Network of sensor devices

Abstract

The invention relates to a network of sensor devices (10) for monitoring a room of a building (1000), each sensor device (10) having a sensor (200), which is located in the room (20) of the building (1000) and is adapted so as to detect data relating to objects (35) located in the room (20) within the detection region (215) of the sensor device (10), and a processing unit (400), which is adapted so as to process the data and determine the number of objects (35) in the detection region, wherein the objects (35) are assigned associated data vectors with determined object parameters, and the detection regions (215a, 215b) of the sensor devices (10) at least partly overlap. The processing units (400) of adjacent sensor devices (10a, 10b) are also adapted so as to communicate with one another in order to avoid double-counting objects (35) located in the overlapping region (215c) of the detection regions (215a, 215b) of the adjacent sensor devices (10a, 10b).

Description

[0001] Network of sensor devices

[0002] The invention relates to a network of sensor devices for room monitoring.

[0003] In building automation and modern building management, sensor devices for detecting living beings, especially people or animals, are becoming increasingly important. In addition to traditional RF Doppler motion sensors or IR pyro sensors, there are also sensor devices that can detect not only the movement of people in a room, but also their mere presence even without movement. These so-called presence sensors or occupancy detectors can be used in buildings, for example, to switch on lights in rooms, ensuring that the lights do not go out even if movement is not constantly detected.

[0004] These high-frequency (HF) occupancy detectors operate like HF motion detectors on the Doppler principle, detecting not only macro-movements but also micro-movements, for example, to detect the presence of a breathing person in a room who is not moving based on lung micro-movements. These HF occupancy detectors can only detect the presence of living beings in the detection area, but cannot count or record the number of living beings in the room.

[0005] Camera systems with appropriate image processing are used for this purpose. Since these camera systems operate in the optical range and usually at high resolution, they initially generate raw image data, which generally allows the identification of people within the camera's detection range. With these room camera systems, image analysis takes place directly within the system, so that only processed position and movement data, without the possibility of identifying the people in the room, for example, in a company conference room, leave the sensor. In other words, the identification of the people in the room is carried out directly by the camera system, without the need to transmit personal data to an external unit for identification.The so-called "Human Presence Detection" (HPD) sensors used in these camera systems can therefore not only detect but also count people present in a room, regardless of whether they are moving or not, without raising data protection concerns. The present invention is based on the object of creating a network of sensor devices, in particular sensor devices with an infrared array sensor, that reliably provides information about the number and position of people in a room for building automation functions in compliance with the General Data Protection Regulation (GDPR).

[0006] This object is achieved by the inventive network of sensor devices according to the independent claims. Advantageous embodiments and further developments of the invention are set forth in the subclaims.

[0007] According to the invention, a decentralized network of sensor devices for monitoring the space of a building is provided. The sensor devices each comprise a sensor arranged in the space of the building, which is adapted to acquire data from objects located in the space within a detection range of the sensor device. The sensor devices further comprise a processing unit adapted to process the data and determine the number of objects in the detection range, wherein data vectors associated with the objects are assigned to the objects and wherein the detection ranges of the sensor devices at least partially overlap. The processing units of adjacent sensor devices are further adapted to communicate with one another in order to avoid double counting of objects located in the overlapping area of ​​the detection ranges of the adjacent sensor devices.

[0008] It is advantageous if the objects are living beings, especially people or animals.

[0009] It is expedient if the sensor comprises an infrared array sensor with an IR pixel array and an associated IR optics for spatially resolved recording of IR image pixel data.

[0010] It is advantageous if the sensor comprises a camera sensor for recording camera image pixel data, an array of IR sensors, or an array of RF sensors. It is advantageous if the determined object parameters of the data vectors assigned to the objects include a position of the moving objects in the detection area, a direction of movement of the objects in the detection area, a speed of the moving objects in the detection area, and a timestamp.

[0011] It is expedient if the processing unit of each of the sensor devices is further adapted to track the direction of movement of the objects into the respective detection area based on the data vectors and to detect that objects are approaching the overlap area with a neighboring sensor device and / or are located in the overlap area.

[0012] It is advantageous if a processing unit of two adjacent sensor devices is further adapted to send a signal to the other processing unit when one of the objects enters the overlapping area, wherein the processing unit of the other sensor device does not count the object when determining the number of objects.

[0013] It is advantageous if adjacent sensor devices are further adapted to continuously exchange synchronization signals so that they are synchronized in time.

