Control of HVAC system based on radar sensor detection

Abstract

A method and system for controlling an HVAC system in a space. A radar sensor is arranged in a ceiling of the space such that side lobes of emitted radar energy are directed upwards, towards ventilation ducts located above the ceiling. A processing unit receives an intermediate frequency, IF, signal from the radar sensor, and for consecutive time frames of the IF signal, determines a frequency spectrum, and determines, for each time frame, that at least two predefined frequency components have an amplitude exceeding a respective threshold. In response to determining that the at least two predefined frequency components have an amplitude exceeding a respective threshold in a number of consecutive time frames corresponding to a first predefined time period, it is determined that an HVAC system is active. The invention allows an improved HVAC control.

Description

[0001] CONTROL OF HVAC SYSTEM BASED ON RADAR SENSOR DETECTION

[0002] FIELD OF THE INVENTION

[0003] The present invention generally relates to control of HVAC (Heating, Ventilation and Air Conditioning) systems based on radar sensor detection. Specifically, the present invention proposes using a radar sensor to determine if a HVAC system is active.

[0004] BACKGROUND OF THE INVENTION

[0005] Most office buildings have HVAC systems. Many HVAC systems are centrally controlled via e.g., a building management system (BMS) with a scheduled ON / OFF for energy saving. Other HVAC systems only have local control via e.g., wall panel. Some HVAC systems have both central control and local control for both energy saving and flexibility. This allows employees that work overtime to manually override the central control, and turn on the HVAC via local control.

[0006] For HVAC systems with local control, it may happen that the HVAC of an office space such as meeting room is active also during non-working hours such as night or weekend, simply because the last user forgot to turn off the HVAC before leaving the meeting room. The option of manual control requires a significant effort to avoid such unnecessary power consumption.

[0007] CN116123613A provides an air conditioner and a control method thereof, and relates to the technical field of air conditioners. The method is applied to a controller of the air conditioner indoor unit, the air conditioner indoor unit further comprises a human body sensor, an air guide blade motor and a draught fan, and the controller is electrically connected with the human body sensor, the air guide blade motor and the draught fan. When the human body sensor is used for detecting a human body signal, the air guide blade motor and / or the fan are / is controlled to stop working; and after signal detection is completed, the air guide blade motor and / or the fan are / is controlled to work. The air conditioner and the control method thereof have the advantage of being higher in detection precision.

[0008] EP3993565A1 discloses an enhanced signal processing method for a radarbased presence sensor enables to control a building automation system using at least one sensor device. A sensor signal processing system for a building automation system comprises a signal processing circuit, which is configured to obtain a sensor signal acquired by a sensor device and to evaluate the sensor signal to detect characteristics indicating a heartbeat or respiration in the sensor signal. The signal processing circuit is configured to generate a detection signal comprising information on presence of at least one person based on the evaluated sensor signal, and to output the detection signal to the building automation system. The signal processing circuit is configured to determine different heartbeat rates or respiration rates in the sensor signal, and to determine a number of persons in the surveillance area based on the determined different heartbeat rates or respiration rates. The signal processing circuit maybe arranged locally in the sensor device, or alternatively in a remote server, which performs processing on one or more sensor signals acquired from corresponding sensor devices via a network.

[0009] It would be desirable to have a simple, yet reliable control system that could improve control of manually controlled HVAC systems.

[0010] SUMMARY OF THE INVENTION

[0011] It is an object of the present invention to provide an improved control of HVAC systems, to mitigate or eliminate the problems discussed above. This and other objects may be achieved by a method and a system in accordance with the independent claims. Embodiments of the present invention are defined in the dependent claims.

[0012] Hence, according to a first aspect of the present invention, there is provided a method controlling an HVAC system in a space, comprising receiving an intermediate frequency, IF, signal from a radar sensor arranged in a ceiling of the space such that side lobes of emitted radar energy are directed upwards, into ventilation ducts located above the ceiling, for consecutive time frames of the IF signal, determining a frequency spectrum, determining, for each time frame, that at least two predefined frequency components have an amplitude exceeding a respective threshold, and in response to determining that the at least two predefined frequency components have an amplitude exceeding a respective threshold in a number of consecutive time frames corresponding to a first predefined time period, determining that an HVAC system is active.

