Fall detection system and method

Abstract

A computer implemented method of identifying falls using a ranging active reflective wave detector, the method comprising: identifying at least one first location of a respective fall detection event, determined from ranging active reflected wave detector measurements, that is determined to be a false fall detection; subsequent to identifying the at least one first location, using the ranging active reflective wave detector to monitor for falls and generating a fall detection action; and determining whether a second location derived from ranging active reflected wave detector measurements is in a region of interest that is outside an exclusion zone, the exclusion zone being associated with the at least one first location; wherein at least one of the generating of a fall detection action based on a detected fall and / or detecting a fall using the ranging active reflective wave detector is conditional upon the second location being in the region of interest.

Description

[0001] FALL DETECTION SYSTEM AND METHOD

[0002] RELATED APPLICATION / S

[0003] This application claims the benefit of priority of British Patent Application No. 2309131.7, filed on June 17, 2023, the contents of which are incorporated herein by reference in their entirety.

[0004] TECHNICAL FIELD

[0005] The present invention relates generally to a method, computer-readable storage medium and device for monitoring an environment to detect falls.

[0006] BACKGROUND

[0007] There is a need to use a monitoring system to automatically detect and identify a state or activity of a person in a designated space, for example in an interior of a building. One example is when a person has fallen. For example, an elderly person may end up in a hazardous situation when they have fallen and are unable to call for help, or at least unable to do so quickly.

[0008] Some known systems have been developed in which the person wears a pendant that has an accelerometer in it to detect a fall based on kinematics. The pendant can transmit an alert signal upon detecting a fall. However the person may not want to wear, or may be in any case not wearing, the pendant.

[0009] Other systems can monitor a person in a space. For example an active reflected-wave based system (i.e. a system that generates waves and measures reflections of such waves, e.g. a radar, lidar or sonar), may be used to determine an activity or a state of a person. However, in such systems, it can be beneficial to reduce false fall detections.

[0010] SUMMARY

[0011] Various aspects of the present invention are defined in the independent claims. Some preferred features are defined in the dependent claims.

[0012] In some scenarios, it can be difficult to distinguish if a person has fallen or if the person is just purposefully lying down, e.g. to rest. One way to make that distinction is to measure radar data indicative of a person’s height from a floor. That data can be used to at least distinguish between the person having fallen so that they are lying on the floor and a person simply lying on a bed or sofa. However a person lying on an item such as a mattress or sleeping matt that sits directly on the floor, as is common in Japanese culture for example, can make it difficult to distinguish between a person on the floor and a person resting on the item. As such, resting surfaces such as floor mattresses, low beds and the like can lead to false positive fall detections. Further, false positive detections may occur for other reasons.

[0013] A first example of the present disclosure provides a computer implemented method of identifying falls in an environment using a ranging active reflective wave detector, the method comprising: identifying at least one first location of a respective fall detection event, determined from ranging active reflected wave detector measurements, that is determined to be a false fall detection; subsequent to identifying the at least one first location, using the ranging active reflective wave detector to monitor for falls and generating a fall detection action in response to a fall detected using the ranging active reflective wave detector; and determining whether a second location derived from ranging active reflected wave detector measurements is in a region of interest that is outside an exclusion zone, the exclusion zone being associated with the at least one first location; wherein at least one of the generating of a fall detection action based on a detected fall and / or detecting a fall using the ranging active reflective wave detector is conditional upon the second location being in the region of interest.

[0014] The fall detection action may be or comprise a fall detection alert or may comprise generating a fall detection alert. The fall detection action, e.g. the fall detection alert, may comprise transmitting a notification to another device to identify the detected fall.

[0015] Optionally, the determining whether the second location derived from ranging active reflected wave detector measurements is in a region of interest that is outside an exclusion zone may be comprised in the using the ranging active reflective wave detector to monitor for falls or may be performed before or after the using the ranging active reflective wave detector to monitor for falls.

[0016] In some embodiments the second location is the location of the detected fall. In such embodiments the generating of the fall detection alert based on the detected fall is conditional upon the second location being in the region of interest.

[0017] In some embodiments the second location is a location of an object that is determined without, or at least prior to, detecting whether a fall is associated with the object. Optionally in such embodiments, in response to determining that the second location is inside an exclusion zone, the ranging active reflective wave detector may be deactivated and / or a fall detection classifier may be disabled or otherwise not used. Thus, in such embodiments not only is no fall detection alert generated, fall detection may itself be disabled or inoperative, unless the second location is in the region of interest.

[0018] The ranging active reflected wave detector measurements used to determine the fall detection event from the at least one first location may be from said active reflected wave detector.

[0019] The region of interest may be a fall detection region of interest, e.g. a region in which falls are to be detected and optionally to which fall detection is limited.

[0020] The region of interest and / or the exclusion zone may be adaptively detected. The method may comprise identifying at least one, e.g. a plurality of, first events, which may be identified from measurements from the ranging active reflective wave detector. The method may comprise identifying a plurality of first events as respective fall detection events from measurements from the ranging active reflective wave detector.

[0021] In some embodiments, the identifying at least one first location comprises: identifying at least one first event as a respective fall detection event from measurements from the ranging active reflective wave detector; determining, from measurements from the ranging active reflective wave detector, at least one event location, the at least one event location respectively corresponding to the at least one first event, determining whether the at least one first event is a false fall detection; and in an event that the at least one first event is determined to be a false fall detection, identifying the at least one event location as being said at least one first location.

[0022] The identifying at least one first location may comprise identifying a plurality of the first locations. The identifying at least one first location may comprise identifying a plurality of first events as respective fall detection events from measurements from the ranging active reflective wave detector. The identifying at least one first location may comprise determining, from measurements from the ranging active reflective wave detector, a respective event location corresponding to each first event of the plurality of first events. The identifying at least one first location may comprise determining whether each first event of the plurality of first events is a false fall detection. The identifying at least one first location may comprise, for any first event determined to be a false fall detection, identifying the event location corresponding to that first event as being a first location of the plurality of first locations.

[0023] In some embodiments, the determining that the second location is in a region of interest that is outside an exclusion zone comprises determining that the second location is not proximate to the at least one first location.

[0024] The method may comprise basing or otherwise deriving the exclusion zone from a cluster of the first locations associated with the plurality of first events. In some embodiments, determining that the second location is not proximate to the at least one first location may comprise determining that the second location is not part of a cluster of locations comprising or consisting of the second location and at least a set or preset minimum number of first locations, such as at least two first locations.

