A network control device for controlling a network that performs radio frequency sensing

Abstract

The invention relates to an apparatus for controlling a network 100 comprising sensing areas 110, 120. The apparatus 130 comprises a) a providing unit 131 for providing a sensing result of a radio frequency sensing in a sensing area; b) a determining unit 132 for determining whether the detection result is potentially unreliable in a spatial domain; and c) a defining unit 133 for defining a new sensing area 150 corresponding to a part of the spatial domain in which the detection result is potentially unreliable. The providing unit 131 is adapted to provide a new sensing result in the new sensing area and to determine a detection result for at least a part of the spatial domain for which the detection result is determined to be potentially unreliable based on the sensing result and the new sensing result. Thus, the invention provides an apparatus allowing to improve the reliability of a radio frequency sensing.

Description

Technical Field

[0001] This invention relates to a network control device, a network control method, and a computer program product for controlling a network including network devices that perform radio frequency sensing, and to a network including the control device. Background Technology

[0002] Radio frequency sensing refers to a technique that senses the presence of a person in an area, for example, based on the influence of a human body on wireless signals in that area. This technique is particularly relevant to systems that already include a certain density of wireless devices in an area, because in these cases, the system can utilize its existing ability to transmit radio frequency signals to sense the presence of a person in the area, in addition to their primary functionality (such as lighting).

[0003] Radio frequency (RF) sensing is typically applied to a variety of tasks, such as motion detection, presence detection, breathing detection, fall detection, privacy / security monitoring, and people counting. However, particularly in open areas, such as kitchens, dining rooms, living room combinations, or large office spaces, issues may arise regarding the accuracy of RF sensing. For example, such open areas may be functionally divided into different zones, creating different sensing areas requiring varying levels of detection granularity, with no clear physical boundaries between these areas. This can lead to overlap in the detection areas of the RF sensing network, causing the same event, such as a person entering the living room, to be detected twice—in the sensing area corresponding to the living room and in the sensing area corresponding to the dining room. In another example, an employee in one area might be detected and counted twice because they are in another sensing area, potentially causing the network system to overestimate the number of people in that area.

[0004] For example, the problem of overlapping sensing areas can be addressed by reducing the detection sensitivity within the sensing area. However, this solution may result in gaps or blind spots where the presence of a person will not be detected, which could lead to undesirable performance. Other solutions to this problem involve increasing the amount of data generated for the sensing area, enabling techniques such as machine learning and pattern recognition to be applied to massive amounts of data to improve detection accuracy. However, such solutions are typically computationally expensive in terms of both processing and storage capabilities, and are therefore generally unsuitable for operation on network devices where RF sensing is performed only as an add-on function to the primary functionality and is typically resource-constrained.

[0005] For radio frequency (RF) sensing, it has been suggested that the aforementioned problem could be addressed by allowing network devices to simultaneously be part of multiple sensing areas, thus enabling continuously overlapping sensing regions. While this suggestion may solve the problem of double detection, it also leads to a significant increase in the necessary computational resources provided by the network devices, such as CPU time and memory. It may also potentially cause a further increase in data traffic within the network, which could lead to network congestion. Since many network devices using RF sensing (such as lighting equipment) have significant limitations on the resources available to them, it would be advantageous if a solution existed that allows for rapid confirmation or rejection of detection results without increasing the computational resources required for that solution.

[0006] WO 2020 / 043592A1 discloses a system for selecting one or more devices in a wireless network to transmit, receive, and / or process radio frequency signals for presence and / or location detection. The system includes at least one processor configured to determine the suitability of each of a plurality of devices for transmitting, receiving, and / or processing the radio frequency signals for presence and / or location detection, select a subset of devices from the plurality of devices based on the suitability determined for each of the plurality of devices, and instruct at least one of the subset of devices to act as a device for transmitting, receiving, and / or processing the radio frequency signals for presence and / or location detection. Summary of the Invention

[0007] One object of the present invention is to provide a network control device, a network including the network control device, a network control method, and a computer program that allow for improved sensing reliability of radio frequency sensing networks without substantially increasing the necessary processing resources.

[0008] A network control device for controlling a network including network devices performing radio frequency sensing is provided in a first aspect of the invention, wherein the network includes sensing regions defined by network devices of the network, network devices being assigned to each sensing region such that each of the sensing regions corresponds to a spatial domain, and wherein radio frequency sensing is performed independently of radio frequency sensing in other sensing regions in each sensing region. The network control device includes: a) a sensing result providing unit for providing sensing results of radio frequency sensing in each of the sensing areas; b) a potentially unreliable detection result determining unit for determining, based on a comparison between sensing results of at least two sensing areas, whether a detection result determined based on the sensing results is potentially unreliable in a spatial field that is part of at least one of the sensing areas and / or adjacent to at least one of the sensing areas; and c) a new sensing area defining unit for defining a new sensing area by assigning network devices of the network to the new sensing area to perform radio frequency sensing in the new sensing area, wherein the new sensing area corresponds to at least a part of a spatial field in which the detection results are potentially unreliable, and wherein radio frequency sensing in the new sensing area is independent of radio frequency sensing in other sensing areas, wherein the sensing result providing unit is adapted to provide new sensing results of radio frequency sensing in the new sensing area, and wherein a detection result is determined for at least a part of a spatial field in which the detection result is determined to be potentially unreliable based on at least one sensing result and the new sensing result.

[0009] Since the potentially unreliable detection result determination unit determines the potentially unreliable detection result, and the new sensing region definition unit defines a new sensing region for performing radio frequency sensing in at least a portion of the spatial domain where the detection result is potentially unreliable, additional information about the region with potentially unreliable detection results can be provided without increasing the necessary processing load. Furthermore, since the detection result can be determined based on the additional information from the new sensing region—that is, based on the new sensing result—and further based on the sensing result of the original sensing region, the detection result becomes more reliable. Therefore, the network control device allows for improved sensing reliability of the radio frequency network without increasing processing resources.

[0010] The network includes network devices suitable for performing radio frequency (RF) sensing. Specifically, the network includes multiple network devices distributed across a spatial domain, enabling these network devices to perform RF sensing within that spatial region. However, in addition to network devices suitable for performing RF sensing, the network may also include network devices not suitable for performing RF sensing. Typically, network devices can be any device that includes a network device communication unit and therefore network capabilities. The network device communication unit is adapted to receive and transmit wireless signals, particularly radio frequency (RF) signals and / or wired signals. For example, the network device communication unit may include a network device transceiver for receiving and transmitting RF signals, or a transmitter for transmitting RF signals and a receiver for receiving RF signals. Besides network capabilities, i.e., in addition to the network device communication unit, the network devices may be adapted to perform other functionalities, preferably lighting functionalities. A network can be formed between network devices by utilizing any known network protocol, such as Wi-Fi, Zigbee, Thread, Bluetooth, etc.

[0011] A network control device is suitable for controlling a network by controlling at least some of a plurality of network devices. The network control device can be considered part of the network, particularly as communicating with at least some of the network devices. For example, the network control device can be a standalone device including communication capabilities, for example, in the form of a communication unit, such that the network control device can communicate with at least some of the network devices using radio frequency signals. However, the network control device can also communicate with at least some of the network devices without being directly part of the network; for example, the network control device can include direct connections to some network devices for sending signals to control the network devices, without being part of the network itself. The network control device can also be provided as part of one of the network devices, wherein in this case, it is preferred that the network control device is integrated with the network device and is adapted to use the communication unit of the network device for communicating with other network devices to control other network devices. Furthermore, the network can include more than one network control device, wherein in this case, different network control devices are adapted to communicate with each other to control the network, such that in this case, the network control device can be considered as a single network control device distributed across different physical entities. Furthermore, the network control device itself can be distributed across multiple different independent or network devices, where the independent devices are not directly associated with the network but can communicate with the network, for example, via a server, gateway, or proxy. For instance, the sensing result providing unit of the network control device can be part of one network device, while the new sensing area defining unit can be an independent device or part of another network device. The network control device is then constructed through the communication between network devices or independent devices, including the distributed characteristics of the network control device, and / or with the network.