[0013] According to a second aspect, there is provided a HVAC control system comprising a radar sensor arranged in a ceiling of a space such that side lobes of emitted radar energy are directed upwards, into ventilation ducts located above the ceiling, the radar sensor being arranged to provide an intermediate frequency, IF, signal, a processing unit configured to: for consecutive time frames of the IF signal, determine a frequency spectrum, determine, for each time frame, that at least two predefined frequency components have an amplitude exceeding a respective threshold, and in response to determining that the at least two predefined frequency components have an amplitude exceeding a respective threshold in a number of consecutive time frames corresponding to a first predefined time period, determine that an HVAC system is active.

[0014] According to a third aspect, there is provided a computer program product, comprising computer program code configured to, when executed on a computer processor connected to a radar sensor, execute the method of the first aspect.

[0015] The approach according to the present invention provides a reliable, and independent detection of whether an HVAC system is active in the space. As will be discussed, this detection may be used in various ways to provide an improved HVAC control, also when the HVAC system itself does not provide for such improved control.

[0016] The invention is based on the realization that a radar sensor, also when installed primarily for the purpose of detecting presence of people in the space, will be suitable for detecting vibrations caused by an HVAC system. The vibrations may be caused by mechanical movement in the HVAC system (e.g. fans etc.) or be AC mains induced vibrations. The vibration will be detectable as a frequency component of a time frequency representation of the intermediate frequency signal from the radar sensor.

[0017] In many cases, the frequency component is not isolated, but appears in connection to other, related components. For example, the radar sensor may receive EMI from the HVAC, which EMI typically features the mains frequency and one or several multiples of the mains frequency. In principle it is possible to identify and detect only one such component to confirm a presence of HVAC, but using two or more synchronized components may improve robustness. For example, identifying multiple synchronized components may allow excluding other potential EMI sources that may be picked up by the radar sensor.

[0018] The first predefined time period is selected such that a reliable and robust detection of an active HVAC system can be made. For example, the first predefined time period can be at least 30 seconds, or at least one minute.

[0019] In one embodiment, it is detected, based on the IF signal, absence of a person in said space, and in response to detecting absence of a person in the space, initiating a HVAC control action. By “absence of a person” is intended a detection determining that there is no person present in the space. A person in the space will typically cause presence of specific frequency components in the IF spectrum. By selecting appropriate thresholds, it can be reliably determined if there is a person in the space. If this is not the case, absence of a person is detected.

[0020] In another embodiment, predefined activity periods associated with the space are accessed, and in response to detecting that the HVAC system is active outside the predefined activity periods, initiating a HVAC control action.

[0021] The HVAC control action may be deactivation of the HVAC system, or communicating a signal to a building management system connected to the HVAC system.

[0022] In a further embodiment, the IF signal is used to detect presence of a person immediately outside the space. This is possible e.g. if the space has a window, and a person passes by outside the window. In response to detecting presence of a person immediately outside the space, the HVAC control action may include generating a visible signal indicating a recommended deactivation of the HVAC system. The visible signal may be generated using a lighting system in the space.

[0023] With this approach, a person passing by outside the space may be alerted that an HVAC system is active even though there is no-one in the space. The passing person can then enter the space and manually deactivate the HVAC system.

[0024] In some embodiments, the method includes an initialization stage, comprising obtaining a time-frequency representation of the IF signal, the time-frequency representation including a sequence of time frames and a frequency spectrum for each time frame, identifying at least two frequency components in a frequency range 0-500 Hz, each frequency component having a substantially constant amplitude for a plurality of consecutive time frames corresponding to a period of time exceeding a second predefined time period, wherein the periods of time of each of the at least two frequency components are synchronized with each other, and selecting the at least two frequency components as the at least two predefined frequency components. During such an initialization stage it can be determined whether there is a HVAC installed in the space.

[0025] The at least two frequency components may be multiples of a base frequency. This may be the case, for example, when the frequency components are AC mains induced. In this case, there will typically be several synchronized frequency components, one corresponding to the AC mains frequency (e.g. 50 Hz) and several others corresponding to multiples of the AC mains frequency.

