detector 200 analyzes the reflected audio signal to determine, for example, the position of the head position and the hip position of the user. If this first analysis step does not result in sufficient accuracy, the sitting detector can control the sound generator 120 in the network device 100 to emit a second audio waveform, for example, in the inaudible frequency range. Then, the sitting detector analyzes the reflected sound signal again to determine the position of the person in view of the incorrect sitting posture. This two-step approach is advantageous because it will reduce the amount of ultrasonic waves generated by the network device. Therefore, ultrasonic noise in the sensing area 20 can be reduced.
[0101] According to one embodiment, sound waveforms within the audible frequency range may be hidden in white noise emitted by a noise-generating device (eg, the white noise may be utilized to suppress invisible speech disturbances in an office).
[0102] According to one embodiment, the sound generator 120 of at least one network device 100 in the sensing area 20 can generate an audio signal designed to suppress audible speech interference in the office. The audio signal can be suitable for shielding any audio signals generated by the network device for audio sensing purposes. This is advantageous because, for example, a person who is working in the sensing area will not be disturbed by audio-based sensing for sitting detection.
[0103] Alternatively, audio sensing may also be embedded in a public service announcement or in the reproduction of music. This may be advantageous because it will allow the distraction of people 30 in the sensing area 20 to be reduced by audio-based sensing. According to one embodiment, audio-based sensing may be performed, for example, in parallel with the reproduction of the announcement message or while the music is being reproduced. Thus, audio-based sensing may be unobtrusively embedded in any announcement or music.
[0104] According to one embodiment, a network device in or at a table (e.g. implemented as a desk lamp) can emit an inaudible audio waveform (>18KHz) that can be used for head position estimation. Preferably, the network device 100 on the desk should be arranged at a distance of 50 to 80 cm to achieve good position estimation.
[0105] In step S21 and step S22, the head position and the chair surface position may be fused based on the head position and the chair surface position obtained by the first audible audio sensing channel and the second inaudible audio sensing channel. Here, a Bayesian method may be applied.
[0106] According to one embodiment, for example, the hidden marker module HMM can be used to track the head and chair surface positions by analyzing the historical data of an individual. If multiple people occupy the sensing area, one of the network devices may cause the monitoring of the sitting posture of more than one person. Preferably, a tracking algorithm can be used to effectively reduce the interference of the position estimation on the positions of other people in the sensing area 20.
[0107] According to one embodiment, the sitting posture detector 200 calculates the relative distance between the position of the pressure point of the head and body on the chair (hip position) and the touch position of the arm on the desk surface. The machine learning module can be trained with labeled data to indicate whether the sitting posture is healthy. Alternatively, a rule-based algorithm can be used, which uses a distance threshold that depends on the lumbar angle, cervical angle, etc.
[0108] According to one embodiment, audio sensing can also be embedded in the application of a mobile phone or laptop. Optionally, the mobile phone or laptop can act as a network device. Optionally, the audio signal received from the sound sensor 130 in the network device 100 can be used to detect the user's location. In addition, this information can be used to guide the person using the mobile phone or laptop to locate the mobile phone or laptop. Here, the mobile phone or laptop can emit an audio waveform (within an audible or inaudible frequency range). Then, the reflected signal can be received by the sound sensor 130 in the network device to enable position detection. Alternatively, the smart device or laptop can include a microphone, which can be used as a sound sensor to detect the reflected audio signal. Then, the signal can be used by the sitting posture detector to detect the sitting posture of the person. Therefore, the smart device or laptop can be used as a network device as described above.
[0109] Assuming the network has M speakers and N microphones, the received signal is
[0110]
[0111] Where yn(t) is the received signal at the nth microphone, xm(t) is the unit transmission signal from the mth speaker, am is the transmission gain, nm(t) is the noise at the microphone, and hn,m(t) is the channel response of the signal from the mth speaker to the nth microphone, where the arrival angles of different paths are embedded in hn,m(t). Assuming a microphone array is used at each node, the signal we receive at the pth element sensor is,
[0112]
[0113] Where θ is the AoA, and d is the distance, Δ, λ, B, T, and v s are the MIC spacing, wavelength, bandwidth, chirp length, and speed of sound propagation. φ is a constant phase term. We see that the received signal is a 2D sinusoid with a frequency of 2πΔcosθ / λ.