[0025] In some embodiments, said at least one event location comprises a plurality of event locations and the method comprises running a clustering algorithm on the said at least one event locations wherein a plurality of first locations constitutes a cluster identified using the clustering algorithm.

[0026] The clustering algorithm may be or comprise an unsupervised learning algorithm. The clustering algorithm may be a density based clustering algorithm. The clustering algorithm may be configured to identify one or more clusters of first locations for fall detection events determined to be false fall detections. Each cluster may correspond to a continuous region having a predefined minimum density of the first locations for fall detection events determined to be false fall detections, and / or the density is more dense (by a predefined measure) relative to other regions.

[0027] The clustering algorithm may identify clusters based on one or more parameters of the algorithm. The one or more parameters may comprise one or more of: a minimum or threshold number of first locations clustered together for a region to be considered to have a high density of first locations, a distance measure used to locate other first locations in a neighborhood of a given first location, and / or the like. The minimum or threshold number of first locations clustered together for a region to be considered to have a high density of first locations may be in a range from two to four points, such as three points. The distance or threshold measure (epsilon) used to locate other first locations in a neighborhood of a given first location may be in a range from 40 to 80 cm, e.g. 60cm. The clustering algorithm may be or comprise a DBSCAN algorithm. In other examples, the clustering algorithm may be or comprise K-means clustering or a different clustering algorithm.

[0028] The determining that the second location is not part of a cluster comprising or consisting of the second location and at least two first locations may comprise determining that the second location is not part of a cluster of locations comprising or consisting of the second location and at least two first locations using the clustering algorithm.

[0029] The determining that the second location is not proximate to the at least one first location may comprise running the clustering algorithm on a plurality of the first event locations and the second location and determining whether the second location forms a cluster with at least the set or preset minimum number of first locations, such as at least two first locations, and determining that the second location is not proximate to the at least one first location if clustering algorithm determines that the second location does not form a cluster with at least the set or preset minimum number of first locations, e.g. with at least two first locations. The set or present number of locations may be at least two, at least three, at least four or at least five first locations. By requiring that the second location form a cluster with a plurality of locations of prior false fall detections, then false positives may be reduced.

[0030] In other examples, instead of or in addition to determining that the second location is not part of a cluster of locations, the determining that the second location is not proximate to the at least one first location may comprise determining that the second location is not within a predetermined distance of one or more, e.g. a cluster of, first locations.

[0031] For example a region surrounding the at least one first location may be determined, and it may be determined whether the second location is within that region, irrespective of whether the second location is itself part of the cluster. The method may in such cases comprise using a clustering algorithm to determine whether there exists any clusters of first locations. Determining that the second location is not proximate to the cluster may comprise determining that the second location is not within a predetermined distance from a location corresponding to or associated with the cluster of locations. In some embodiments, the location associated with the cluster is a determined center of the cluster. In some embodiments, the determined center of the cluster is an averaged location of the first locations that constitute the cluster. In some embodiments the second location comprises a multi-dimensional (e.g. 3 -dimensional) coordinate. In some embodiments each at least one first location comprises a multi-dimensional (e.g. 3 -dimensional) coordinate. Optionally the multi-dimensional coordinate may be or represent a representative center, optionally based on a weighted center, of reflections from an object. Optionally the multi-dimensional coordinate may be determined using a tracking algorithm to locate the object using a plurality of time- sequential measurement frames from the ranging active reflective wave detector. The object may be the body of a person.

[0032] The multi-dimensional coordinate, e.g. the center or weighted center of reflections for the object, may be determined from the positions of reflections of the active wave from the ranging active reflective wave detector, and which may have regard to the intensity and / or magnitude of the reflections (e.g. a centre location comprising an average of the locations of the reflections weighted by their intensity and / or magnitude).

[0033] In one or more embodiments, the object’s centre or the centre of the part of the object may be a weighted centre of the reflections from the object. The locations may be weighted according to an Radar Cross Section (RCS) estimate of each measurement point, where for each measurement point the RCS estimate may be calculated as a constant, which may be determined empirically, and may be a function of one, two or more of: the signal, the signal to noise ratio and / or the distance from the ranging active reflected wave detector to the position corresponding to the measurement point.

[0034] The object may, for the at least one first location, be defined when a person is in a fallen position. The object for the second location may be defined when a person is in a fallen position.

[0035] In an embodiment, the at least one first location that the exclusion zone is associated with may consist of a set of first locations of a respective fall detection event determined to be a false fall detection, wherein the set of first locations may be limited to a maximum size. The maximum size may, for example, limit the number of locations in the set to a value between 7 and 12, e.g. 10 locations. The set of first locations may be a set of first locations determined by a sliding window of a set or preset number of most recently identified first locations.

[0036] The exclusion zone may be adaptable to take into account newly identified first locations. Additionally or alternatively, the exclusion zone may be adaptable to disregard or deemphasize relatively older first locations. For example, the locations in the set may be comprise most recently identified first locations of a respective fall detection event determined to be a false fall detection. For example, when the set of locations has a size equal to the maximum size, as a new first location where a respective fall detection event is determined to be a false fall detection is identified, the new first location may replace an oldest first location in the set.

[0037] In some embodiments, event locations are disqualified from being a first location if more than a predetermined amount of time (e.g. 2 months) has elapsed since the event location was identified.

[0038] In some embodiments, an exclusion zone may be retired as an exclusion zone in an event that a predefined condition is met.

[0039] The predefined condition may for example comprise that a predetermined amount of time (e.g. 2 months) has elapsed since a most recent first event has occurred within the exclusion zone and / or has at least partly defined the exclusion zone. For example, if within the predetermined amount of time there have been no first events that have caused the exclusion zone to be adapted, then the exclusion zone may in some embodiments be retired and / or each of the at least one first location associated with the exclusion zone may be retired so that an exclusion zone is no longer defined or definable based on the at least one first location.

[0040] In some embodiments, the method comprises identifying the second location, wherein the second location is a location associated with detected fall, the fall being detected based on ranging active reflected wave detector measurements, wherein in response determining that to the second location is not outside the exclusion zone, in the method a fall detection alert is not generated in response to said detected fall.