[0026] The second predefined time period is selected such that the presence of an HVAC system in the space can be reliably determined. For example, the second predefined time period may be at least 30 minutes, and preferably at least one hour. The initialization stage may further comprise determining an average amplitude of each of said at least two frequency components, and associating each detected average amplitude with each respective predefined frequency component. The detected average amplitude may be used to determine the threshold used to detect whether the HVAC is active. So, by associating each predefined frequency component with an average amplitude, each respective threshold can be based on the associated amplitude value.

[0027] In some embodiments, the initialization stage further includes detecting, based on the IF signal, motion activities caused by a person in the space, accumulating detected motion triggers in a histogram, and identifying the predefined activity periods as time periods during which the number of motion triggers are greater than a predefined threshold. If the space is an office space of a company, the activity periods will typically be correlated to regular office hours of the company.

[0028] This approach allows the processing unit to determine when the HVAC system is expected to be activated, even in a situation where the processing unit does not have an accurate timer for determining time or even date information. The occurrence frequency of motion triggers from people will typically correspond to the working hours. During the working hours of a day, presence of human activities in / close to the meeting room would be significantly higher than the off-working hours. Then the motion detection results of the radar sensor can be used to estimate the working hours.

[0029] BRIEF DESCRIPTION OF THE DRAWINGS

[0030] This and other aspects of the present invention will now be described in more detail, with reference to the appended drawings showing embodiments of the present invention.

[0031] Fig. 1 schematically shows a room equipped with an HVAC system and a control system according to an embodiment of the present invention.

[0032] Figs. 2a-b are two flow charts of methods according to embodiments of the present invention.

[0033] Figs. 3a-f show a set of diagrams along a common time axis.

[0034] DETAILED DESCRIPTION

[0035] Figure 1 shows a general illustration of a space or room 1, e.g. a conference room, equipped with a HVAC system 2. The HVAC system 2 is very schematically illustrated as including a duct and one or several fans or other means to circulate air for heating or cooling the room 1. The room may have a window 3 facing e.g. a corridor or an adjacent room. In practice, the means to circulate air may be located elsewhere. There may or may not be a person 4a present in the room 1, and there may or may not be a person 4b present outside the window 3. Further, a radar sensor 5 is mounted in the ceiling of the room. The radar sensor 5 may be installed for various reasons, e.g. to detect when someone enters the room 1. For the present disclosure, the radar sensor 5 forms part of a system 10 for control of (or assisting control of) the HVAC system 2, and may be installed for this sole purpose.

[0036] The radar sensor may use any type of Doppler radar, such as a pulsed Doppler radar or an unmodulated continuous wave (CW) Doppler radar. In the latter example, a constant frequency carrier wave in the GHz range is emitted and its reflection received. Transmitted and reflected waves are mixed to form an intermediate frequency (IF) signal in the time domain. In available radar motion sensors, the carrier frequency may be for example 5.8 GHz, 24 GHz or 60 GHz. The IF signal frequency content will be proportional to the carrier frequency, and to the detected motion. As an example, for a carrier frequency of 5.8 GHz, motion with a velocity of 0.5 m / s results in an IF signal frequency of 19 Hz, while “fast” motion, around 4 m / s results in an IF signal frequency of 152 Hz. The analogue IF signal is typically sampled by an A / D converter to form a digital IF signal.

[0037] As indicated in Figure 1, the radar sensor 5 is mounted in the ceiling 6 with the main lobe 7 of emitted radar energy facing down to the interior of the room 1. Any motion within the meeting room can be thus detected by the radar sensor 5. However, some emitted radar energy will be emitted in several side lobes 8 facing up above the ceiling 6. This means that the radar sensor 5, to a certain extent, can also ‘see’ motion above the ceiling, e.g. vibration of ventilation ducts. When the HVAC system 2 is operated, it will generate movement and vibration that may be picked up by the radar sensor 5. For example, AC mains induced EMI and vibration will cause the IF signal to include frequency components corresponding to the mains frequency (e.g. 50Hz / 60Hz), and multiples thereof (100Hz / 120Hz, 150Hz / 180Hz, etc.).