[0012] The network comprises at least two sensing regions, each defined by a network device assigned to it. Assigning network devices to sensing regions can also be viewed as grouping network devices, where each device is assigned to a group, and each group performs radio frequency (RF) sensing within a sensing region. A sensing region for a group of network devices corresponds to a spatial domain in which the network devices can perform RF sensing. The spatial domain corresponding to a sensing region is defined as an area in which the network device assigned to the sensing region can provide a signal that allows the network to perform its detection task with sufficient reliability. For example, a sensing region can be defined such that a signal, for example, provided by at least one of the network devices and affected by an object such as a person, can still be detected with sufficient reliability by at least one network device assigned to that sensing region, for example, if the changes in the intensity, phase, or amplitude of the RF signal are strong enough to be compared with a reference RF pattern. Typically, a sensing region is a three-dimensional domain in space; however, it can also be interpreted as a two-dimensional projection or two-dimensional slice of the three-dimensional domain relative to an existing or virtual horizontal plane in the spatial domain (e.g., the floor, ceiling, a plane passing through the center of a room, etc.). The sensing region for a group of network devices is physically determined by, for example, the spatial distribution of the group of network devices, sensing sensitivity, and the directionality of RF sensing. In most cases, the sensing area is roughly defined by the spatial distribution of network devices assigned to it. However, in some cases, the sensing area may deviate from the spatial distribution of network devices. For example, in some situations, a group of network devices may include a "blind spot" directly beneath the network devices. Typically, the sensing area can be estimated with sufficient accuracy by considering the known transmission characteristics of the network devices' radio frequency signals, particularly their transmission power and radiation patterns, the known locations of the network devices, and the known layout of the surrounding space, for example, using information about wall locations. However, for the purposes of this invention as defined herein, a precise definition of the sensing area's boundaries is not necessary; it is sufficient to estimate approximate boundaries, for example, through experimentation or experience. For instance, it is sufficient if, for example, it can be determined through experimentation or experience, that a person in a certain area is detected as part of at least two sensing areas, indicating that at least two sensing areas overlap, or if a person in a certain area is not detected in any sensing area, indicating that no sensing area corresponds to the location of the person's presence, and a network blind spot can be found in that area. Furthermore, the effect of assigning one or more network devices to the sensing area can also be estimated based on such experiments, experience, or generally known assumptions.For example, experiments show that a sufficiently reliable sensing can be performed within a certain range around each pair of network devices contributing to the sensing is a good estimate, where the sensing area can then be roughly determined by providing that area around all network devices contributing to the sensing of the network devices assigned to the sensing area. However, other possibilities for estimating the sensing area can also be utilized.

[0013] The sensing area is further defined as performing radio frequency (RF) sensing in each area independently of RF sensing in other sensing areas. In this context, independent execution of RF sensing specifically refers to independent processing of RF sensing in different sensing areas. For example, a network device may be assigned to two sensing areas to provide and receive RF signals from other network devices assigned to the respective sensing areas. However, the network device independently processes the signals received from the two sets of network devices; that is, processing performed on RF signals belonging to the first sensing area has no effect on processing RF signals belonging to the second area. Therefore, each sensing area provides sensing results or sensing outcomes independent of other sensing areas.

[0014] The sensing result providing unit is adapted to provide sensing results of radio frequency sensing in each sensing area. The sensing result providing unit may be, for example, a storage unit storing the sensing results of each sensing area, or may be connected to a storage unit storing the sensing results of each sensing area. The sensing result providing unit may also be directly connected to at least one network device in each sensing area and receive sensing results directly from each of these network devices, wherein the sensing result providing unit is then adapted to provide the sensing results. The sensing result refers to the result of radio frequency sensing performed in the corresponding sensing area, and may refer to one or more quantities indicating predetermined events that can be detected in the radio frequency signal. Based on the task given to the network, these events, and therefore the sensing results, may indicate, for example, the number of people in the sensing area, the presence or absence of one or more people in the sensing area, respiratory movements of people in the sensing area, fall detection in the sensing area, changes in the position of objects or people in the sensing area, etc. In a preferred embodiment, the sensing result for each sensing area indicates the number of people detected as being present in the sensing area.

[0015] The potentially unreliable detection result determination unit is adapted to determine whether a detection result is potentially unreliable in a spatial area that is part of at least one of at least two sensing areas and / or adjacent to it. Typically, the detection result itself is determined based on sensing results provided by the sensing areas. Based on a task assigned to at least one of the network and / or network devices, the detection result may then relate, for example, an indication of the number of people present in different areas of a room, an indication of the presence or absence of a person or object in a specific area of ​​the room, an indication of a location where a fall has been detected, an indication of an area where no one is present, etc. However, in some cases, detection results in a spatial area may be unreliable.

[0016] The potentially unreliable detection result determination unit is adapted to determine whether a detection result is potentially unreliable in a spatial domain based on a comparison between sensing results from at least two sensing areas. For this comparison, logical rules can be applied, such as logical expectations regarding sensing results in different sensing areas. For example, a first sensing area covers the only entrance to a room, and a second sensing area covers the domain of the room excluding the only entrance. In this case, if the second sensing area determines the sudden presence of a person as a sensing result, while the first sensing area does not detect a person as a sensing result, the potentially unreliable detection result determination unit can be adapted to determine that the sensing results in the second and / or first sensing areas are potentially unreliable, and therefore also detection results based on unreliable sensing results. Typically, the logical rules applied to the comparison of sensing results to determine whether a detection result based on sensing results is potentially unreliable can refer to predetermined logical rules based on past experience, the network environment, the distribution of sensing areas, the detection result to be achieved (i.e., the task given to the network), the known reliability of the sensing results, etc. Furthermore, the potentially unreliable detection result determination unit can also be adapted to learn from presented past cases when a detection result is potentially unreliable. For example, a dataset containing multiple previous sensing results of a sensing area and indicating which of the sensing results have led to reliable detection results and which of the sensing results have led to unreliable detection results can be adapted to learn rules that should be applied to the determination.

[0017] If the potentially unreliable detection result determination unit determines, based on a comparison between sensing results from at least two sensing areas, that the detection result is not potentially unreliable (i.e., likely reliable), then the definition of the new sensing area and the further determination of the detection result based on the new sensing result, as described below, can be omitted. Conversely, a network utilizing a detection result determination unit provided, for example, in a control device, can be adapted to determine the detection result directly based on the sensing results.

[0018] The new sensing area definition unit is adapted to define a new sensing area by assigning network devices of the network to the new sensing area for performing radio frequency sensing in the new sensing area. For example, network devices previously assigned to one of the original sensing areas can be assigned to the new sensing area by the new sensing area definition unit. However, network devices that were not previously assigned to any sensing area can now also be assigned to the new sensing area. The new sensing area definition unit is adapted to assign network devices such that the new sensing area corresponds to at least a portion of a spatial region where the detection results are potentially unreliable. The new sensing area definition unit can be adapted to assign network devices, for example, based on the known spatial distribution of network devices near the spatial region where the detection results are potentially unreliable, based on the known sensing characteristics of the network devices (which indicate that they can cover the spatial region where the detection results are potentially unreliable), etc. Furthermore, the new sensing area definition unit can be adapted to assign network devices based on predetermined rules. These rules can be predetermined and provided to the new sensing area definition unit, for example, by the user during installation or after working with the network for a period of time. The rules can then be based on the user's experience with the detection results provided by the network. For example, if a user determines that detection results are frequently unreliable in a certain area, the user can provide rules to the new sensing area definition unit to assign network devices to a new network. These network devices are specifically selected by the user to cover the area where the most unreliable detection results occur. Furthermore, to select network devices for defining a new sensing area, the new sensing area definition unit may be adapted to utilize one or more of the following information: a) a layout / plan of the area where the network is installed, b) a network neighbor table provided by the network, c) signal characteristics of the network devices indicating their spatial relationships, particularly signal strength and / or signal quality, and / or information about which network device contributes most to the sensing results in the sensing area.

[0019] Then, the network device assigned to the new sensing area can perform radio frequency sensing in the new sensing area independently of radio frequency sensing in other sensing areas. In this context, independent sensing also refers to the independent processing of radio frequency signals, that is, the sensing results in other sensing areas have no impact on the sensing results in the new sensing area.

[0020] The sensing result providing unit is then also adapted to provide new sensing results of radio frequency sensing in the new sensing area. Therefore, specific information about the spatial domain where the overall detection results may be unreliable can be provided. To better distinguish between the sensing results provided for the new sensing area (i.e., the new sensing results) and the sensing results provided for the sensing area defined before the new sensing area was defined, the latter sensing result will sometimes be referred to as the original sensing result, and the corresponding sensing area as the original sensing area, in the following text.

[0021] Then, based on the new sensing result and at least one sensing result from the original sensing area, the detection result is determined, at least for the potentially unreliable detection result determination unit, which has already determined a portion of the spatial domain of the potentially unreliable detection result. For example, the detection result can be determined by a dedicated detection result determination unit that is part of the network or network control device. However, the detection result can also be determined by any other unit of the network or network control device, or by a unit that is not itself part of the network or network control device but communicates with the network or network control device, such that the detection result is determined as described above. In a preferred embodiment, such a detection result determination unit is part of the network control device and is adapted to determine the detection result based on the new sensing result and at least one original sensing result, and optionally also based on the original sensing result, for example, if the potentially unreliable detection result determination unit determines that there is no potentially unreliable detection result.