[0041] In other embodiments, the method comprises identifying the second location, wherein the second location represents a location of a person and monitoring for falls comprises operating a fall detection process based on ranging active reflected wave detector measurements conditional at least upon a predefined criterion being met, wherein the predefined criterion comprises that the person is determined based on the second location to be in a region of interest that is outside the exclusion zone. Optionally said fall detection process comprises: operating the ranging active reflected wave detector to gather ranging active reflected wave measurements for a fall detection classifier and operating the fall detection classifier to detect a fall. The ranging active reflected wave detector may require a large amount of power to operate relative to other forms of detector. However, the ranging active reflected wave detector may be particularly effective in determining falls. The method may comprise using other detectors, such as audio detectors, PIR detectors, imaging devices, tracking devices, and / or the like, to determine that the person is likely in the region of interest, e.g. that the person is, based on the second location, likely to be in the region of interest that is outside the exclusion zone. Additionally or alternatively, the ranging active reflected wave detector may itself be used to determine whether the person is the region of interest outside the exclusion zone. The method may comprise operating the ranging active reflected wave detector (or if it was already used to determine that the person is the region of interest outside the exclusion zone, then further operating the ranging active reflected wave detector) to gather ranging active reflected wave measurements for a fall detection classifier and / or operating the fall detection classifier to detect a fall conditional on it being determined that the person is likely in the region of interest, e.g. using the other detectors. In this way, power consumption may be reduced, which may be especially beneficial in battery operated or other devices not reliant on a physically connected electrical supply.

[0042] Generating the fall detection alert may be or comprise one or both of: transmitting a notification to identify the detected fall, or outputting a visual and / or audible alert in relation to the detected fall.

[0043] The method may comprise implementing a plurality of exclusion zones, each exclusion zone being associated with at least one first location, optionally with a plurality of first locations. Each exclusion zone may be associated with a different cluster of first locations. Each exclusion zone may be associated with different first locations, e.g. with a different plurality of first locations, to each other exclusion zone. Determining that the second location is in a region of interest that is outside an exclusion zone may comprise determining that the second location is not proximate to any of the at least one first locations, for example not part of or proximate to any clusters of first locations. The determining that the second location is in a region of interest outside an exclusion zone may comprise determining that the second location is not part of a cluster of locations associated with any of the exclusion zones.

[0044] In some embodiments, determining whether the at least one first event is a false fall detection comprises, in response to identifying each of the at least one first event: generating a fall detection alert in response to the first event; receiving a response to the fall detection alert indicative of at least one of: whether the fall detection alert represented a fall event, whether the fall detection alert did not represent a false event, and in an event at least one of: receiving that the response indicative of the fall detection alert having not represented a fall event, or not receiving a response indicative of the fall detection alert having represented a fall event, the method comprises determining that the first event is a false fall detection. The response may, for example, be a manually input response or may be received via wireless communication, e.g. using a transceiver on the device.

[0045] Determining whether or not the fall detection alert represented a fall event may be inferred without a response for a person to which the fall detection alert relates. Determining whether the at least one first event is a false fall detection may comprise accruing a plurality of the first events over a period of time (e.g. between 5 days and 15 days, e.g. 1 week). The determining whether the at least one first event is a false fall detection may comprise inferring after the period of time one or more of the at least one first locations from respective ones of the at least one event locations, for example based on at least one of: an assumption that no falls occurred during the period of time, or receiving a confirmation signal indicating that no falls occurred during the period of time. That is, during a period where it is assumed or confirmed that no falls have occurred, event locations giving rise to at least one first event, particularly to a plurality of first events or clusters of first events, may be indicative of something at the one or more of the at least one first locations that gives rise to false fall detection events, such that subsequent events detected at those event locations are more likely to be fall detection events falsely detecting a fall. As such, first events during the period of time may be treated as false fall detection events. The above approach may be a convenient way of identifying whether the at least one first event is a false fall detection regardless of the mechanism that causes the false detection.

[0046] In embodiments where the exclusion zone is based on a cluster of first locations, assuming that no falls occurred during the period of time may be justified as, even if one of the first locations represented an actual fall, since the exclusion zone requires a cluster of first locations, it may be concluded that it is unlikely a plurality of falls occurred during the period of time. Hence the likelihood of an exclusion zone being unjustifiably formed is low.

[0047] The period of time may optionally be a set-up period during which fall detection alerts in response to the respective first events are at least one of: not generated or generated but not treated as an alert, or not as real alert, by a device receiving the alert. In the latter case, an operator of the device may not be notified of the alert and / or may not be prompted to respond to the alert. The set-up period may be a period during which no prompt is provided, verbal or otherwise, for a person to provide a response indicating that a first event is a false fall detection. The method may comprise not listening for responses from a person regarding a false fall detection during the set-up period.

[0048] The method may comprise implementing at least one of the exclusion zone only during one or more a set or preset time periods.

[0049] The determining of the at least one region of interest may comprise automatically inferring at least one of the one or more regions of interest or the at least one exclusion region, e.g. by learning using one or more learning algorithms. Optionally the device may switch to be limited to the region of interest after a learning period. The learning period may be or may be comprised in the set-up period. The determining of the at least one region of interest may comprise inferring, e.g. automatically inferring, at least one of the one or more regions of interest or the at least one exclusion region based on one or more of the first events. The one or more of the first events may be at least one of: detected prior to the monitoring of the at least one region of interest to detect a fall and / or detected during the monitoring of the at least one region of interest to detect a fall. The automatically inferring the at least one of the one or more regions of interest and / or the at least one exclusion region may be carried out before use of the device to detect a fall and / or during use of the device to detect a fall. For example, the automatic inferring of the at least one of the one or more regions of interest or the at least one exclusion region may comprise creating and / or varying the at least one of the one or more regions of interest and / or the at least one exclusion region.

[0050] The ranging active reflective wave detector may comprise one of: radar, lidar, sonar or the like.

[0051] The first locations may, in some cases, correspond to a resting surface. The resting surface may be or comprise a surface for reclining on top of. The resting surface may be or comprise a surface for sitting or lying on top of. The resting surface may be or comprise a surface elevated from a floor for resting on top of. The resting surface may be or comprise a surface of any one of: a mat, bed, mattress, futon, shikibuton, chair, seat, sofa, couch, chaise lounge, bench, stool, beanbag, cushion, floor cushion, pouffe, or the like.