[0014] It is advantageous if the processing units of adjacent sensor devices are further adapted to perform a handover process of moving objects in the overlapping area.

[0015] It is advantageous if the data vector assigned to the moving objects further comprises a sensor flag.

[0016] It is useful if the overlap area between adjacent sensor units defines a fixed boundary area.

[0017] It is advantageous if the sensor devices are arranged on a ceiling and / or a wall of a room in the building. Further advantages, features, and details of the invention will become apparent from the following description of preferred embodiments and from the drawings. These show:

[0018] Fig. 1A, 1B: a top view and a side view of a sensor device with an infrared array sensor according to an embodiment of the invention,

[0019] Fig. 2: a schematic block view of a known sensor device with a

[0020] Infrared array sensor according to an embodiment of the invention,

[0021] Fig. 3A, 3B: a schematic cross-sectional view and a plan view of a conference room with a ceiling-mounted sensor device according to an embodiment of the invention,

[0022] Fig. 4A, 4B: a user interface on a computer screen displaying thermal pixel raw data and processed people counting data, as well as an illustration of people counting data from a single infrared array sensor,

[0023] Fig. 5A, 5B: a schematic cross-sectional view and a plan view of a warehouse with two sensor devices mounted on the ceiling according to an embodiment of the invention,

[0024] Fig. 6A, 6B: Illustrations of people counting data of a known network of infrared array sensors with erroneous total people counting results and of an inventive network of infrared array sensors with correct total people counting results.

[0025] In the figures, identical components and components with the same function are marked with the same reference numerals.

[0026] Figures 1A and 1B show a top view and a side view of a sensor device 10 according to an embodiment of the invention. The sensor device 10 has an infrared array sensor 200, which enables, for example, the counting of people in a room. Figure 2 shows a schematic block view of a sensor device 10 with an infrared array sensor 200 according to an embodiment of the invention. The sensor device 10 optionally has a sensor housing 100 that can be mounted on a ceiling 21 or wall 22 of a room 20 (Fig. 3A). However, the sensor does not have to be mountable on the ceiling or wall, but can also be attached to a tripod in the room, for example. It is important that the sensor can see into the room from above. Various sensor technologies can be used in the infrared array sensor 200 according to the invention.For example, an IR pixel array 210 may comprise a thermopile matrix based on a stack of thermoelectric elements that generate a current proportional to thermal radiation. However, it is also possible for the IR pixel array 210 to comprise a plurality of microbolometers that measure a change in electrical resistance between two electrodes due to thermal radiation.

[0027] However, the infrared array sensor 200 can also operate with a micro-optical mechanical system (MOMs). This means that the IR pixel array 210 of the infrared array sensor 200 has a micromechanical infrared matrix arrangement as pixels 215, which measures a mirror deflection of the respective micromechanical pixel elements due to incident infrared radiation. However, a camera can also be used that has an image pixel array with associated optics for spatially resolved recording of image pixel data. Furthermore, the invention is not limited to the use of an infrared array sensor 200; for example, an array of spatially distributed IR sensors or RF sensors can also be used to form the inventive network of sensor devices 10. However, the network of sensor devices 10 with an infrared array sensor 200 will be described below as a concrete embodiment of the invention.

[0028] Due to the native low resolution of the IR pixel array 210, such systems are particularly suitable for protecting privacy. With a native low resolution of 28 x 15 pixels 215 of the IR pixel array 210, the infrared array sensor 200 generates a thermal image that complies with data protection regulations; therefore, no personal data is ever recorded, not even as raw pixel data. In addition to the IR pixel array 210, the infrared array sensor 200 also has additional peripheral components 220, which enable, for example, preprocessing of the pixel data of the pixels 215, controlling the IR pixel array 210, or reading and transmitting the pixel data of the pixels 215.

[0029] Furthermore, the sensor device 10 has an IR optics 240 associated with the IR pixel array 210 for spatially resolved recording of thermal image pixel data. The IR pixel array 210 is connected via a data bus 230 to an image processing unit 300 for image processing and image analysis of the thermal image pixel data of the infrared array sensor 200. The sensor device 10 further comprises a central processing unit 400 for object detection of objects in the image pixel plane of the thermal image pixel data processed by the image processing unit. The central processing unit 400 can also be connected via a data bus system 610 to a multi-sensor unit 600, which sends additional sensor data such as brightness, room temperature, humidity, air quality, volatile organic compounds (VOCs), and CO2 content to the central processing unit 400 for further processing.