[0022] To determine the detection result, information provided by the new sensing result can be considered to make the detection result reliable. For example, logical rules based on the task provided to the network can be applied to the sensing results of the original sensing area and the new sensing result of the new sensing area to determine the reliable detection result. For example, if the sensing result of the first original sensing area indicates the presence of a person in the spatial domain, while the sensing result of the second original sensing area indicates the absence of a person in the same spatial domain, the conflict can be resolved based on the new sensing result, which relates to a sensing area that specifically covers the spatial domain where the conflict has been determined.

[0023] In one embodiment, at least two spatial fields corresponding to the sensing area overlap, enabling at least two sensing results to be provided for the overlapping portion, and / or a portion of the spatial field in which network devices are distributed does not correspond to the sensing area, making it impossible to provide sensing results for the uncovered portion, wherein the portion of the spatial field in which potentially unreliable detection results are determined corresponds at least partially to the overlapping portion and / or the uncovered portion, respectively.

[0024] For example, if the network devices assigned to sensing areas are located near each other, such that a signal indicating the presence of an object, based on a selected sensitivity of radio frequency sensing, can still be determined by at least one network device in both sensing areas, this may result in an overlap of the spatial domains corresponding to the sensing areas. This causes people appearing in the overlapping area of ​​the two sensing areas to be counted by both sensing areas, thus being counted twice. If a detection result, such as the total count of all people in the network's area, is determined based on these sensing results, the result will be unreliable, at least for the overlapping area. In the case of overlapping sensing areas, the new sensing area defining unit can be adapted to define a new sensing area by assigning network devices to the new sensing area, such that at least a portion of the overlapping area is covered by the new sensing area. For example, the new sensing area defining unit can select network devices near the overlapping area to assign them to the new sensing area to cover at least a portion of the overlapping area with the new sensing area. Regarding the example given above where people appear in the overlapping area, the new sensing result of the new sensing area can then provide, for example, a result where only one person appears in the overlapping area, where this new sensing result can then be taken into account when determining the detection result.

[0025] Overlapping areas can typically be avoided by reducing the sensitivity of radio frequency sensing in the sensing area. However, this reduction in sensitivity can, in some cases, result in the inability to provide sensing results for the spatial domain between at least two sensing areas—that is, areas in the spatial domain that generate the network distribution that are not covered by at least one sensing area. Therefore, for example, if the network includes uncovered areas, a person present in an uncovered area will not be sensed in any sensing area of ​​the network. Thus, in this case, the detection result based on the sensing results of all sensing areas will again be unreliable. Therefore, also for this situation, a new sensing area defining unit can be adapted to define a new sensing area such that previously uncovered areas are now at least partially covered by the new sensing area, for example, by assigning network devices to new sensing areas near the uncovered areas.

[0026] In one embodiment, comparing sensing results between at least two sensing regions includes determining the temporal correlation and / or activity fingerprint of the at least two sensing results. For example, comparing sensing results includes determining the temporal correlation, which may include determining at what point in time an event such as the presence of a person, fall detection, or breathing event has been detected in one of the at least two sensing regions, and may further include comparing that point in time with a point in time when a similar event has been detected in the other of the at least two sensing regions. If similar events have been detected at different points in time, the detection is unlikely to be caused by the same event. However, if similar events are detected at substantially the same point in time, the potentially unreliable detection result determination unit may be adapted to determine that the detection in the two sensing regions is likely the result of the same event that has already occurred in the overlapping region. In this case, the potentially unreliable detection result determination unit may be adapted to determine that the detection result determined based on these sensing results is potentially unreliable. However, the temporal correlation of the sensing results may also be used to determine the time difference between events detected in the sensing regions or whether events were detected in the sensing regions in a specific order, and this time difference may be used to determine potentially unreliable detection results. For example, if the sensing result refers to an event where a person leaves the sensing area, and the sensing results of adjacent sensing areas do not indicate the presence of the corresponding person, or indicate that the person entered at a much later time, the unreliable detection result determination unit may be adapted to determine that the detection result is unreliable, for example, because it is inferred from the above example that there is an uncovered area between the two sensing areas.

[0027] Additionally or alternatively, the comparison between sensing results may also include determining the activity fingerprints of the two sensing results. An activity fingerprint may refer, for example, to determining a signal that changes over time, indicating the activity of a person present in the sensing area at that moment. For example, a person simply sitting at a desk working on a computer will display a different activity fingerprint, i.e., another signal indicating his / her activity, compared to a person walking around or standing somewhere chatting with another person. The activity fingerprints determined based on the sensing results of the sensing areas can then be compared. For example, if the fingerprints are very different from each other, they are unlikely to have been generated by the detection of the same person, while if the activity fingerprints are substantially similar, it is likely the result of the detection of the same person by at least two sensing areas, e.g., a person appearing in an overlapping area. Therefore, the potentially unreliable detection result determination unit may be adapted to determine that a detection result is potentially unreliable if the activity fingerprints determined in at least two sensing areas (preferably at least two adjacent sensing areas) are substantially similar. Furthermore, the potentially unreliable detection result determination unit may be adapted to determine a similarity measure based on the comparison of the sensing results of at least two sensing areas, and may determine whether a detection result determined based on the sensing results is potentially unreliable based on the determined similarity measure, e.g., whether the determined similarity measure is higher than a predetermined threshold.

[0028] In one embodiment, the new sensing area definition unit is adapted to assign network devices to a new sensing area based on known characteristics of the network. Preferably, in one embodiment, the known characteristics of the network include information indicating the spatial distribution of network devices, the current state of network devices, environmental conditions of network devices, radio noise, interference, and / or the general capabilities of network devices, and the new sensing area definition unit is adapted to assign network devices to the new sensing area based on the indicated spatial distribution, current state, environmental conditions, radio noise, interference, and / or the capabilities of network devices. More preferably, the spatial distribution of the network includes information indicating the relative and / or absolute locations of network devices in the network, wherein the new sensing area definition unit is then adapted to determine which network devices should be assigned to the new sensing area based on the location information of the network devices. Preferably, the new sensing area definition unit is adapted to assign network devices near areas where detection results are potentially unreliable to the new sensing area. Furthermore, the network characteristics may also include information about the state of network devices, information about the types of network devices present in the network, information about the radio frequency sensing sensitivity of network devices, information about the radio frequency sensing directionality between network devices, etc., wherein the new sensing area definition unit may also be adapted to assign network devices to the new sensing area based on the above information. In addition, network characteristics may include environmental conditions of the network devices, such as temperature and humidity, and wireless noise near the network devices, such as interference from other Wi-Fi devices and / or known interference, such as the presence of animals, or windows that may cause airflow and thus move objects such as curtains. This information can also be taken into account when assigning network devices to new sensing areas.

[0029] In one embodiment, the new sensing area defining unit is adapted to allocate network devices to a new sensing area based on previous detection results and / or previous sensing results and / or expectations of sensing results and / or detection results. For example, previous detection results may include information about the number of people present in the area during previous detection, where the new sensing area defining unit may be adapted to allocate network devices to the new sensing area based on the number of people in the sensing area. For example, if previous detection or sensing results indicate that the number of people in an area exceeds a certain threshold, the new sensing area defining unit may be adapted to allocate fewer network devices to the new sensing area compared to the case where the number of people in the area is below a certain threshold, because detecting more people is easier than, for example, detecting only one person sleeping in the area. Furthermore, in one embodiment, the new sensing area defining unit may be adapted to decide not to allocate network devices to a new sensing area based on previous detection results and / or previous sensing results and / or expectations of sensing results and / or detection results, i.e., omitting the provision of a new sensing area, even though the detection results may be potentially unreliable. For example, if the detection results and / or previous sensing results indicate that the detection area is very crowded, it may not be necessary to know the exact number of people, so that even if the number is unreliable, it may not interfere with the task of providing network devices. Conversely, when only a few people are distributed in the detection area, reliable human detection is more desirable. Expectations for sensing and / or detection results can be based on, for example, past experience or provided knowledge about the environment in which the network is installed. For instance, if the network has identified a room as being occupied by one person at certain times of the day and by more than one person at other times over the past few days, this expectation can be taken into account when defining new sensing areas. Furthermore, this information can be considered if it is known in advance that the network equipment is installed in an office environment where a certain number of people work at their desks. Additionally, expectations for sensing and / or detection results can also be based on sensing and / or detection results in other spatial areas including similar environments and similar networks containing network equipment. For example, if an office building includes open-plan offices where the spatial arrangement of network equipment and desks is repeated within rooms or even on different floors, the sensing and / or detection results of networks installed in one of the repeated office arrangements in the building can also be used to determine expectations. The potentially unreliable detection result determination unit may also consider information about previous and / or sensing results and / or expected detection and / or sensing results, for example, by comparing at least two sensing results with previous and / or expected sensing results to determine potentially unreliable detection results.