[0038] The system 10 in Figure 1 further includes a processing unit 9 connected to a memory 11. The processing unit 9 is configured to receive the digital IF signal from the radar sensor 5, and to process this signal to determine if the HVAC system is active. The processing unit 9 may also be configured to take appropriate control action in response to detecting that the HVAC system is active. Finally, the processing unit 9 may be configured to perform an initialization stage, for configuration of the system. Figure 2a-b shows an example of processing performed by the processing unit 9. Figure 2a is an initialization stage, which may be performed only once. One purpose of the initialization stage is to determine if the room is equipped with an HVAC system, and to determine some characteristics of the HVAC system. Another purpose is in this case to determine typical activity periods of the room 1, e.g. regular office hours in the case of an office space. Figure 2b is a control stage, which is performed continuously. During this stage, it is determined if a HVAC system is active, and suitable control action is taken, depending also on other factors.

[0039] Figure 3 shows various signals involved in the processes in Figures 2a-b. All signals are here illustrated for a period of one week.

[0040] Diagram 3a shows an example of a time domain IF signal provided by the radar sensor 5, in this case a 5.8GHz Doppler radar sensor, here sampled with a sampling frequency of 1 kHz.

[0041] Diagram 3b is a time-frequency representation of the IF signal. The representation is formed by dividing the IF signal into a series of consecutive time frames, and perform a Fast-Fourier transform (FFT) of each time frame, thereby forming a frequency spectrum for each time frame. In the diagram, time is on the x-axis, frequency is on the y- axis, and the grey-scale represents the amplitude of the spectrum.

[0042] In the illustrated example, the FFT operates on 128 sample points. With a sampling rate of 1 kHz, this indicates an FFT time window of 128 ms. However, the FFT typically operates on a sliding time window, so that each frequency spectrum time frame in the figure might represent less than 100 ms. It is noted that the time frame can be varied between tens of millisecond up to e.g. one second by using different sampling rate, FFT window size and overlapping ratio.

[0043] As shown in Figure 3b, the time-frequency representation of the IF signal comprises a set, in this case two, of relatively weak, but consistent frequency components 31, 32 that exist horizontally, and which start and stop synchronously. Starting points 31a, 32a and ending points 31b, 32b of these components have been marked in Figure 3. These components are caused by the HVAC system 2. When the HVAC is active (ON), these components appear, while they disappear when HVAC is turned OFF. In the illustrated example, the components 31, 32 are AC mains induced vibrations, at 50 Hz and 100 Hz, but also other frequencies are possible.

[0044] In order to identify these components, the following characteristics may be useful: 1. The frequency components evenly distributed over the frequency range at least between 0Hz and 500Hz,

[0045] 2. They are typically N times (N is an integer) of a basic frequency of 50Hz (or 60Hz) which is same as the AC mains frequency (50Hz / 60Hz),

[0046] 3. They always appear / disappear simultaneously,

[0047] 4. They can last for more than hours with very stable amplitude over the time. Further, it is noted that for a central controlled HVAC via BMS with pre-set scheduling, the duration of these frequency components as well as the relative time of appear / disappear for everyday is almost fixed. However, for a locally controlled HVAC, the start / stop of these frequency components will happen when the radar sensor detects the presence of people (someone needs to be in the room to do the manual ON / OFF of the HVAC). So, for a locally controlled HVAC, the frequency components can never start / stop when the radar sensor 5 does not detect any person in the room.

[0048] Turning to the process in Figure 2a, in a first step SI the time frequency representation 3b is generated by the processing unit 9. Then, in step S2, HVAC -related frequency components are identified using the above characteristics. For example, the processing unit 9 can be configured to identify at least two frequency components in a frequency range 0-500 Hz, each frequency component having a substantially constant amplitude for a plurality of consecutive time frames corresponding to a period of time exceeding a predefined time period, e.g. 30 minutes or one hour, and which frequency components are synchronized with each other.

[0049] Depending on the environment and relative position of the radar sensor with respect to the HVAC, the frequency components may have different amplitude and relative strength. However, it is symptomatic that the relevant frequency components are regular and consistent in time, between clearly distinguishable ON and OFF times.

[0050] If such frequency components 31, 32 can be identified, with synchronized start and end points 31a, 32a and 3 lb, 32b, it is likely that there is a HVAC system installed in the room 1. Even further reliability can be obtained if the frequency components 31, 32 appear several times during a week.

[0051] After such ‘stable’ and synchronized frequency components 31, 32 have been identified, their average amplitude may be determined, and stored together with the frequencies of the identified components. These amplitudes can be used to generate a threshold TH for following detection, e.g. for example TH = amplitude - a, or TH = p x amplitude, where a and P are appropriate constants representing a predefined margin (either a constant margin a, or a proportional margin P). So, for component 31, the processing units 9 stores a frequency fsi and an average amplitude A31, while for component 32, the processing units 9 stores a frequency fs2 and an average amplitude A32.