[0030] Figures 3A and 3B show a schematic cross-sectional view and a plan view of a room 20 of a building 1000, in particular a conference room 20, with a sensor device 10 mounted on the ceiling 21 according to an embodiment of the invention. The conference room 20 has various objects 40 such as a table, chairs, a cupboard, or a display device. People 30 are present in the room 20, who can be detected and counted by the sensor device 10, as described below. Figure 3B shows a plan view of the conference room 20, wherein, for illustration purposes, the depicted pixels 215 of the IR pixel array 210 of the infrared array sensor 200 of the sensor devices 10 are schematically superimposed over the objects 40 and people 30 present in this room 20.

[0031] Figures 4A and 4B show a user interface 50 on a computer screen displaying raw thermal pixel data from pixels 215 and processed people counting data, as well as an illustration of people counting data from people 30 from a single infrared array sensor 200. Figure 4A shows, on the lower left side, a so-called "heat map" generated from the raw pixel data of the IR pixel array 210. This raw pixel data from pixels 215 is sent via the data bus 230 to the image processing unit 300 for image processing and image analysis of the thermal image pixel data from the infrared array sensor 200. Using artificial intelligence (KI), the raw pixel data from pixels 215 is analyzed and segmented to identify objects 35, such as items 40 or people 30, in the room 20.

[0032] As shown in the user interface 50 in Figure 4A, based on the evaluation of the raw pixel data of the pixels 215 by means of blob analysis by the image processing unit 300 or the central processing unit 400, individual people 30 can be distinguished, tracked, and counted based on the body temperature of 37°C in corresponding pixel regions of the pixels 215. Blob analysis is one of the basic image processing functions and is based on extracting features from connected pixels 215 that share the same logical state (blobs) to ensure easy segmentation and analysis of many different object properties of the segmented objects 35, such as items 40 or people 30.

[0033] Furthermore, not only can the number of people 30 be determined at a specific time using blob analysis, as shown in the user interface 50 at the bottom right, but a time course of the number of people 30 in the room 20 can also be determined, as shown in the top center of the user interface 50. Thus, as shown in Figure 4B, the sensor device 10 is capable of counting people 30 in a room 20 and tracking their movement. In the case shown in Figure 4B, two people are outside the room 20, so that three people 30 are determined by the central processing unit 400 to be in the room.

[0034] Figures 5A and 5B show a schematic cross-sectional view and a plan view of a warehouse with two sensor devices 10 mounted on the ceiling according to an embodiment of the invention. These two sensor devices 10 form, by way of example, the network of sensor devices 10 according to the invention. However, the invention is not limited to a network of two sensor devices 10; a plurality of sensor devices 10 can also be used. Although in most applications an entire room 20 can be monitored by just one sensor device 10, there is the case according to the invention in which a network of sensor devices 10a, 10b must be used, in which a room 20, such as a warehouse of a logistics building 1000, must be monitored by means of the network according to the invention with multiple sensor devices 10a, 10b. This is illustrated in Figures 5A and 5B.Here, two sensor devices 10a and 10b are used as a network to monitor people 30 and objects 40, such as a storage shelf or a storage box 40, and their movement within the space. As shown in Figure 5B, the left area of ​​the storage room 20 is monitored by pixel array 215a, and the right area of ​​the storage room 20 is monitored by pixel array 215b. Pixel array 215a and pixel array 215b have an overlap area 215c.

[0035] According to the invention, a network of sensor devices 10 is provided which achieves an intelligent fusion of many simple sensor data, which can be combined into an array in delimited fields and evaluated across multiple fields. An array of simple infrared sensors will be examined in more detail below as an example. These infrared sensors can either be arranged centrally (e.g., as in the Panasonic GridEye or the Calumino CTS) or distributed spatially across many individual IR sensors. It is therefore advantageous if the sensor 200 comprises an infrared array sensor with an IR pixel array 210 and an associated IR optics 240 for the spatially resolved recording of IR image pixel data. However, it is also expedient if the sensor 200 comprises a camera sensor for recording camera image pixel data, an array of IR sensors, or an array of RF sensors.