[0030] In one embodiment, the potentially unreliable detection result determination unit is adapted to further determine a confidence level for the determined detection result for each portion of the spatial domain formed by the sensing area based on the sensing result, wherein the new sensing area definition unit is adapted to define a new sensing area based on the confidence value. For example, the potentially unreliable detection result determination unit may be adapted to compare the sensing result of an area with the expected sensing result of that area, and based on the result of the comparison, may then be adapted to determine the confidence level of the sensing area. In another example, the potentially unreliable detection result determination unit may be adapted to take into account knowledge of the overall confidence level of the sensing results of network devices in the sensing area and / or the network. For example, if it is known that a network device assigned to the sensing area provides unreliable results, for example due to generally low RF sensing sensitivity, the sensing results of the sensing area to which the unreliable network device is assigned may be given a low confidence level, which indicates a low confidence level in the reliability of the sensing results. Furthermore, the potentially unreliable detection result determination unit may be adapted to provide a low confidence level only to the portion of the sensing area contributed by the unreliable network device, while other portions of the sensing area may include a higher confidence level. Therefore, the confidence level of the overall sensing results for the sensing area can be provided based on whether potentially unreliable network devices contribute to the sensing results.

[0031] In one embodiment, the sensing result refers to at least one physical event, and a comparison is performed between the sensing results of at least two sensing areas within the sensing area to determine whether the same physical event has been detected at least twice and / or whether the physical event has been missed by one of the sensing areas. Preferably, the determination of whether the same physical event has been detected twice and / or whether the physical event has not yet been detected by one of the sensing areas is based on predefined rules applied to the comparison of the sensing results. The predefined rules may be provided by the user, for example, during network installation, or may be learned by the network, for example, based on known machine learning algorithms. These rules may take into account, for example, knowledge about the environment in which the network is installed, knowledge about the spatial area covered by the sensing areas of the network, knowledge about the sensing sensitivity of the network devices assigned to the sensing areas, etc. For example, if it is known that two sensing areas are adjacent to each other, and one sensing area covers the only door of the room, a rule can be applied that if the sensing area covering the room door has not previously provided a predetermined event, the sensing result of the sensing area not covering the door is potentially unreliable. This rule would cover the logic that a person must pass through the door to enter the room, and therefore must first be sensed in the sensing area covering the door before they can be sensed in the sensing area away from the door. However, entirely different rules can also be applied, which can be determined and learned by the network itself when processing the sensing results after installation.

[0032] In one embodiment, the new sensing region defining unit is adapted to define a new sensing region such that the sensing sensitivity in the new sensing region is lower than the sensing sensitivity in other sensing regions, and / or to make the new sensing result refer to another type of sensing result. Specifically, the sensing sensitivity of the new sensing region can be determined by the new sensing region defining unit based on a comparison of at least two sensing results provided by the potentially unreliable detection result determination unit. For example, if the comparison indicates that the detection result is potentially unreliable due to counting people present in the overlapping area twice, the new sensing region defining unit can be adapted to reduce the sensitivity of the new sensing region relative to the sensitivity used in other sensing regions (i.e., the original sensing region), since the simple presence or absence of a person may be easily determined with lower sensitivity. However, if the detection result and therefore the task of the network involves more subtle detections, such as respiration detection, the sensitivity in the new sensing region can be provided to be the same as the sensitivity in other sensing regions. Furthermore, in one embodiment, the new sensing region defining unit can also be adapted to define a new sensing region such that the sensing sensitivity in the new sensing region is higher than the sensing sensitivity in other sensing regions. Specifically, if the new sensing area defining unit determines that the distribution of network devices near the spatial area that the new sensing area should cover is not optimal, such as sparse, the new sensing area defining unit can be adapted to increase the sensitivity of the network devices allocated to the new sensing area, so that the desired spatial area can still be covered even with increased sensitivity. Therefore, the new sensing area defining unit can also be adapted to determine the sensitivity of the new sensing area based on known network characteristics (e.g., spatial distribution of network devices).

[0033] In one embodiment, the new sensing region defining unit is adapted to define a new sensing region such that the sensing sensitivity in the new sensing region is higher than the sensing sensitivity in other sensing regions to determine a specific sensing outcome. For example, as a sensing outcome, determining the breathing pattern of a person present in the sensing region requires a higher sensitivity than determining if a person has fallen in the sensing region. The new sensing region defining unit is accordingly adapted to determine whether the new sensing outcome needs to involve the same task as the overall detection outcome, and to determine the sensitivity based on the sensing outcome that should be provided by the new sensing outcome. For example, if the overall task of the network is to determine the number of people in an area requiring high sensing sensitivity as a detection outcome, the new sensing region defining unit may be adapted to determine, based on a comparison provided by a potentially unreliable detection outcome determining unit, that for the new sensing region, it is only necessary to simply determine whether a person is currently moving in that area, where, for this task, the sensing sensitivity of the new sensing region can be reduced compared to the sensing sensitivity of other sensing regions.

[0034] In one embodiment, the new sensing region defining unit is adapted to assign network devices to new sensing regions such that radio frequency (RF) sensing is not performed in the sensing regions previously assigned to the network devices, as long as RF sensing is performed in the new sensing regions. After the RF sensing performed in the new sensing regions is completed, the new sensing region defining unit is adapted to reassign the network devices to their previously assigned sensing regions. Therefore, in this embodiment, the new sensing regions only exist for the amount of time required to perform their task of providing new sensing results. For example, the new sensing region defining unit can be adapted such that the new sensing regions exist only for a predetermined period of time, and after the end of that period, the new sensing regions gradually change back to the previously existing sensing regions. For example, the new sensing region defining unit can be adapted to store information about assigning network devices to sensing regions, for example, in the form of a list, before assigning network devices to new sensing regions, and can then be adapted to dissolve the new sensing regions and assign the network devices to their original sensing regions based on the stored information. Furthermore, the new sensing area definition unit can also be adapted to store the allocation of network devices to the new sensing area, for example, in the form of a list. This allows the new sensing area definition unit to allocate network devices to the new sensing area based on the stored information if the same spatial area requires new sensing results again, thus eliminating the need to determine which network devices are near the spatial area. This has the advantage of making the allocation of network devices more efficient, faster, or less resource-intensive.

[0035] In one embodiment, the new sensing area defining unit is adapted to control the network devices allocated to the sensing area and the new sensing area such that, whenever radio frequency sensing is performed in the new sensing area, the amount of processing and / or the total communication load of the network devices allocated to the new sensing area and the sensing area remains the same or lower compared to the total amount of processing and / or the total communication load performed before the definition of the new sensing area, and / or wherein the new sensing area defining unit is adapted to control the network devices allocated to the sensing area and the new sensing area such that, whenever radio frequency sensing is performed in the new sensing area, the time not spent on sensing by the network devices remains the same or higher compared to the time not spent on sensing before the definition of the new sensing area. An advantage of this embodiment is that radio frequency sensing can be performed simultaneously in the sensing areas (i.e., the original sensing area and the new sensing area). Furthermore, the new sensing area can remain active for a longer period without requiring higher processing power or providing a higher communication load to the network devices, which could lead to performance losses in the network, such as time delays when turning on lights based on network detection results. In one example, if the radio frequency (RF) signal transmitted and / or processed by the network device remains constant or is reduced, the amount of processing and / or total communication load performed by the network device in one of the original sensing regions, the new sensing region and the original sensing region, can be reduced or kept the same. To this end, the new sensing region defining unit can be adapted to control the network devices allocated to the sensing region and the new sensing region such that the possible processing communication volume is divided between the new sensing region and the sensing region to which the network device is allocated. For example, the communication volume or processing volume utilized by the sensing region can be reduced by half, so that the remaining amount can be used for RF sensing in the new sensing region. However, other ratios for dividing the performed processing volume or total communication load, such as 2:1, 3:1, etc., can also be used. The same principle as described above can be applied to embodiments in which the new sensing region defining unit is adapted to control the network devices allocated to the sensing region (i.e., allocated to the original sensing region) and the new sensing region such that the time not spent on sensing in the new sensing region and the original sensing region remains the same or becomes longer. In particular, the time not spent on sensing can be determined by the amount of processing or communication load used for RF sensing. However, the time not spent on RF sensing can also be determined by other factors, in which the new sensing area definition unit is then adapted to also control these other factors.

[0036] In another aspect of the invention, a network is proposed, wherein the network comprises a) a plurality of network devices distributed in a spatial domain and adapted to perform radio frequency sensing, and b) a network control device according to any one of the preceding claims.