[0052] The exclusion zone may consist of or comprise the resting surface, but in some other embodiments, the region being excluded from the region of interest may additionally or alternatively, comprise a horizontally adjacent region (e.g. abutting in the horizontal sense). For example, the surface for resting thereon may be a seating surface of a chair, wherein the region encompassing the surface may be any horizontal location at, above (and optionally below) the seating surface, and the horizontally adjacent region may be a region of the floor where the feet (and optionally other lower parts of the legs) of a person may be while the person is sitting on the seating surface. This can be advantageous because it may be that the seating surface is outside or partly outside the observable region, and as a result the ranging active reflected wave detector may be able to measure reflections only coming from the person’s feet or the lower part of the person’s legs, which could present as reflections that might otherwise be confusable with a person lying on the floor.

[0053] Even if the seating surface is entirely within the region of interest, it may be that the person, being supported by the chair, may present very little movement above their feet and lower legs, at least for some times in which the ranging active reflected wave detector is collecting measurements. Their feet and potentially their lower legs may, however, during some of those times move and represent the only reflected wave measurement(s) from a moving reflection point(s). In embodiments in with static reflected wave measurements are disregarded, this could the only measured reflections being from the person’s feet or the lower part of the person’s legs, which could present as reflections that might otherwise be confusable with a person lying on the floor.

[0054] In either case by excluding the region where the feet and optional where the legs or lower part thereof may be while the person is sitting, fewer false detections may be achievable.

[0055] The or each at least one fall detection region of interest may be determined by two-dimensional limitations (e.g. limitations to horizontal dimensions) or three-dimensional limitations. In either case the or each at least one region may be defined as covering a horizontal area. In the case of having three-dimensional limitations the or each at least one fall detection region of interest may additionally or alternatively cover a volume.

[0056] Preferably, a total volume or total area (e.g. horizontal area) of the at least one region of interest (i.e. for a single fall detection region of interest, the volume or area that it covers; for a plurality of regions a volume or area covered by the regions of interest collectively) that is the same or preferably less than a volume or area (e.g. horizontal area) of a region that is observable by the ranging active reflective wave detector (i.e. a “radar observable region”). The fall detection region of interest may be a subset of the region that is observable by the ranging active reflective wave detector. The method may comprise configuring the device to store a spatial characterization of a single fall detection region of interest that may be smaller than the region that is observable by the ranging active reflective wave detector. The method may comprise configuring the device to store a spatial characterization of a plurality of regions of interest, wherein the total volume or area (e.g. horizontal area) of the regions of interest may be smaller than the volume or area (e.g. horizontal area) of the region that is observable by the ranging active reflective wave detector.

[0057] The region that is observable by the ranging active reflective wave detector may be a region for which the ranging active reflective wave detector is capable of identifying wave reflection locations from any location in said region. The region that is observable by the ranging active reflective wave detector may comprise a region for which the ranging active reflective wave detector is capable of identifying wave reflection locations (e.g. coordinates) caused by reflections from a moving point of reflection, e.g. causing a measurable Doppler effect, from any location in said region.

[0058] The region that is observable by the ranging active reflective wave detector may be a region in the environment that the ranging active reflective wave detector is capable of observing, when positioned in the environment and operational, which may be defined by distance limitations (e.g. the maximum distance that is observable by the radar) and its field of view. The region that is observable by the ranging active reflective wave detector may, for example, define the region in which objects are detectable by the ranging active reflected wave detector. Thus the region that is observable by the ranging active reflective wave detector may be a maximum region that the ranging active reflected wave detector can monitor when positioned for use. Such a maximum region may also be limited by a region for which the ranging active reflected wave detector is able to ensure that one or more minimum detection performance level requirements are met. In addition to any limitations on the maximum range of the detector (e.g. the maximum identifiable distance to a location of wave reflection) the distance limitations of the observable region may also be limited by environmental factors, for example by any walls or floors of a room that reflect the waves of the detector, preventing objects behind the wall / beneath the floor from being observed.

[0059] The spatial characterization of the at least one fall detection region of interest may comprise data defining at least one virtual fence. Each virtual fence of the at least one virtual fence may define a boundary of a respective portion of a region that is observable by the ranging active reflective wave detector.

[0060] The spatial characterization of the at least one fall detection region of interest may comprise a characterization of at least one exclusion region, whereby at least part of the least one fall detection region of interest may be defined based on excluding the at least one exclusion region from the at least one fall detection region of interest.

[0061] The method may comprise determining the at least one fall detection region of interest and / or the at least one exclusion zone at least one of: during installation or after installation, e.g. whilst the monitoring device is in a fall detection region of interest input mode, being a mode of the device that accepts input of the at least one fall detection region of interest.

[0062] A second example of the present disclosure provides a processing system for identifying falls in an environment using output of a ranging active reflective wave detector, the processing system comprising a processor configured to: identify at least one first location where a respective fall detection event, determined from ranging active reflected wave detector measurements, is determined to be a false fall detection; subsequent to identifying the at least one first location, use the ranging active reflective wave detector to monitor for falls and generate a fall detection action in response to a fall detected using the ranging active reflective wave detector, and determine whether a second location derived from ranging active reflected wave detector measurements is in a region of interest that is outside an exclusion zone, the exclusion zone being associated with the at least one first location; wherein at least one of the generating of a fall detection action based on a detected fall and detecting a fall using the ranging active reflective wave detector is conditional upon the second location being in the region of interest.

[0063] The processor may be configured to perform any of the methods described herein, including those of the first example.

[0064] The processing system may be provided in the device that comprises the ranging active reflective wave detector or may be remote from the device, or may be distributed between a processing system provided on the device and a processing system remote from the device.

[0065] The fall detection action may be or comprise a fall detection alert.

[0066] In some embodiments the second location is the location of the detected fall. In such embodiments the generating of the fall detection alert based on the detected fall is conditional upon the second location being in the region of interest. In some embodiments the second location is a location of an object that is determined without, or at least prior to, detecting whether a fall is associated with the object. Optionally in such embodiments, in response to determining that the second location is inside an exclusion zone, the ranging active reflective wave detector may be deactivated and / or a fall detection classifier may be disabled or otherwise not used. Thus, in such embodiments not only is no fall detection alert generated, fall detection may itself be disabled or inoperative, unless the second location is in the region of interest.

[0067] The ranging active reflected wave detector measurements used to determine the fall detection event from the at least one first location may be from said active reflected wave detector.

[0068] The region of interest may be a fall detection region of interest, e.g. a region in which falls are to be detected and optionally to which fall detection is limited.