[0036] According to the invention, a network of sensor devices for monitoring the space of a building is provided. The sensor devices 10 each comprise a sensor 200 arranged in the room 20 of the building 1000, which is adapted to acquire data from objects 35 located in the room 20 within a detection range 215 of the sensor device 10. The sensor devices 10 further comprise a processing unit 400 adapted to process the data and determine the number of objects 35 in the detection range, wherein data vectors with determined object parameters are assigned to the objects 35, and wherein the detection ranges 215a, 215b of the sensor devices 10 at least partially overlap.The processing units 400 of adjacent sensor devices 10a, 10b are further adapted to communicate with each other in order to avoid double counting of objects 35 located in the overlapping area 215c of the detection areas 215a, 215b of the adjacent sensor devices 10a, 10b. It is advantageous if the objects 35 are living beings, in particular people 30 or animals. An essential component of the system is the implementation of a decentralized solution at the product level, which can be based on a Bluetooth mesh network. The central processing unit 500 can thus communicate with other sensor devices 10 via a decentralized radio network, such as Bluetooth mesh, for example, in order to compare people counting data with other sensor devices 10 located in the room 20.

[0037] Figure 6A shows an illustration of people counting data from a known network of infrared array sensors with inaccurate overall people counting results. Sensors are known to be able to locate, count, and track people within their detection range. For this purpose, an analysis of the sensor data for each individual sensor runs on the sensors or sometimes in a cloud service. This becomes problematic when a single sensor cannot cover the entire area. People at the edge of the detection range may then no longer be counted or may even be counted twice, depending on whether the detection ranges are adjacent or overlap, as illustrated in Figures 5B and 6A.

[0038] Figure 6B shows an illustration of people counting data from a network of infrared array sensors according to the invention with correct overall people counting results. In order to record the exact number and position of all people 30 in room 20, this error must be corrected at a higher analysis level. Since people 30 outside the detection range obviously cannot be counted, the areas of the sensors 200 must overlap to avoid undercounting. The problem of double counting must now be solved. To do this, the overlap of the areas must first be defined and identified as a corresponding zone in both sensors. Each person in the detection range of each sensor now receives a quadruple data set consisting of position, direction of movement (vector), speed, and timestamp.

[0039] As shown in Figure 6B, the 4 people from Sensor 1 at time t have the data:

[0040] (1.1); direction (0,2); V=0.2m / s

[0041] (5.1); direction (6,2); V=0.3m / s

[0042] (2,5); direction (3,6); V=0.1 m / s

[0043] (4,3); direction (3,3); V=0.2m / s The 3 people from sensor 2 have the following data at time t:

[0044] (1,6); direction (0.6); V=0.2m / s

[0045] (4,2); direction (3,3); V=0.6m / s

[0046] (4,4); direction (4,5); V=0.05m / s

[0047] With this data and the information that the sensors overlap at coordinates (5,5) to (0,5), a higher-level logic can use the timestamps and direction vectors to determine that two people are moving into the overlap area and will therefore soon be counted twice. These two people, marked in red, are now tracked by the higher-level logic, and the sum of both sensors is adjusted accordingly. A precise time base, which is regularly wirelessly synchronized between the two sensors, is crucial for this.

[0048] It is therefore advantageous if the determined object parameters of the data vectors assigned to the objects 35 include a position of the moving objects 35 in the detection area 215, a direction of movement of the objects 35 in the detection area 215, a speed of the moving objects 35 in the detection area 215, and a timestamp. The processing unit 400 of each of the sensor devices 10 can be adapted to track the direction of movement of the objects 35 into the respective detection area 215a, 215b based on the data vectors and to detect that objects 35 are approaching the overlap area 215c with a neighboring sensor device 10b and / or are located in the overlap area 215c.The processing unit 400 of two adjacent sensor devices 10a, 10b may further be adapted to send a signal to the other processing unit 400 when one of the objects 35 enters the overlapping area 215c, wherein the processing unit 400 of the other sensor device 10b does not count the object 35 when determining the number of objects 35.

[0049] It is advantageous if neighboring sensor devices 10a, 10b are further adapted to continuously exchange synchronization signals so that they are synchronized in time. It is further advantageous if the processing units 400 of neighboring sensor devices 10 are further adapted to perform a handover process of moving objects 35 in the overlap area 215c. The data vector assigned to the moving objects 35 can optionally further comprise a sensor flag. It is advantageous if the overlap area 215c between neighboring sensor units 10 defines a specified boundary area. It is advantageous if the sensor devices 10 are arranged on a ceiling 21 and / or a wall 22 of a room 20 of the building 1000.