[0037] In another aspect of the invention, a network control device is proposed for controlling a network including a network method for performing radio frequency (RF) sensing, wherein the network includes sensing regions defined by network devices of the network, network devices being assigned to each sensing region such that each of the sensing regions corresponds to a spatial domain, and wherein RF sensing is performed independently of RF sensing in each sensing region. The network control method includes: a) providing sensing results of RF sensing in each of the sensing regions; b) determining, based on a comparison between sensing results of at least two sensing regions, whether a detection result determined based on the sensing results is potentially unreliable in a spatial domain that is part of at least one of the sensing regions and / or adjacent to at least one of the sensing regions; c) defining a new sensing region by assigning network devices of the network to a new sensing region to perform RF sensing in the new sensing region, wherein the new sensing region corresponds to at least a portion of a spatial domain in which the detection result is potentially unreliable, and wherein RF sensing in the new sensing region is independent of RF sensing in other sensing regions; d) providing new sensing results of RF sensing in the new sensing region; and e) determining a detection result for at least a portion of the spatial domain in which the detection result is determined to be potentially unreliable, based on at least one sensing result and the new sensing result.

[0038] In another aspect of the invention, a computer program product for controlling a network is provided, wherein the computer program product includes program code means that causes the network control device described above to perform the network control method described above.

[0039] It should be understood that network control devices, networks, network control methods, and computer program products have similar and / or equivalent preferred embodiments, specifically as defined in the dependent claims.

[0040] It should be understood that the preferred embodiments of the present invention may also be any combination of the dependent claims or the above embodiments and their respective independent claims.

[0041] These and other aspects of the invention will become apparent and will be explained with reference to the embodiments described below. Attached Figure Description

[0042] In the following figures:

[0043] Figure 1 The network and the network control equipment used to control the network are illustrated schematically and demonstratively.

[0044] Figure 2 The diagram illustrates and demonstrates a network control method for controlling the network, and

[0045] Figure 3A -C、 Figure 4and Figure 5 The working principle of some exemplary embodiments is illustrated schematically and demonstratively. Detailed Implementation

[0046] Figure 1 Network 100 and network control device 130 for controlling network 100 are illustrated schematically and exemplary. Network 100 includes a plurality of network devices 111, 121. In a preferred embodiment, network 100 refers to a lighting system, and therefore at least some of the network devices 111, 121 also include lighting functionality in addition to network communication and radio frequency (RF) functionality. More preferably, network 100 refers to a system that allows control of additional functionality (e.g., lighting functionality) of network devices 111, 121 based on RF sensing results and / or detection results determined based on those sensing results. Network devices 111, 121 are adapted to perform RF sensing using RF signals 112, 122 transmitted and received by network devices 111, 121. Typically, network 100 can be any network including network devices adapted to perform RF sensing, wherein network 100 can be based on, for example, Wi-Fi communication protocols, Zigbee communication protocols, Bluetooth communication protocols, etc. Network devices 111, 121 of network 100 are assigned to sensing areas 110, 120. In this exemplary configuration, network devices 111 are all assigned to sensing area 110, and network devices 121 are all assigned to sensing area 120. Sensing areas 110 and 120 each correspond to a spatial domain for which the network devices 111 and 121 assigned to the respective sensing area can perform radio frequency sensing with a suitable, preferably predetermined, sensitivity. In this embodiment, the sensing areas 110 and 120 corresponding to the two spatial domains overlap in an overlapping domain 140.

[0047] Typically, network 100 is given one or more tasks related to radio frequency sensing and is adapted to provide detection results based on sensing results of sensing areas 110, 120 corresponding to a given task. The task may be, for example, counting people present in areas of network 100, detecting presence in at least a portion of area 100, detecting breathing in at least a portion of area 100, detecting falls in at least a portion of area 100, etc. In the example below, the task of network 100 and the detection results thus provided by network 100 involve determining the number of people in different parts of areas of network 100. Therefore, the number of people in these sensing areas is also provided as a result of radio frequency sensing in sensing areas 110, 120. Typically, network 100 is then adapted to determine the number of people in different areas of network 100, for example, by utilizing a detection result determination unit based on the sensing results of sensing areas 110, 120, which may be part of network 100, for example, a network device of the network, a server connected to the network, or some other dedicated hardware connected to the network. Furthermore, such a detection result determination unit may also optionally be part of network control device 130.

[0048] Network control device 130 may be considered part of network 100, or it may not be considered part of network 100 but communicate with at least one network device of network 100. Network control device 130 is adapted to control network 100, particularly by controlling one or more network devices 111, 121 of network 100. To perform control of network devices 111, 121, network control device 130 is adapted to communicate with at least one of network devices 111, 121, for example, using radio frequency communication signal 135. However, network control device 130 may also use any other communication means to communicate with at least one of network devices 111, 121, for example, wired communication, optical communication, ultrasonic communication, etc. Furthermore, network control device 130 may also be part of at least one of network devices 111, 121, for example, by hard-wired connection to the network device or by being housed in the same enclosure as the network device. In this case, network control device 130 can use the communication possibilities of the network device providing it to control network 100.

[0049] The network control device 130 includes a sensing result providing unit 131, a potentially unreliable detection result determining unit 132, a new sensing area defining unit 133, and optionally also includes... Figure 1 The detection result determination unit is not shown in the figure.

[0050] The sensing result providing unit 131 is adapted to provide sensing results for each of the sensing areas 110, 120. Specifically, in one example, the sensing result providing unit 131 is adapted to provide the number of people present in the spatial domain corresponding to each sensing area 110, 120, detected by radio frequency sensing of network devices assigned to the sensing areas. The sensing result providing unit 131 may, for example, communicate with at least one network device 111, 121 of each sensing area 110, 120 to receive the sensing results. However, the sensing result providing unit 131 may also be adapted to communicate with only one network device, wherein the network device receives the sensing results of all sensing areas 110, 120 via network communication of network 100. Furthermore, the sensing result providing unit 131 may also be adapted to communicate with a server or another control device that communicates with at least one of the network devices 111, 121 and is adapted to forward the sensing results of the sensing areas 110, 120 to the sensing result providing unit 131. The sensing result providing unit 131 is then adapted to provide the sensing result to at least the potentially unreliable detection result determining unit 132.

[0051] The potentially unreliable detection result determination unit 132 is adapted to determine whether a detection result determined based on the sensing result provided by the sensing result providing unit 131 is potentially unreliable. In particular, the potentially unreliable detection result determination unit 132 is adapted to compare the sensing results of at least two sensing regions 110, 120 to determine whether the sensing results may potentially lead to unreliable detection results, for example, in the overlapping region 140.

[0052] For example, if a person is present in the overlapping region 140, both network device 111 assigned to sensing region 110 and network device 121 assigned to sensing region 120 can detect the person present in the overlapping region 140. If this is indeed the case, the sensing results of the two sensing regions 110, 120 will be that one person is present in each sensing region 110, 120. Therefore, in this case, network 100 will determine, for example, that two people are present in the network region, or whether the sensing results include an indication that two people are present in the overlapping region 140 in the region where that person is present. To avoid these erroneous, i.e., unreliable detection results from network 100, in this example, the potentially unreliable detection result determination unit 132 can be adapted such that the comparison between the sensing results of sensing region 110 and sensing results of sensing region 120 includes determining the temporal correlation of the sensing results. In the example above, such temporal correlation indicates that the person detected in each sensing region of sensing regions 110, 120 was detected by network devices 111, 121 assigned to each sensing region of sensing regions 110, 120 at the same time. Therefore, in this example, the potentially unreliable detection result determination unit 132 can be adapted to determine, for example, that the temporal correlation of sensing results exceeding a predetermined threshold indicates a potentially unreliable detection result. However, the potentially unreliable detection result determination unit 132 can also use other measures to compare sensing results and determine that the detection results are potentially unreliable. For example, the potentially unreliable detection result determination unit 132 can be adapted to compare sensing results in a sensing area using activity fingerprints, predetermined knowledge, and / or network characteristics, such as the distribution of network devices, past experience with network results, etc.

[0053] In one example, the potentially unreliable detection result determination unit 132 may include a trained artificial intelligence system, such as a neural network, which is trained using previous sensing results of the sensing area and, for example, detection results marked by the user as reliable or unreliable, such that the artificial intelligence component can determine whether the detection result is potentially unreliable based on the learned rules integrated into the artificial intelligence component.