[0069] The region of interest and / or the exclusion zone may be adaptively detected.

[0070] In some embodiments, the identifying at least one first location comprises: identifying at least one first event as a respective fall detection event from measurements from the ranging active reflective wave detector; determining, from measurements from the ranging active reflective wave detector, at least one event location, the at least one event location respectively corresponding to the at least one first event, determining whether the at least one first event is a false fall detection; and in an event that the at least one first event is determined to be a false fall detection, identifying the at least one event location as being said at least one first location.

[0071] The identifying at least one first location may comprise identifying a plurality of the first locations. The identifying at least one first location may comprise identifying a plurality of first events as respective fall detection events from measurements from the ranging active reflective wave detector. The identifying at least one first location may comprise determining, from measurements from the ranging active reflective wave detector, a respective event location corresponding to each first event of the plurality of first events. The identifying at least one first location may comprise determining whether each first event of the plurality of first events is a false fall detection. The identifying at least one first location may comprise, for any first event determined to be a false fall detection, identifying the event location corresponding to that first event as being a first location of the plurality of first locations.

[0072] In some embodiments, the determining that the second location is in a region of interest that is outside an exclusion zone comprises determining that the second location is not proximate to the at least one first location.

[0073] In some embodiments, determining that the second location is not proximate to the at least one first location may comprise determining that the second location is not part of a cluster of locations comprising or consisting of the second location and at least a set or preset minimum number of first locations, such as at least two first locations.

[0074] In some embodiments, said at least one event location comprises a plurality of event locations and the processor is configured to run a clustering algorithm on the said at least one event locations wherein a plurality of first locations constitutes a cluster identified using the clustering algorithm.

[0075] The clustering algorithm may be or comprise an unsupervised learning algorithm. The clustering algorithm may be a density based clustering algorithm. The clustering algorithm may be configured to identify one or more clusters of first locations for fall detection events determined to be false fall detections. Each cluster may correspond to a continuous region having a predefined minimum density of the first locations for fall detection events determined to be false fall detections, and / or the density is more dense (by a predefined measure) relative to other regions.

[0076] The clustering algorithm may identify clusters based on one or more parameters of the algorithm. The one or more parameters may comprise one or more of: a minimum or threshold number of first locations clustered together for a region to be considered to have a high density of first locations, a distance measure used to locate other first locations in a neighborhood of a given first location, and / or the like. The clustering algorithm may be or comprise a DBSCAN algorithm. In other examples, the clustering algorithm may be or comprise K-means clustering or a different clustering algorithm.

[0077] The determining that the second location is not part of a cluster comprising or consisting of the second location and at least two first locations may comprise determining that the second location is not part of a cluster of locations comprising or consisting of the second location and at least two first locations using the clustering algorithm. The determining that the second location is not proximate to the at least one first location may comprise running the clustering algorithm on a plurality of the first event locations and the second location and determining whether the second location forms a cluster with at least the set or preset minimum number of first locations, such as at least two first locations, and determining that the second location is not proximate to the at least one first location if clustering algorithm determines that the second location does not form a cluster with at least the set or preset minimum number of first locations, e.g. with at least two first locations. The set or present number of locations may be at least two, at least three, at least four or at least five first locations. By requiring that the second location form a cluster with a plurality of locations of prior false fall detections, then false positives may be reduced.

[0078] In other examples, instead of or in addition to determining that the second location is not part of a cluster of locations, the determining that the second location is not proximate to the at least one first location may comprise determining that the second location is not within a predetermined distance of one or more, e.g. a cluster of, first locations.

[0079] For example a region surrounding the at least one first location may be determined, and it may be determined whether the second location is within that region, irrespective of whether the second location is itself part of the cluster. The processor may be configured to implement a clustering algorithm to determine whether there exists any clusters of first locations. Determining that the second location is not proximate to the cluster may comprise determining that the second location is not within a predetermined distance from a location corresponding to or associated with the cluster of locations. In some embodiments, the location associated with the cluster is a determined center of the cluster. In some embodiments, the determined center of the cluster is an averaged location of the first locations that constitute the cluster.

[0080] In some embodiments the second location comprises a multi-dimensional (e.g. 3 -dimensional) coordinate. In some embodiments each at least one first location comprises a multi-dimensional (e.g. 3 -dimensional) coordinate. Optionally the multi-dimensional coordinate may be or represent a representative center, optionally based on a weighted center, of reflections from an object. Optionally the multi-dimensional coordinate may be determined using a tracking algorithm to locate the object using a plurality of time- sequential measurement frames from the ranging active reflective wave detector. The object may be the body of a person. The multi-dimensional coordinate, e.g. the center or weighted center of reflections for the object, may be determined from the positions of reflections of the active wave from the ranging active reflective wave detector, and which may have regard to the intensity and / or magnitude of the reflections (e.g. a centre location comprising an average of the locations of the reflections weighted by their intensity and / or magnitude).

[0081] In one or more embodiments, the object’s centre or the centre of the part of the object may be a weighted centre of the reflections from the object. The locations may be weighted according to an Radar Cross Section (RCS) estimate of each measurement point, where for each measurement point the RCS estimate may be calculated as a constant, which may be determined empirically, and may be a function of one, two or more of: the signal, the signal to noise ratio and / or the distance from the ranging active reflected wave detector to the position corresponding to the measurement point.

[0082] The object may, for the at least one first location, be defined when a person is in a fallen position. The object for the second location may be defined when a person is in a fallen position.

[0083] In an embodiment, the at least one first location that the exclusion zone is associated with may consist of a set of first locations of a respective fall detection event determined to be a false fall detection, wherein the set of first locations may be limited to a maximum size. The maximum size may, for example, limit the number of locations in the set to a value between 7 and 12, e.g. 10 locations. The set of first locations may be a set of first locations determined by a sliding window of a set or preset number of most recently identified first locations.

[0084] The exclusion zone may be adaptable to take into account newly identified first locations. Additionally or alternatively, the exclusion zone may be adaptable to disregard or deemphasize relatively older first locations.

[0085] For example, the locations in the set may be comprise most recently identified first locations of a respective fall detection event determined to be a false fall detection. For example, when the set of locations has a size equal to the maximum size, as a new first location where a respective fall detection event is determined to be a false fall detection is identified, the new first location may replace an oldest first location in the set. In some embodiments, event locations are disqualified from being a first location if more than a predetermined amount of time (e.g. 2 months) has elapsed since the event location was identified.