[0050] To improve the data vector, it can optionally be extended to include the identification number (ID) and, further optionally, visibility by sensors. This visibility is indicated by flags for sensors 1, 2, 3, and 4. An identification (ID) can be assigned to a person 30 based on their entry location and entry time. This is done by assigning a consecutive number within a sensor device. Defining a door as a threshold allows crossing a threshold to be used as a measurement point for people counting. However, this method is only moderately effective because it is susceptible to inaccuracies in the detection.

[0051] The handover process is initiated as soon as a person enters the sensor range. A check is made to determine whether the person is new or someone already moving at the boundary between the sensor ranges. If the person is new, a new ID is assigned. However, if they are a border crosser, the corresponding flag is set at the new sensor. The presence of duplicate flags indicates that the person is in the handover process. This process also works in a four-sensor configuration.

[0052] The handover process is considered complete as soon as a person leaves the sensor area. In this context, a decision must be made whether the person has completely left the area, for example, through a door, or whether they are a former border crosser. If the person has left the area, their ID is deleted. However, if they are a former border crosser, the flag is only set for one sensor. If multiple flags exist, the system waits until only one flag is active to successfully complete the handover process.

[0053] In the improved surveillance system, each person within the detection range of a sensor now receives a data quadruple. This quadruple consists of the person's position, their direction of movement (specified as a vector), their speed, and a timestamp. In addition, each person is assigned a

[0054] Assigned an identification number (ID) to uniquely identify them.

[0055] The system comprises two sensors 10 that are part of a larger array of sensors. This array consists of 32 individual pyroelectric sensors (pyros), with the times of all sensors synchronized with each other. An existing sensor is responsible for tracking the ID of a person 30 until they are no longer visible within the detection area. Building up the tracking data for each person 30 is a lengthy process, as for each detected movement, the data quadruple consisting of ID, position, direction of movement, speed, and timestamp must be re-captured and updated. The use of the sensors 10 enables precise detection and tracking of people through the exact determination of the data quadruple, which is crucial for monitoring and analyzing movement patterns within the detection area.

[0056] The handover process between the sensors is ensured by a sequence of steps. First, the boundaries between the sensors 10 are defined, with each sensor knowing exactly who its neighbors are. When a person moves from sensor 1 towards a neighboring sensor, for example, sensor 2, sensor 1 sends a notification to sensor 2. This step is crucial for informing the sensors of possible handover events. A key aspect of the process is the synchronization of the time base between the sensors to ensure that all timestamps used to capture the movement data are consistent. This synchronization ensures that when a person handovers from sensor 1 to sensor 2, the timestamp is updated and thus considered new.

[0057] As soon as a person enters the detection area of ​​sensor 2, sensor 2 is already informed by the previous notification and "knows" that this person came from sensor 1. This means that the person is not recorded as new by sensor 2, thus avoiding double counting. However, sensor 2 detects the person and takes over tracking without adding them to the total count. As soon as the person has completely disappeared from sensor 1's detection area, sensor 2 is informed by sensor 1, so that the detected person is now counted. Since this person is no longer counted by sensor 1, a correct total count of people from sensors 1 and 2 can be achieved.The advantage of this handover process according to the invention is that a certain hysteresis occurs; sensor 2 does not immediately count a person entering its detection range, so that a reversal of the person's movement within the detection range of sensor 1 does not immediately result in the total number of people jumping back and forth. Rather, a reliable total number of people is output by the network according to the invention in a decentralized manner, which does not exhibit any oscillations in the total number of people as people move through the room 20 and through the detection ranges 215a, 215b of the sensor devices 10 of the network, without the network having to combine all the raw pixel data from all sensor devices 10 in a central unit through a stitching process.The network thus operates reliably by already using the people counting data from the individual sensor devices, which it simply corrects in decentralized sensor-to-sensor communication for the decentralized network within the sensor devices 10 themselves to obtain a correct total number of people in the network. Thus, the network according to the invention can be expanded as required and without great effort, which, in a central stitching process, would require a complete reinstallation of the system consisting of known sensor devices connected in a star pattern to the central unit.

[0058] Conventional sensors are installed in the room and activated, for example, via smartphones. Geographic parameters (installation height, orientation, detection radius, positions of relevant locations, etc.) are transferred to them by measuring / determining the distance and angle to a reference point or GPS coordinates. This method is very complex and requires a great deal of time and precision.