[0054] Then, the new sensing region definition unit 133 is adapted to define a new sensing region 150, which corresponds to a spatial region or at least a portion of which the potentially unreliable detection result determination unit 132 has determined to be potentially unreliable. For example, in the above example, the new sensing region definition unit 133 may be adapted to define a new sensing region 150 including an overlapping region 140. Specifically, the new sensing region definition unit 133 defines the new sensing region 150 including the overlapping region 140 by assigning corresponding network devices 111, 121 of the network 100 to the new sensing region 150. In the above example, the new sensing region definition unit 133 may be adapted to assign four network devices 111, 121 adjacent to the overlapping region 140 to the new sensing region 150. In a preferred embodiment, the new sensing area definition unit 133 assigns network devices 111 and 121 to the new sensing area 150 based on known characteristics of the network 100, such as the known spatial distribution of network devices 111 and 121, the known relationships between network devices 111 and 121, and the known characteristics of network devices 111 and 121, particularly their sensing characteristics. However, the new sensing area definition unit 133 may also be adapted to assign network devices 111 and 121 to the new sensing area 150 based on a predetermined plan or past experience. For example, a user who knows the layout of the network 100 can determine that, in all cases where the potentially unreliable detection result determination unit 132 determines a potentially unreliable detection result, a particular network device 111 or 121 should be assigned to the new sensing area 150 by the new sensing area definition unit 133. Furthermore, the new sensing region definition unit 133 can be adapted to learn rules based on the training data of the correspondingly provided network 100. These rules allow the new sensing region definition unit 133 to assign network devices 111, 121 to the sensing region 150 based on a spatial domain, for which potentially unreliable detection results are determined without direct knowledge of the network characteristics. However, note that through the learning process, indirect knowledge of the characteristics of the network 100 is integrated into the rules learned by the new sensing region definition unit 133.

[0055] Then, the sensing result providing unit 131 may be adapted, for example, as described above, to provide new sensing results for the new sensing area 150 by communicating with one of the network devices 111, 121 assigned to the new sensing area 150, or in any other manner already described above.

[0056] Then, based on at least one of the sensing results (i.e., the original sensing results) and the new sensing result, a detection result is determined for at least a portion of the spatial domain, for which potentially unreliable detection results are determined, for example, for at least the overlapping domain 140. For example, a logical rule can be applied that allows determining how the original sensing results must be considered when determining the detection result based on the new sensing result. In the above example, which includes a person present in the overlapping domain 140, the new sensing result could indicate that only one person is present in the new sensing area 150. Network 100 (e.g., the detection result determination unit is part of network 100 or optionally part of network control device 130) can then be adapted to apply a logical rule to determine that the two sensing results actually refer to the same person at least in the overlapping portion 140, such that the sensing results for at least that portion of sensing areas 110, 120 cannot be simply added together, but must be corrected based on the new sensing result for the new sensing area 150. Thus, in this case, network 100, for example, determines by the detection result determination unit, that only one person is indeed present in the area of ​​network 100.

[0057] Figure 2A method for controlling a network (such as network 100) as described above is illustrated schematically and exemplary. In a first step 210, method 200 includes providing sensing results of radio frequency sensing in each sensing region of the network (e.g., sensing regions 110, 120 of network 100). As described above, the sensing results may be provided by sensing result providing unit 131. In a next step 220, it is determined whether the detection result determined based on the sensing results is potentially unreliable in a spatial region that is part of at least one of the sensing regions 110, 120 and / or adjacent to it. For example, according to the principles described above, in step 210, by comparing the sensing results of regions 110 and 120, it can be determined that the detection result is potentially unreliable for the overlapping region 140. As described above, the potentially unreliable detection result may be determined, for example, by potentially unreliable detection result determining unit 132. If no potentially unreliable detection result is determined, method 200 may include determining the detection result based on the sensing results, and subsequent steps 230, 240, 250 may be omitted. However, if it is determined that the detection result is potentially unreliable in at least a portion of the area of ​​network 100, method 200 performs steps 230, 240, and 250. In this case, in the next step 230, a new sensing area is defined by assigning network devices 111 and 121 to a new sensing area 150 to perform radio frequency sensing in the new sensing area 150, wherein the new sensing area 150 corresponds to at least a portion of the spatial region where the detection result is potentially unreliable. This step can also be performed according to the principles and examples given above regarding the new sensing area definition unit 133. In the next step 240, new sensing results of radio frequency sensing in the new sensing area 150 are provided, for example, by sensing result providing unit 131. In the final step 250 of method 200, a detection result is determined for at least a portion of the spatial region where the potentially unreliable detection result was determined, wherein the detection result is also determined based on at least one of the sensing results (i.e., the original sensing result) and the new sensing result, according to the principles already described above.

[0058] The following provides more detailed examples of how the invention applies to some frequently occurring scenarios. These examples relate to a space comprising multiple network devices capable of operating radio frequency (RF) sensing and forming a network, where the space is larger than the area that a normal RF sensing area can cover with acceptable performance. Therefore, at least two RF sensing areas are provided, which can partially overlap when set to a desired sensitivity level for reliable event detection in each area, such that events occurring in the partially overlapping area are picked up by both areas. The main idea of ​​the invention then involves determining in this situation whether two events might correspond to the same physical event, for example, determining whether the two events are caused by a single physical event, where each event is detected by one of the sensing areas. Furthermore, the invention proposes to reconfigure (i.e., reassign) the network devices in the space to create new sensing areas and determine the results of RF sensing based on the sensing results in the new sensing areas. The sensing performance of the new sensing areas may be lower than that of the original sensing areas for the original sensing targets of the network. However, for example, for the detection of a specific event at a specific location that triggers such reconfiguration, such as breathing detection, the performance of the new sensing areas may also be higher than that of the original sensing areas.

[0059] Figure 3A The example of the application and operation of the invention shown relates to an office space 300, which includes wireless lighting devices 311 and 312 installed in a regular pattern in the ceiling as network devices 1, 2, 3, 4, 5, 6, 7, and 8 to provide consistently consistent lighting. The network devices are adapted to perform radio frequency sensing to count the number of people in each area, allowing newcomers to be guided to less crowded areas, thus reducing noise and / or increasing the availability of desks. This means that the number of people in each area and across the entire floor needs to be updated rapidly, as it is unknown when a new colleague will need to be guided to a desk. Office space 300 includes two distinct sensing zones 310 and 320. Figure 3A The sensing areas 310 and 320 are shown to overlap. In some cases, it may be desirable to configure adjacent areas such that they do overlap. This will allow the benefits of the invention when potential duplicate counting may occur. However, it must also be considered that over-extending the coverage of the sensing area to increase / create overlap may also lead to undesirable triggering at other edges / adjacent areas of the detection area, such as people being picked up in a corridor. In this example, network devices 1, 2, 3, 4, 5, 6, 7, and 8 are grouped (i.e., assigned) to sensing areas 310 and 320, and their performance parameters are optimized for counting people at their desks. However, at certain specific times, such as entering a room, taking a coffee break, taking a lunch break, returning from a meeting, or standing next to a colleague sitting at a desk chatting, etc., Figure 3BThe diagram illustrates that people can gather in specific small rooms for brief chats and stay there for a few minutes. These areas can be where network devices can still detect lingering individuals and where multiple sensing areas 310 and 320 overlap. For example... Figure 3B As shown, this could result in people in certain areas being counted twice, once for each sensed area. The system might then perceive this as a crowded area and redirect new colleagues to other areas when it is actually a suitable location for them. Furthermore, if the system detects someone lingering in a certain location within the office, it might direct office staff to find a quiet workspace on the appropriate space management app, rather than using that section of the office.

[0060] In this context, the present invention proposes, in a step performed, for example by a potentially unreliable detection result determination unit, to determine that the number of people between regions 310 and 320 is most likely associated with a single, unique event, and that the sensing results will therefore lead to potentially unreliable detection results. For example, this determination can be performed based on noting the temporal correlation of the triggering of the sensing event in the sensing results. Furthermore, it is proposed that, to confirm whether this is a genuine dual event or a single event, the most suitable network device for temporarily performing RF sensing is determined, for example by a new sensing area definition unit. Here, these suitable network devices are network devices 2, 4, 5, and 7. This determination can be based on, for example, an office layout / floor plan, a network neighbor table, analysis of signal strength from nodes to determine proximity, or even after detection to understand which network device is most affected by the event. Then, as... Figure 3C As shown, a new sensing area 330 involving network devices 2, 4, 5, and 7 can be created, where the network devices are ideally suited for counting people in their surrounding space. The original sensing areas 310 and 320 can then be temporarily disabled, for example. Alternatively, the original sensing areas 310 and 320 can continue to provide sensing results, or sensing can be performed alternately between the original sensing areas 310 and 320 and the new sensing area 330. Radio frequency sensing is then performed by the new sensing area 330 for the desired functionality (in this case, counting people) to verify the actual number of people in the area. In this example, the detection result can then be determined, for example, by a detection result determination unit by applying rules based on the new and original sensing results, according to the following scenario:

[0061]

[0062] Once the detection results can be determined with sufficient reliability—for example, after reaching a sufficient level of confidence in the number of people present—the new sensing area 330 can be disabled in the original sensing areas 310 and 320, allowing them to continue their normal operation, or sensing in area 330 can simply be stopped. This depends on which method is used to construct sensing area 330 based on sensing areas 310 and 320. The new sensing results in the new sensing area 330 can then be provided to the system responsible for processing the people count, so that the detection results (here, the total number of people) can be updated if necessary.