[0086] In some embodiments, an exclusion zone may be retired as an exclusion zone in an event that a predefined condition is met.

[0087] The predefined condition may for example comprise that a predetermined amount of time (e.g. 2 months) has elapsed since a most recent first event has occurred within the exclusion zone and / or has at least partly defined the exclusion zone. For example, if within the predetermined amount of time there have been no first events that have caused the exclusion zone to be adapted, then the exclusion zone may in some embodiments be retired and / or each of the at least one first location associated with the exclusion zone may be retired so that an exclusion zone is no longer defined or definable based on the at least one first location.

[0088] In some embodiments, the processor may be configured to identify the second location, wherein the second location is a location associated with detected fall, the fall being detected based on ranging active reflected wave detector measurements, wherein in response determining that to the second location is not outside the exclusion zone, a fall detection alert is not generated in response to said detected fall.

[0089] In other embodiments, the processor is configured to identify the second location, wherein the second location represents a location of a person and monitoring for falls comprises operating a fall detection process based on ranging active reflected wave detector measurements conditional at least upon a predefined criterion being met, wherein the predefined criterion comprises that the person is determined based on the second location to be in a region of interest that is outside the exclusion zone. Optionally said fall detection process comprises: operating the ranging active reflected wave detector to gather ranging active reflected wave measurements for a fall detection classifier and operating the fall detection classifier to detect a fall. The ranging active reflected wave detector may require a large amount of power to operate relative to other forms of detector. However, the ranging active reflected wave detector may be particularly effective in determining falls. The processor may be configured to use other detectors, such as audio detectors, PIR detectors, imaging devices, tracking devices, and / or the like, to determine that the person is likely in the region of interest, e.g. that the person is, based on the second location, likely to be in the region of interest that is outside the exclusion zone. Additionally or alternatively, the ranging active reflected wave detector may itself be used to determine whether the person is the region of interest outside the exclusion zone. The processor may be configured to operate the ranging active reflected wave detector (or if it was already used to determine that the person is the region of interest outside the exclusion zone, then further operating the ranging active reflected wave detector) to gather ranging active reflected wave measurements for a fall detection classifier and / or operating the fall detection classifier to detect a fall conditional on it being determined that the person is likely in the region of interest, e.g. using the other detectors. In this way, power consumption may be reduced, which may be especially beneficial in battery operated or other devices not reliant on a physically connected electrical supply.

[0090] Generating the fall detection alert may be or comprise one or both of: transmitting a notification to identify the detected fall, or outputting a visual and / or audible alert in relation to the detected fall.

[0091] The processor may be configured to implement a plurality of exclusion zones, each exclusion zone being associated with at least one first location, optionally with a plurality of first locations. Each exclusion zone may be associated with a different cluster of first locations. Each exclusion zone may be associated with different first locations, e.g. with a different plurality of first locations, to each other exclusion zone. Determining that the second location is in a region of interest that is outside an exclusion zone may comprise determining that the second location is not proximate to any of the at least one first locations, for example not part of or proximate to any clusters of first locations. The determining that the second location is in a region of interest outside an exclusion zone may comprise determining that the second location is not part of a cluster of locations associated with any of the exclusion zones.

[0092] In some embodiments, determining whether the at least one first event is a false fall detection comprises, in response to identifying each of the at least one first event: generating a fall detection alert in response to the first event; receiving a response to the fall detection alert indicative of at least one of: whether the fall detection alert represented a fall event, whether the fall detection alert did not represent a false event, and in an event at least one of: receiving that the response indicative of the fall detection alert having not represented a fall event, or not receiving a response indicative of the fall detection alert having represented a fall event, the processor is configured to determine that the first event is a false fall detection. The response may, for example, be a manually input response or may be received via wireless communication, e.g. using a transceiver on the device. Determining whether or not the fall detection alert represented a fall event may be inferred without a response for a person to which the fall detection alert relates. In embodiments, determining whether the at least one first event is a false fall detection may comprise accruing a plurality of first events over a period of time (e.g. between 5 days and 15 days, e.g. 1 week), and inferring after the period of time one or more of the at least one first locations from respective ones of the at least one event locations, for example based on at least one of: an assumption that no falls occurred during the period of time, or receiving a confirmation signal indicating that no falls occurred during the period of time. That is, during a period where it is assumed or confirmed that no falls have occurred, event locations giving rise to at least one first event, particularly to a plurality of first events or clusters of first events, may be indicative of something at the one or more of the at least one first locations that gives rise to false fall detection events, such that subsequent events detected at those event locations are more likely to be fall detection events falsely detecting a fall. As such, the above approach may be a convenient way of identifying whether the at least one first event is a false fall detection regardless of the mechanism that causes the false detection.

[0093] In embodiments where the exclusion zone is based on a cluster of first locations, assuming that no falls occurred during the period of time may be justified as, even if one of the first locations represented an actual fall, since the exclusion zone requires a cluster of first locations, it may be concluded that it is unlikely a plurality of falls occurred during the period of time. Hence the likelihood of an exclusion zone being unjustifiably formed is low.

[0094] The period of time may optionally be a set-up period during which fall detection alerts in response to the respective first events are at least one of: not generated or generated but not treated as an alert, or not as real alert, by a device receiving the alert. In the latter case, an operator of the device may not be notified of the alert and / or may not be prompted to respond to the alert.

[0095] The processor may be configured to implement at least one of the exclusion zone only during one or more a set or preset time periods.

[0096] The determining of the at least one region of interest may comprise automatically inferring at least one of the one or more regions of interest or the at least one exclusion region, e.g. by learning using one or more learning algorithms. Optionally the device may switch to be limited to the region of interest after a learning period. The determining of the at least one region of interest may comprise inferring, e.g. automatically inferring, at least one of the one or more regions of interest or the at least one exclusion region based on one or more detected events. The one or more detected events may be at least one of: detected prior to the monitoring of the at least one region of interest to detect a fall and / or detected during the monitoring of the at least one region of interest to detect a fall. The automatically inferring the at least one of the one or more regions of interest and / or the at least one exclusion region may be carried out before use of the device to detect a fall and / or during use of the device to detect a fall. For example, the automatic inferring of the at least one of the one or more regions of interest or the at least one exclusion region may comprise creating and / or varying the at least one of the one or more regions of interest and / or the at least one exclusion region.

[0097] The ranging active reflective wave detector may comprise one of: radar, lidar, sonar or the like.