[0052] Another possible part of the initialization stage is step S4, where motion triggers are identified and counted in consecutive bins to form a motion trigger histogram, e.g. as illustrated in Figure 3d. By setting an appropriate threshold 33, the processing unit may, in step S5, identify primary periods of activity 34, as indicated in Figure 3e. If the room is an office apace, the activity periods 34 may be assumed to correspond to working hours.

[0053] Turning now to the continuously performed process in Figure 2b, the frequencies fsi, £32 and amplitudes A31, A32 are used in step SI 1 to detect the predefined frequency components which are related to operation of the HVAC system 2.

[0054] For detection of the HVAC, a set of separate criteria may be used, based on the stored frequencies and associated amplitudes. For example, a first criterion may stipulate that the amplitude around a first center frequency (e.g. 50 Hz) exceeds the associated threshold TH1 AND that the amplitude around a second center frequency (e.g. 100 Hz) exceeds the associated threshold TH2. A second criterion may then stipulate that the first criteria must be met for a number of consecutive frames, e.g. 1000 frames which, with 100 ms frames as in the example above, would correspond to almost 2 minutes. The first criterion ensures synchronized components, while the second criterion excludes the potential false trigger from other non-HVAC generated signal, including other mains induced components.

[0055] As mentioned above, the thresholds TH1 and TH2 may be correlated to the average amplitudes A31, A32, for example TH1 = A31 - a, or TH1 = PA31 where a and pis are appropriate constants.

[0056] By monitoring these frequency components 31, 32, and their start and end times 3 la, 3 lb, 32a, 32b, it is established in step SI 1 when the HVAC 2 is turned ON and when the HVAC 2 is turned OFF. Figure 3c shows a series of periods 35, between start time toN and stop time toFF, when the HVAC system 2 is active.

[0057] Based on the identified periods 35, the processing unit 9 may make various assessments in steps S12 - S13 and possibly initiate a HVAC control action in step S14.

[0058] One simple assessment, in step S12, is to determine, e.g. based on lack of motion triggers detected by the radar sensor 5, that no-one present in the room 1. In the absence of anyone in the room, the control action in step S14 is initiated.

[0059] Another assessment, in step S13, is to cross-reference the activity of the HVAC system 2 with predefined (expected) activity periods associated with the room 1. These predefined activity periods may be programmed into the processing unit 9, e.g. via a suitable interface (not shown). Or, in the case where the initialization stage includes steps S4- S5 above, the predefined activity periods may be the working hours 34 in Figure 3e.

[0060] If the HVAC system 2 is found to be active outside the predefined activity periods (e.g. office hours) then the control action in step S14 is initiated. Figure 3f illustrates a comparison with HVAC active periods 35 with working hours 34, resulting in periods 36 where the control action in step S14 will be initiated.

[0061] Turning now to the HVAC control action in step S14, this may be various things. In a most simple case, when the processing unit 9 is connected to the HVAC system 2, the control action in step S14 may be to deactivate the HVAC system 2. Alternatively, where the processing unit 9 is connected to a building management system (BMS) which in turn is connected to the HVAC system, the HVAC control action may be communication of a signal to the BMS.

[0062] In yet another situation, in particular when the HVAC system 2 is controlled locally, inside the room, the radar sensor 5 may monitor the presence of a person 4b outside the room, e.g. passing the window 3. Such detection is often possible, as motion triggers caused by such a person are weaker than motion triggers caused by someone inside the room. In response to detecting someone outside the room 1 during a period 36, the HVAC control action in step S14 may include generating a visible signal, indicating a recommended deactivation of the HVAC system 2.

[0063] Especially, if the processing unit 9 is connected to a lighting system the visible signal may involve controlling the lamps / luminaires 12 inside the room 1 to provide a visible alert to a person outside the room 1.

[0064] The person skilled in the art realizes that the present invention by no means is limited to the preferred embodiments described above. On the contrary, many modifications and variations are possible within the scope of the appended claims. For example, the processing unit 9 may be configured to perform other, and more complex, control operations based on the underlying HVAC detection disclosed herein.