[0063] Figure 4 Another embodiment of this scenario is shown. In this case, not only is a new sensing area provided, but two smaller new sensing areas 340 and 350 are also provided in the overlapping area. Therefore, a more detailed impression of where a person is can be obtained. In this example, by assigning only the two network devices 2 and 5, and 4 and 7 shown, to the two new sensing areas 340 and 350 respectively, detection can be focused on the path between the two network devices. In this case, the radio frequency sensing sensitivity can be set low enough that only people at specific locations along the path are detected. In this case, not only can the location of a person be reliably determined, but more details about the presence of the person can also be determined. For this example, Figure 4 The presence of an additional person in the second, newly sensed area is shown. To determine the detection results in this scenario, rules based on the example above can be provided.

[0064] In the same office setting, if RF sensing is used for occupancy-based lighting control in addition to occupancy counting, the system can alternatively be configured, for example, by setting sensitivity to ensure no overlapping areas exist, to avoid turning on lights in adjacent areas, for example. However, as Figure 5 As shown, this lack of overlap results in people going undetected in the blind spot created between sensing areas 310' and 320'. This could cause the system to direct new colleagues to the area assuming there are available seats, when the area may already be nearing saturation when you take into account one or more additional people in the blind spot.

[0065] In this context, the present invention as described above can be applied. Specifically, for example, a potential unreliable detection result determination unit can determine whether a potential double negation has occurred that could lead to an unreliable detection result. This determination can be based, for example, on the determination that a person previously in detection area 310' has left, has nowhere else to go except detection area 320', and is not currently being picked up in either sensing area. People may also arrive from top or bottom areas, for example, through a lobby. Furthermore, for example, a new sensing area definition unit determines the most suitable network device for temporarily performing RF sensing to confirm whether this is a genuine absence that could be triggered by a previous false alarm or double negation. In this case, the most suitable network devices remain network devices 2, 4, 5, and 7. Similarly, the most suitable network devices can be determined based on, for example, an office layout / floor plan, a network neighbor table, analysis of signal strength from nodes, or even after detection, understanding which network devices are most affected by the event. Then, based on the most suitable network devices, a new sensing area 330 including sensing devices 2, 4, 5, and 7 can be created, ideally suited for counting people in the space around them. Then you can temporarily disable the original detection areas 310' and 320'.

[0066] Then, radio frequency (RF) sensing is performed for the desired functionality (in this case, people counting) to verify the actual number of people in the area. In this case, it is advantageous to switch to a different RF sensing functionality, i.e., to perform RF sensing for a different task. For example, a simple motion sensing algorithm could be applied instead of people counting. However, it is also possible to maintain the same functionality, such as the people counting function provided in the example here. Then, to determine the final detection result, rules can be implemented, for example, based on the scenario utilizing people counting:

[0067] 0 There were no omissions or oversights. X X people were missed in both regions.

[0068] Once the detection results can be determined with sufficient reliability, for example, after sufficient confidence in the number of people present, the new sensing area 330 can be disabled in the original sensing areas 310' and 320', allowing them to continue their normal operation, or sensing in the new sensing area 330 can simply be stopped. This depends on which method is used to establish the new sensing area 330 based on sensing areas 310' and 320'. The new sensing results in the new sensing area 330 can then be provided to the system responsible for processing the personnel counting statistics, so that the detection results (here, the total number of people) can be updated if needed.

[0069] In addition to the embodiments described above, the network control device can also be adapted to indicate a confidence level for each detection type, rather than an ambiguous binary value. For example, a potentially unreliable detection result determination unit can be adapted to indicate a 95% confidence level for the presence of two people in an area and a 35% confidence level for the presence of a third person. To ensure appropriate performance, the network can discard the presence of a third person based on a low confidence level when determining the detection result. However, the confidence level can be used to determine whether a new sensing area should be defined by a new sensing area definition unit. Furthermore, for some applications, the confidence level determined by the potentially unreliable detection result determination unit can be used to perform personnel counting, for example, by determining the presence of three to five people in an area with a 95% confidence level. For such applications, a confidence threshold can be provided, allowing the step of defining a new sensing area to be omitted if a result with a confidence level higher than the threshold is provided. The confidence threshold can be determined based on the application of the network or the functional tasks assigned to the network. For example, a coarse personnel count is already very useful for optimizing indoor airflow, cleaning frequency, etc., while for such applications, an exact number may not be necessary.

[0070] Furthermore, as mentioned above, the new sensing area definition unit can further optimize the creation, i.e., definition, of new sensing areas by utilizing confidence levels. For example, new sensing areas can be created to increase the probability of confirming the presence of a person in the area. In particular, if a sensing area has a higher confidence level compared to neighboring sensing areas, a new sensing area can be defined to utilize more network devices originally assigned to more reliable sensing areas. This helps to confirm the presence / count of people with a higher probability more quickly.

[0071] In another embodiment, when performing RF sensing in a new sensing area, losing track of what is happening in the non-overlapping partitions of the original sensing area can be inconvenient. For example, if the network is used to perform functionalities such as occupancy detection, gesture recognition, fall detection, or breathing detection, losing track of it may be unacceptable. In this case, the new sensing area definition unit can be adapted not to dismantle the original sensing areas. In this case, if the resulting increase in network load and / or device resource usage is acceptable, the transmission rate for all original sensing areas can remain the same as before the introduction of the new sensing area. However, if the original sensing areas are not disabled, it is preferable to change the sensing characteristics of these areas, such as their transmission rates. For example, the transmission rates can be changed such that they satisfy one or more of the following conditions. Preferably, the transmission rates are changed such that the total amount of network traffic seen by network devices allocated to the original sensing areas and the new sensing areas remains the same or lower within a specific time window. For example, if the original sensing areas reduce their transmission rates from the original value S to S / 2, then the new sensing areas can coexist simultaneously, and as long as they also transmit at a rate of S / 2, there will be no significant increase in network load. This is preferred when network load is more important than detection latency. In another preferred embodiment, the transmission rate can be varied such that the total traffic seen by other network devices not belonging to the sensing area is affected by the definition of the new sensing area, but the total traffic does not increase significantly within the wireless range of such network devices. This is preferably applied when there are two or more adjacent areas, and network devices in those other sensing areas may be affected by increased network traffic. In another preferred embodiment, the network traffic can be varied such that the total network traffic seen by network devices is a function of, for example, the relative confidence level of each original sensing area determined as described above. For example, if the confidence level of the first sensing area is twice that of the second sensing area, both of which involve the new sensing area, the network traffic can be segmented such that the first sensing area reduces its network traffic to S / 4, the second sensing area to S / 4, and the new sensing area to S / 2, etc.

[0072] In one embodiment, the original sensing areas can remain enabled, thus generating the raw volume of their traffic, but can be configured to perform more basic functionality, such as simply sensing presence rather than counting people. This means there may be higher network traffic compared to them continuing to count, but with lower processing requirements.

[0073] Although in the above embodiments only two sensing regions are defined for the network in most cases, in other embodiments, more than two, such as three or any other number of sensing regions, can be defined. In this case, the more than two sensing regions can also overlap or can be adjacent to regions where radio frequency sensing cannot be performed. The above principle can then be determined based on more than two sensing regions; for example, if three sensing regions overlap, the potentially unreliable detection result determination unit can be adapted to compare the sensing results of the three sensing regions.

[0074] Although the above embodiments are described for lighting systems or office distribution systems, in other embodiments, in addition to general network communication and radio frequency sensing capabilities, the network may include other functionalities such as audio capabilities and sensor capabilities, and can be applied to other situations, such as users' private residences, public spaces in cities, high-security environments, etc. Therefore, although the tasks of the network and the resulting sensing outcomes are exemplarily described in the above embodiments as counting the number of people in the network area, other tasks and therefore other detection outcomes can be envisioned in other applications and in combination with other network functionalities, based on the same principle. For example, the network could provide tasks such as: determining the presence or absence of a person in security applications, detecting a fall in a home environment for the elderly, detecting breathing in a child care environment, and controlling lighting in private homes or public spaces.

[0075] By studying the accompanying drawings, the disclosure, and the appended claims, those skilled in the art can understand and implement other variations of the disclosed embodiments in practicing the claimed invention.

[0076] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude multiple.

[0077] A single unit or device can perform the functions of several items listed in the claims. The mere fact that certain measures are referenced in mutually different dependent claims does not imply that a combination of these measures cannot be used advantageously.