[0114] In the time domain, according to one embodiment, orthogonal audio pulses (such as direct sequence spread spectrum) are designed for each speaker, so the speakers will not cause or minimize negligible interference to each other and themselves, such as delayed signals due to multiple reflections. The identity of the signal received at the network device (lamp-based microphone) (i.e., it can be detected whether the signal originated from speaker 1, or 2, or a combination of 1 and 2). The microphone can use orthogonal matched filters to filter out the signals of the sound generators for networks 1 and 2, respectively. Therefore, time beamforming can be achieved.
[0115] According to one embodiment, it is proposed to preferably use a lighting system embedded with a large number of microphone sensors as detection units distributed throughout the room to monitor object status and / or position. Sound generators can also be integrated into a subset of lighting fixtures. For example, very affordable ultra-cheap audio transmission elements (capable of sending only one beep at a pre-selected fixed frequency) can be easily obtained from children's toys at very low cost and can be used as sound generators. If more advanced programmable audio frequencies are required as predetermined sounds to further improve the audio sensing performance, a series of suitable, very affordable programmable speaker products can be used as sound generators.
[0116] According to one embodiment, the network devices may be part of a lighting system and may each include at least one light unit. The network devices may also include a speaker (sound generator) and a microphone (sound sensor). In particular, the network devices may be adapted to perform audio sensing via at least two different audio sensing channels to detect a sitting posture of a user or object. Optionally, the first audio sensing channel may involve a first short-range audio sensing and the second audio sensing channel may involve a second long-range audio sensing channel formed by a plurality of lighting embedded speakers and microphones. Advantageously, the combination of two different audio sensing channels (a first channel for determining the head position and a second channel for determining the hip position) will significantly improve the accuracy of the audio sensing.
[0117] According to one embodiment, the waveform of one sound generator in the audio sensing channel can be controlled. However, preferably, two different audio sensing channels can be used, each with a different frequency. For example, to detect the head position, a first audible sound emitted from a ceiling light and a second inaudible sound emitted, for example, from a desk lamp can be used to achieve high accuracy of head detection.
[0118] According to one embodiment, a plurality of network devices (sensing nodes) may be provided. Here, the network device may include a sound generator and / or a sound sensor. Optionally, the sound sensor and / or the sound generator may operate in two different audio channels. Preferably, two pairs of nodes may be allocated for the first audio sensing channel and the second audio sensing channel.
[0119] According to one embodiment, in an environment with multiple speakers and microphones (sound generators; sound sensors) that can be close to, for example, a table or desk, the best choice of frequency and sensing nodes to use can be determined. In particular, when selecting the corresponding sound generator and sound sensor, the distance of the microphone and speaker to the chair, as well as the interference effect of the body movement of a second person nearby on the sensing signal can be taken into account. In particular, the longer the distance to the chair where the person is sitting, the lower the sensing frequency should be adjusted. If the signal strength received at the microphone (sound sensor) is insufficient, for example if a low sensing frequency has been selected, another speaker and / or microphone can be assigned to perform gesture detection.
[0120] According to one embodiment, not only the absolute sitting posture of a person is detected, but also abnormalities in the sitting posture (compared to the usual sitting posture) can be detected. This can be used as a sign of health problems, for fatigue detection in cars, trains, buses, for fatigue detection or accident prevention in workshops or factories.
[0121] In order to avoid the above disadvantages of the prior art, the present invention particularly proposes to utilize a distributed microphone grid, i.e. a sound detector grid, integrated in a luminaire in order to monitor a person's sitting posture. The proposed audio sensing solution is capable of monitoring a person's real state.
[0122] Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.
[0123] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality.
[0124] A single unit or device may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.
[0125] Procedures performed by one or several units or devices, such as control of a sound detector or a sound generator, provision of a baseline, determination of the state and / or position of an object, etc., may be performed by any other number of units or devices. These procedures may be implemented as program code means of a computer program and / or as dedicated hardware.
[0126] The computer program may be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state medium, provided together with or as a part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems. Any reference signs in the claims shall not be construed as limiting the scope.