[0098] The first locations may, in some cases, correspond to a resting surface. The resting surface may be or comprise a surface for reclining on top of. The resting surface may be or comprise a surface for sitting or lying on top of. The resting surface may be or comprise a surface elevated from a floor for resting on top of. The resting surface may be or comprise a surface of any one of: a mat, bed, mattress, futon, shikibuton, chair, seat, sofa, couch, chaise lounge, bench, stool, beanbag, cushion, floor cushion, pouffe, or the like.

[0099] The exclusion zone may consist of or comprise the resting surface, but in some other embodiments, the region being excluded from the region of interest may additionally or alternatively, comprise a horizontally adjacent region (e.g. abutting in the horizontal sense). For example, the surface for resting thereon may be a seating surface of a chair, wherein the region encompassing the surface may be any horizontal location at, above (and optionally below) the seating surface, and the horizontally adjacent region may be a region of the floor where the feet (and optionally other lower parts of the legs) of a person may be while the person is sitting on the seating surface. This can be advantageous because it may be that the seating surface is outside or partly outside the observable region, and as a result the ranging active reflected wave detector may be able to measure reflections only coming from the person’s feet or the lower part of the person’s legs, which could present as reflections that might otherwise be confusable with a person lying on the floor.

[0100] Even if the seating surface is entirely within the region of interest, it may be that the person, being supported by the chair, may present very little movement above their feet and lower legs, at least for some times in which the ranging active reflected wave detector is collecting measurements. Their feet and potentially their lower legs may, however, during some of those times move and represent the only reflected wave measurement(s) from a moving reflection point(s). In embodiments in with static reflected wave measurements are disregarded, this could the only measured reflections being from the person’s feet or the lower part of the person’s legs, which could present as reflections that might otherwise be confusable with a person lying on the floor.

[0101] In either case by excluding the region where the feet and optional where the legs or lower part thereof may be while the person is sitting, fewer false detections may be achievable.

[0102] The or each at least one fall detection region of interest may be determined by two-dimensional limitations (e.g. limitations to horizontal dimensions) or three-dimensional limitations. In either case the or each at least one region may be defined as covering a horizontal area. In the case of having three-dimensional limitations the or each at least one fall detection region of interest may additionally or alternatively cover a volume.

[0103] Preferably, a total volume or total area (e.g. horizontal area) of the at least one region of interest (i.e. for a single fall detection region of interest, the volume or area that it covers; for a plurality of regions a volume or area covered by the regions of interest collectively) that is the same or preferably less than a volume or area (e.g. horizontal area) of a region that is observable by the ranging active reflective wave detector (i.e. a “radar observable region”). The fall detection region of interest may be a subset of the region that is observable by the ranging active reflective wave detector. The system is configured to store a spatial characterization of a single fall detection region of interest that may be smaller than the region that is observable by the ranging active reflective wave detector. The system is configured to store a spatial characterization of a plurality of regions of interest, wherein the total volume or area (e.g. horizontal area) of the regions of interest may be smaller than the volume or area (e.g. horizontal area) of the region that is observable by the ranging active reflective wave detector.

[0104] The region that is observable by the ranging active reflective wave detector may be a region for which the ranging active reflective wave detector is capable of identifying wave reflection locations from any location in said region. The region that is observable by the ranging active reflective wave detector may comprise a region for which the ranging active reflective wave detector is capable of identifying wave reflection locations (e.g. coordinates) caused by reflections from a moving point of reflection, e.g. causing a measurable Doppler effect, from any location in said region. The region that is observable by the ranging active reflective wave detector may be a region in the environment that the ranging active reflective wave detector is capable of observing, when positioned in the environment and operational, which may be defined by distance limitations (e.g. the maximum distance that is observable by the radar) and its field of view. The region that is observable by the ranging active reflective wave detector may, for example, define the region in which objects are detectable by the ranging active reflected wave detector. Thus the region that is observable by the ranging active reflective wave detector may be a maximum region that the ranging active reflected wave detector can monitor when positioned for use. Such a maximum region may also be limited by a region for which the ranging active reflected wave detector is able to ensure that one or more minimum detection performance level requirements are met. In addition to any limitations on the maximum range of the detector (e.g. the maximum identifiable distance to a location of wave reflection) the distance limitations of the observable region may also be limited by environmental factors, for example by any walls or floors of a room that reflect the waves of the detector, preventing objects behind the wall / beneath the floor from being observed.

[0105] The spatial characterization of the at least one fall detection region of interest may comprise data defining at least one virtual fence. Each virtual fence of the at least one virtual fence may define a boundary of a respective portion of a region that is observable by the ranging active reflective wave detector.

[0106] The spatial characterization of the at least one fall detection region of interest may comprise a characterization of at least one exclusion region, whereby at least part of the least one fall detection region of interest may be defined based on excluding the at least one exclusion region from the at least one fall detection region of interest.

[0107] The processor may be configured to determine the at least one fall detection region of interest and / or the at least one exclusion zone at least one of: during installation or after installation, e.g. whilst the monitoring device is in a fall detection region of interest input mode, being a mode of the device that accepts input of the at least one fall detection region of interest.

[0108] A third example of the present disclosure provides a system comprising the processing system of the second example and a ranging active reflective wave detector, wherein the processing system is configured to identify falls in an environment using output of the ranging active reflective wave detector. The processing system may be integrated into a single common device with the ranging active reflective wave detector. The processing system and the ranging active reflective wave detector may be contained within a common housing and / or physically coupled together.

[0109] A fourth example of the present disclosure provides a computer implemented method of operating a device comprising a ranging active reflective wave detector to detect a fall in an environment in which the device is installed, the method comprising: receiving a signal from a sensor other than the ranging active reflective wave detector, wherein the receiving of the signal from the sensor is indicative that the person is located on an object in the environment; using the ranging active reflective wave detector to monitor for falls; generating a fall detection action in response to a fall detected using the ranging active reflective wave detector; wherein at least one of the generating of a fall detection action based on a fall detected using the ranging active reflective wave detector and / or detecting a fall using the ranging active reflective wave detector is conditional upon the sensor not indicating that a person is located on the object.

[0110] The fall detection action may be or comprise a fall detection alert.

[0111] The object may have a resting surface for resting on top of. In some embodiments the resting surface is elevated from the floor. In some embodiments the resting surface is at or within 50 cm or within 40 cm or within 30 cm of an adjacent floor level. For example, the resting surface may be a top surface of a futon.