[0078] Procedures such as providing sensing results, determining potentially unreliable detection results, defining new sensing areas, providing new sensing results, and determining detection results, which are performed by one or more units or devices, can be performed by any other number of units or devices. These processes can be implemented as program code devices and / or dedicated hardware for computer programs.

[0079] Computer program products may be stored / distributed on suitable media, such as optical storage media or solid-state media, supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems.

[0080] Any reference numerals in the claims should not be construed as limiting the scope.

[0081] This invention relates to an apparatus for controlling a network including a sensing region. The apparatus includes a) a providing unit for providing sensing results of radio frequency sensing in the sensing region; b) a determining unit for determining whether the sensing results are potentially unreliable in a spatial domain; and c) a defining unit for defining a new sensing region corresponding to a portion of the spatial domain where the sensing results are potentially unreliable. The providing unit is adapted to provide new sensing results in the new sensing region and, based on the sensing results and the new sensing results, to determine a sensing result for at least a portion of the spatial domain where the sensing results are determined to be potentially unreliable. Therefore, this invention provides an apparatus that allows for improved sensing reliability in a radio frequency sensing network.

Claims

1. A network control device for controlling a network (100) including network devices (111, 121) performing radio frequency sensing, wherein the network (100) includes sensing regions (110, 120) defined by the network devices (111, 121) of the network (100), the network devices (111, 121) being assigned to each sensing region (110, 120) such that each of the sensing regions (110, 120) corresponds to a spatial domain, and wherein radio frequency sensing in each sensing region (110, 120) is performed independently of radio frequency sensing in other sensing regions (110, 120), the network control device (130) comprising: The sensing result providing unit (131) is used to provide the sensing results of radio frequency sensing in each of the sensing areas (110, 120). A potentially unreliable detection result determination unit (132) is configured to determine, based on a comparison between sensing results from at least two sensing regions (110, 120), whether a detection result determined based on the sensing results is potentially unreliable in a spatial domain that is part of at least one of the sensing regions (110, 120) and / or adjacent to at least one of the sensing regions (110, 120), wherein the sensing result refers to at least one physical event, and wherein a comparison between sensing results from at least two of the sensing regions (110, 120) is performed to determine whether the same physical event has been detected at least twice and / or whether the physical event has been missed by one of the sensing regions (110, 120), and A new sensing region definition unit (133) is configured to define a new sensing region (150) by assigning network devices (111, 121) of the network (100) to perform radio frequency sensing in the new sensing region (150), wherein the new sensing region (150) corresponds to at least a portion of the spatial domain in which detection results are potentially unreliable, and wherein the radio frequency sensing in the new sensing region (150) is independent of the radio frequency sensing in the other sensing regions (110, 120). The sensing result providing unit (131) is adapted to provide new sensing results of the radio frequency sensing in the new sensing area (150), and Detection results are determined for at least a portion of the spatial domain where the detection results are determined to be potentially unreliable, based on at least one sensing result and the new sensing result.

2. The network control device of claim 1, wherein at least two spatial regions in the spatial field corresponding to the sensing areas (110, 120) overlap, enabling at least two sensing results to be provided for the overlapping portion, and / or wherein a portion of the spatial field in which the network device (111, 121) is distributed does not correspond to the sensing areas (110, 120), enabling sensing results to be unavailable for the uncovered portion, wherein portions of the spatial field in which potentially unreliable detection results are determined correspond at least partially to the overlapping portion and / or the uncovered portion, respectively.

3. The network control device of claim 1, wherein the comparison between the sensing results of at least two sensing areas (110, 120) includes determining the temporal correlation and / or activity fingerprint and / or motion trajectory of the at least two sensing results.

4. The network control device according to claim 1, wherein the new sensing area definition unit (133) is adapted to assign network devices (111, 121) to the new sensing area (150) based on known characteristics of the network (100).

5. The network control device according to claim 1, wherein the new sensing area definition unit (133) is adapted to assign the network device (111, 121) to the new sensing area (150) based on previous detection results and / or previous sensing results and / or based on expectations of sensing results and / or detection results.

6. The network control device according to claim 4, wherein the known characteristics of the network (100) include information indicating the spatial distribution of the network devices (111, 121), the current state of the network devices (111, 121), the environmental conditions of the network devices (111, 121), wireless noise, interference and / or the general capabilities of the network devices (111, 121), and wherein the new sensing area definition unit (133) is adapted to assign the network devices (111, 121) to the new sensing area (150) based on the indicated spatial distribution, the current state, the environmental conditions, wireless noise, interference and / or the capabilities of the network devices (111, 121).

7. The network control device according to claim 1, wherein the potentially unreliable detection result determination unit (132) is adapted to further determine the confidence level of the determined detection result based on the sensing result as each part of the spatial domain formed by the sensing regions (110, 120), wherein the new sensing region definition unit (133) is adapted to define the new sensing region (150) based on the confidence value.

8. The network control device according to claim 1, wherein the new sensing area definition unit (133) is adapted to define the new sensing area (150) such that the sensing sensitivity in the new sensing area (150) is lower than the sensing sensitivity in other sensing areas (110, 120), and / or such that the new sensing result refers to another type of sensing result.

9. The network control device according to claim 1, wherein the new sensing area definition unit (133) is adapted to define the new sensing area (150) such that the sensing sensitivity in the new sensing area (150) is higher than the sensing sensitivity in other sensing areas (110, 120) so as to determine a specific sensing result.

10. The network control device according to claim 1, wherein the new sensing area definition unit (133) is adapted to assign the network devices (111, 121) to the new sensing area (150) such that as long as radio frequency sensing is performed in the new sensing area (150), radio frequency sensing is not performed in the sensing areas (110, 120) previously assigned to the network devices (111, 121), wherein after the radio frequency sensing performed in the new sensing area (150) is completed, the new sensing area definition unit (133) is adapted to reassign the network devices (111, 121) to their previously assigned sensing areas (110, 120).

11. The network control device according to claim 1, wherein the new sensing area definition unit (133) is adapted to control network devices (111, 121) allocated to the sensing areas (110, 120) and the new sensing area (150). The new sensing area defining unit (133) is adapted to control the network devices (111, 121) allocated to the new sensing area (150) and the sensing areas (110, 120) such that, when performing radio frequency sensing in the new sensing area (150), the total amount of processing and / or the total communication load of the network devices (111, 121) are the same as or lower than the amount of processing and / or the total communication load of the network devices (111, 121) when compared with the amount of processing and / or the total communication load of the network devices (111, 121) when performing radio frequency sensing in the new sensing area (150), and / or the new sensing area defining unit (133) is adapted to control the network devices (111, 121) allocated to the sensing areas (110, 120) and the new sensing area (150), such that, when performing radio frequency sensing in the new sensing area (150), the time that the network devices (111, 121) do not spend on sensing is the same as or higher than the time that is not spent on sensing before the new sensing area (150) is defined.

12. A network (100) comprising: Multiple network devices (111, 121), distributed in the space domain and adapted to perform radio frequency sensing, and The network control device (130) according to any one of the preceding claims.

13. A network control method for controlling a network (100) including network devices (111, 121) performing radio frequency sensing, wherein the network (100) includes sensing regions (110, 120) defined by the network devices (111, 121) of the network (100), the network devices (111, 121) being assigned to each sensing region (110, 120) such that each of the sensing regions (110, 120) corresponds to a spatial domain, and wherein radio frequency sensing in each sensing region (110, 120) is performed independently of radio frequency sensing in other sensing regions (110, 120), the network control method (200) comprising: Provides the sensing results of radio frequency sensing in each of the sensing regions (210). Based on a comparison between sensing results from at least two sensing regions, it is determined (220) whether a detection result determined based on the sensing results is potentially unreliable in a spatial domain that is part of at least one of the sensing regions and / or adjacent to at least one of the sensing regions (110, 120), the sensing results referring to at least one physical event, and wherein a comparison between sensing results from at least two of the sensing regions (110, 120) is performed to determine whether the same physical event has been detected at least twice and / or whether the physical event has been missed by one of the sensing regions (110, 120). The new sensing region is defined (230) by assigning network devices of the network to perform radio frequency sensing in the new sensing region, wherein the new sensing region corresponds to at least a portion of the spatial field in which detection results are potentially unreliable, and wherein the radio frequency sensing in the new sensing region is independent of the radio frequency sensing in the other sensing regions. Provide new sensing results of radio frequency sensing (240) in the new sensing area, and Based on at least one sensing result and the new sensing result, a (250) detection result is determined for at least a portion of the spatial domain where the detection result is determined to be potentially unreliable.

14. A computer program product comprising a computer program, which, when executed by a processor, implements the network control method according to claim 13.

Images