[0059] According to the invention, one or more sensors (infrared, radar, passive / active) are installed in a room. They are either capable of measuring the distance and / or direction of a movement themselves, or their detection range is so narrowly focused that the position of the sensor represents a position in the room. In this case, a large number of sensors are required, which can, however, also be bundled at a central location with appropriate optics.

[0060] The feedback from each sensor (element) is now transmitted in real time to a display (e.g. an app on a smartphone) via a wireless connection. The installer is then prompted to go to an entrance to the room and start the calibration. This tells the sensor its orientation in the room in relation to the front door. Guided by the display, it then walks around the room (e.g. along all the walls) until it reaches the door again. During this entire time, all sensors record this process and save the respective measured values ​​in a time-synchronized, equidistant grid. In the next step, the sensor data is correlated with the outline of the room visualized on the display. The analytics then “know” the entrance to the room and its boundaries and can locate all sensors accordingly.

[0061] In the next step, the installer is asked to visit important locations (points of interest, POIs). For example, they might sit on an armchair, then on one side of the sofa, then the other, stand in front of the stove, or lie down on the bed. For each of these steps, the sum of all sensor data for the respective POI is saved as a reference, and a typical time spent at the POI can also be defined. The room is thus trained, and the sensor network can very precisely locate people in the room without having to measure a single distance.

[0062] This information is then used in the company to determine the location, length of stay and frequency of use and thus to provide assistance in nursing and elderly care, in schools and kindergartens, offices, restaurants or even retail.

Claims

Patent claims 1. A network of sensor devices (10) for monitoring the space of a building (1000), the sensor devices (10) each comprising: a sensor (200) arranged in the room (20) of the building (1000), which is adapted to acquire data from objects (35) located in the room (20) within a detection range (215) of the sensor device (10), a processing unit (400) adapted to process the data and to determine the number of objects (35) in the detection range, wherein data vectors with determined object parameters are assigned to the objects (35), and wherein the detection ranges (215a, 215b) of the sensor devices (10) at least partially overlap, characterized in that the processing units (400) of adjacent sensor devices (10a, 10b) are further adapted to communicate with each other in order to avoid double counting of objects (35) located in the overlapping range (215c) of the detection areas (215a,215b) of the adjacent sensor devices (10a, 10b) to avoid objects (35).

2. Network according to claim 1, characterized in that the objects (35) are living beings, in particular persons (30) or animals.

3. Network according to one of the preceding claims, characterized in that the sensor (200) comprises an infrared array sensor with an IR pixel array (210) and an IR optics (240) associated therewith for spatially resolved recording of IR image pixel data.

4. Network according to one of the preceding claims, wherein the sensor (200) comprises a camera sensor for recording camera image pixel data, an array of IR sensors, or an array of RF sensors.

5. Network according to one of the preceding claims, characterized in that the determined object parameters of the data vectors assigned to the objects (35) include a position of the moving objects (35) in the detection area (215), a Direction of movement of the objects (35) in the detection area (215), a speed of the moving objects (35) in the detection area (215) and a time stamp.

6. Network according to one of the preceding claims, characterized in that the processing unit (400) of each of the sensor devices (10) is further adapted to track the direction of movement of the objects (35) into the respective detection area (215a, 215b) based on the data vectors and to detect that objects (35) are approaching the overlap area (215c) with an adjacent sensor device (10b) and / or are located in the overlap area (215c).

7. Network according to one of the preceding claims, characterized in that a processing unit (400) of two adjacent sensor devices (10a, 10b) is further adapted to send a signal to the other processing unit (400) when one of the objects (35) advances into the overlapping area (215c), wherein the processing unit (400) of the other sensor device (10b) does not count the object (35) when determining the number of objects (35).

8. Network according to one of the preceding claims, characterized in that adjacent sensor devices (10a, 10b) are further adapted to continuously exchange synchronization signals so that they are synchronized in time.

9. Network according to one of the preceding claims, characterized in that the processing units (400) of adjacent sensor devices (10) are further adapted to carry out a handover process of moving objects (35) in the overlap area (215c).

10. Network according to claim 6, characterized in that the data vector assigned to the moving objects (35) further comprises a sensor flag.

11. Network according to one of the preceding claims, wherein the overlap region (215c) between adjacent sensor units (10) defines a fixed boundary region.

12. Network according to one of the preceding claims, characterized in that the sensor devices (10) are mounted on a ceiling (21) and / or a wall (22) of a 5 room (20) of the building (1000).

Images