Claims
1. A system for performing sound-based sensing of an object (30) in a sensing area (20), comprising: A network (10) comprising at least two network devices (100) configured to perform sound-based sensing in a sensing area (20), wherein the network devices (100) are distributed in the sensing area (20) and are configured to communicate with each other based on RF signals, wherein at least one network device (100) comprises a sound generator (120) configured to generate a sound signal (101), and at least one network device (100) comprises a sound sensor (130) configured to detect the sound signal (101) from the sound generator (120) after propagating through at least a portion of the sensing area (20) and generate a sensing signal indicative of the detected sound signal, wherein the sound generator (120) and the sound sensor (130) are arranged in different network devices (100), wherein the sound generator (120) is configured to generate the sound signal (101) having a frequency that can change over time or to generate different frequencies for different network devices, and wherein the generation of the different frequencies is based on the location of the network device (100) in the sensing area (20), and A sitting posture detector (200) is configured to detect a sitting posture of an object (30) in the sensing area (20) based on a sensing signal from the sound sensor (130).
2. The system according to claim 1, wherein The sitting posture detector (200) is configured to estimate a hip position (31) and / or a head position (32) of an object (30) sitting at a table or desk (40) in the sensing area (20) based on a sensing signal from the sound sensor (130).
3. The system according to claim 1 or 2, wherein The sitting posture detector (200) includes a controller (210) configured to control the operation of at least one of the network devices (100) to perform sound-based sensing in the sensing area (20).
4. The system according to claim 3, wherein The controller (210) is configured to control the operation of the at least one sound generator (120) to determine a sound waveform and / or frequency of a sound signal generated in at least a first and / or second audio channel.
5. The system according to claim 4, wherein The first audio channel is within an audible frequency range and the second audio channel is within an inaudible frequency range.
6. A system according to claim 4 or 5, wherein at least one sound generator (120) configured to generate a sound signal (101) in the first audio channel, at least one sound generator (120) configured to generate a sound signal (101) in the second audio channel, at least one sound sensor (130) configured to detect a sound signal (101) in the first audio channel and to generate a sensing signal indicative of the detected sound signal (101), At least one sound sensor (130) is configured to detect a sound signal (101) in the second audio channel and generate a sensing signal indicative of the detected sound signal (101).
7. The system according to claim 6, wherein The controller (210) is configured to control one of the sound generators (120) to generate an audible sound signal (101) or an inaudible sound signal in the first audio channel and to control one of the sound generators (210) to generate an inaudible sound signal (101) or an audible sound signal (101) in the second audio channel, The controller (210) is configured to initiate generation of the inaudible sound signal or the audible sound signal only when the sitting posture detector (200) has detected a correct head position of the subject (30).
8. The system according to any one of claims 1 to 7, wherein The sitting posture detector (200) includes a fall detector (230) configured to determine that the subject (30) has fallen from the chair (40) based on a sensing signal from the sound sensor (130).
9. The system according to any one of claims 1 to 8, wherein The sitting posture detector (200) is configured to determine an abnormality regarding a sitting posture compared to a previously determined sitting posture based on a sensing signal from the at least one sound sensor (130) and issue a warning if an abnormality is determined.
10. The system according to any one of claims 1 to 9, wherein At least one of the network devices (100) comprises a lighting function.
11. The system according to any one of claims 1 to 10, wherein At least one network device (100) is implemented as a light unit, in particular a ceiling light, a lighting wall switch or a table lamp.
12. The system according to any one of claims 1 to 11, wherein At least one network device (100) is implemented as a smart device, a smart speaker, or a laptop.
13. The system according to any one of claims 1 to 12, wherein One of the network devices (100) is configured to output or display instructions to a user or how best to position the sound generator (110) and / or the sound sensor (130).
14. A method for performing sound-based sensing of an object (30) in a sensing area (20) based on a network (10), the network (10) comprising at least two network devices (100) configured to perform sound-based sensing in the sensing area (20), wherein the network devices (100) are distributed in the sensing area (20) and are configured to communicate with each other based on RF signals, the method The following steps are involved: generating a sound signal by at least one sound generator (120) in one of the network devices (100), wherein the sound generator (120) is configured to generate a sound signal (101) having a frequency that can vary over time or to generate different frequencies for different network devices, and wherein the generation of the different frequencies is based on the location of the network device (100) in the sensing area (20), detecting the sound signal from the at least one sound generator (120) in one of the network devices (100) after propagating through at least a portion of the sensing area (20), generating, by a sound sensor (130) in one of the network devices (100), a sensing signal indicative of a detected sound, The sound generator (120) and the sound sensor (130) are arranged in different network devices (100), and The sitting posture of the object (30) in the sensing area (20) is detected based on the sensing signal from the sound sensor (130).
15. A computer program product for performing sound-based sensing of an object (30) in a sensing area (20), wherein the computer program product comprises program code means for causing a system according to any one of claims 1 to 13 to perform a method according to claim 14.