Claims

CLAIMS:

1. A method for controlling an HVAC system (2) in a space (1), comprising: receiving an intermediate frequency, IF, signal from a radar sensor (5) arranged in a ceiling (6) of the space (1) such that side lobes (8) of emitted radar energy are directed upwards, towards ventilation ducts located above the ceiling (6); for consecutive time frames of said IF signal, determining a frequency spectrum; determining, for each time frame, that at least two predefined frequency components have an amplitude exceeding a respective threshold; and in response to determining that the at least two predefined frequency components (31, 32) have an amplitude exceeding a respective threshold in a number of consecutive time frames corresponding to a first predefined time period, determining that an HVAC system (2) is active.

2. The method according to claim 1, wherein the first predefined time period is at least 30 seconds, and preferably is at least one minute.

3. The method according to claim 1 or 2, wherein each respective threshold is based on an amplitude value (A31, A32) associated with each respective predefined frequency component.

4. The method according to any one of claims 1 - 3, wherein the radar sensor (5) is arranged in the ceiling (6) of the space (1) such that the main lobe (7) of emitted radar energy is facing down to the interior of the space (1), the method further comprising: detecting (step S12) , based on the IF signal, absence of a person (4a) in said space (1); and in response to detecting absence of a person in said space, initiating (step S14) a HVAC control action.

5. The method according to any one of claims 1 - 3, further comprising:accessing (step SI 3) predefined activity periods (34) associated with said space (1); and in response to detecting that said HVAC system (2) is active outside said predefined activity periods, initiating (step S14) a HVAC control action.

6. The method according to claim 5, wherein the space (1) is an office space of a company, and said predefined activity periods (34) are correlated to regular office hours of the company.

7. The method according to any one of claims 4 - 6, wherein the HVAC control action (step S14) includes deactivation of the HVAC system and / or communication of a signal to a building management system connected to the HVAC system.

8. The method according to any one of claims 4 - 6, further comprising: detecting, based on the IF signal, presence of a person (4b) immediately outside said space, and in response to detecting presence of a person immediately outside said space, said HVAC control action (step S14) includes generating a visible signal, indicating a recommended deactivation of the HVAC system.

9. The method according to claim 8, wherein the visible signal is generated using a lighting system (12) in the space.

10. The method according to any one of the preceding claims, further comprising an initialization stage, including the steps of obtaining (step SI) a time-frequency representation of the IF signal, said timefrequency representation including a sequence of time frames and a frequency spectrum for each time frame; identifying (step S2) at least two frequency components in a frequency range 0-500 Hz, each frequency component having a constant amplitude within a predefined margin for a plurality of consecutive time frames corresponding to a period of time exceeding a second predefined time period, wherein the periods of time of each of said at least two frequency components are synchronized with each other; andselecting said at least two frequency components as said at least two predefined frequency components.

11. The method according to claim 10, wherein the second predefined time period is at least 30 minutes, and preferably at least one hour.

12. The method according to one of claims 10 - 11, wherein the initialization stage further comprises determining an average amplitude of each of said at least two frequency components, and storing (step S3) each detected average amplitude with each respective predefined frequency component.

13. The method according to claim 10 and 5, wherein the initialization stage further comprises: detecting, based on the IF signal, motion triggers caused by a person (4a) in said space (1); accumulating (step S4) detected motion triggers in a histogram (Fig. 3e); and identifying said predefined activity periods (34) as time periods during which the number of motion triggers are greater than a predefined threshold (33).

14. A computer program product, comprising computer program code configured to, when executed on a computer processor connected to a radar sensor, execute the method of one of the preceding claims.

15. An HVAC control system, comprising: a radar sensor (5) arranged in a ceiling (6) of the space (1) such that side lobes(8) of emitted radar energy are directed upwards, into ventilation ducts located above the ceiling (6), said radar sensor (5) being arranged to provide an intermediate frequency, IF, signal; a processing unit (9) configured to: for consecutive time frames of said IF signal, determine a frequency spectrum; determine, for each time frame, that at least two predefined frequency components have an amplitude exceeding a respective threshold; and in response to determining that the at least two predefined frequency components (31, 32) have an amplitude exceeding a respective threshold in a number ofconsecutive time frames corresponding to a first predefined time period, determine that an HVAC system (2) is active.

Images