[0112] The ranging active reflective wave detector may comprise one of: radar, lidar, sonar or the like. The sensor may optionally be a different type of device to the ranging active reflective wave detector, e.g. the sensor may not be a ranging active reflective wave detector.

[0113] The resting surface may be or comprise a surface for sitting or lying or otherwise reclining on top of. The resting surface may be or comprise a surface of any one of: a mat, bed, mattress, futon, shikibuton, chair, seat, sofa, couch, chaise lounge, bench, stool, beanbag, cushion, floor cushion, pouffe, or the like. The monitoring of the at least part of the environment may comprise using the ranging active reflective wave detector to monitor the environment to detect a fallen person in at least one fall detection region of interest.

[0114] Preferably, a total volume or total area (e.g. horizontal area) of the at least one region of interest (i.e. for a single fall detection region of interest, the volume or area that it covers; for a plurality of regions, a volume or area covered by the regions of interest collectively) is the same or preferably less than a volume or area (e.g. horizontal area) of a region that is observable by the ranging active reflective wave detector (i.e. a “radar observable region”). The fall detection region of interest may be a subset of the region that is observable by the ranging active reflective wave detector. The method may comprise configuring the device to store a spatial characterization of a single fall detection region of interest that may be smaller than the region that is observable by the ranging active reflective wave detector. The method may comprise configuring the device to store a spatial characterization of a plurality of regions of interest, wherein the total volume or area (e.g. horizontal area) of the regions of interest may be smaller than the volume or area (e.g. horizontal area) of the region that is observable by the ranging active reflective wave detector.

[0115] The object may be located within a region that is observable by the ranging active reflected wave detector, and may optionally be more or mostly within said at least one fall detection region of interest, wherein the fall detection region of interest is a fall detection region of interest in which falls are detectable based on measurements by the ranging active reflected wave detector. The fall detection region of interest and / or the region that is observable by the ranging active reflected wave detector may be a fall detection region of interest for detecting one or more activities of a person and / or states of a person that may include non-fall states. Thus, even if, based on the sensor, a person is determined to be resting on the surface and therefore not in a fall condition, the active reflected wave detector may in some embodiments still operate to determine a state or activity of the person while the person is on the surface.

[0116] The sensor may be a pressure sensor. The signal from the sensor may be indicative that the person is on the surface by detecting that at least a minimum amount of force is upon sensor. The sensor may be in, on or under the resting surface, e.g. under a mattress, cushion, shikibuton, mat, or the like.

[0117] The sensor may be any other presence sensor, or a motion sensor. The sensor may be a bed sensor for detecting occupancy of a bed. Optionally it may be determined that a person is on the resting surface based on the sensor detecting a presence of a person above the surface that is indicative of the person being on the surface. For example, the detecting the presence of a person above the surface may comprise detecting the presence of a person above the surface using a proximity sensor that detects when a person is above surface but not when they are not above the surface. In another example, the detecting the presence of a person above the surface may comprise detecting the presence of a person above the surface using a motion sensor looking downward towards the surface with a field of view that does not extend laterally beyond the surface.

[0118] With the method describe above, a different sensor, such as but not limited to a pressure sensor, is provided for determining whether a person is resting on a resting surface of an object, where the person would be expected to be prone, sitting or otherwise reclining. In this way, even though the ranging active reflective wave detector may in some embodiments provide a signal that is indicative that a person has fallen, no fall detection is output or consequential action taken, optionally at least without qualification or further checks being performed. In other embodiments operation of the active reflected wave detector may be inhibited (e.g. whereby it not does not provide a signal) in response to the signal from the sensor other than the ranging active reflective wave detector being indicative that the person is located on the object in the environment.

[0119] In either case, when a person is lying or otherwise reclining on a resting surface of an object that may otherwise give rise to a false determination that the person has fallen when they are merely resting, no alarm may be raised.

[0120] A fifth example of the present disclosure provides a processing system for identifying falls in an environment using output of a ranging active reflected wave detector, the processing system comprising a processor configured to: receive a signal from a sensor other than the ranging active reflective wave detector, wherein the receiving of the signal from the sensor is indicative that the person is located on an object in the environment; use the ranging active reflective wave detector to monitor for falls; generate a fall detection alert in response to a fall detected using the ranging active reflective wave detector; wherein at least one of the generating of a fall detection alert based on a fall detected using the ranging active reflective wave detector and / or detecting a fall using the ranging active reflective wave detector is conditional upon the sensor not indicating that a person is located on the object.

[0121] The processing system may be configured to control the ranging active reflective wave detector to monitor at least part of the environment, optionally conditional upon the sensor not indicating that a person is located on the object.

[0122] The ranging active reflective wave detector may comprise one of: radar, lidar, sonar or the like. The sensor may optionally be a different type of device to the ranging active reflective wave detector, e.g. the sensor may not be a ranging active reflective wave detector.

[0123] The object may have a resting surface for resting on top of. The resting surface may be or comprise a surface for sitting or lying or otherwise reclining on top of. The resting surface may be or comprise a surface of any one of: a mat, bed, mattress, futon, shikibuton, chair, seat, sofa, couch, chaise lounge, bench, stool, beanbag, cushion, floor cushion, pouffe, or the like.

[0124] The monitoring of the at least part of the environment may comprise using the ranging active reflective wave detector to monitor the environment to detect a fallen person in at least one fall detection region of interest.

[0125] Preferably, a total volume or total area (e.g. horizontal area) of the at least one fall detection region of interest (i.e. for a single fall detection region of interest, the volume or area that it covers; for a plurality of regions, a volume or area covered by the fall detection regions of interest collectively) is the same or preferably less than a volume or area (e.g. horizontal area) of a region that is observable by the ranging active reflective wave detector (i.e. a “radar observable region”). The fall detection region of interest may be a subset of the region that is observable by the ranging active reflective wave detector. The method may comprise configuring the device to store a spatial characterization of a single fall detection region of interest that may be smaller than the region that is observable by the ranging active reflective wave detector. The method may comprise configuring the device to store a spatial characterization of a plurality of fall detection regions of interest, wherein the total volume or area (e.g. horizontal area) of the fall detection regions of interest may be smaller than the volume or area (e.g. horizontal area) of the region that is observable by the ranging active reflective wave detector. The object may be located within a region that is observable by the ranging active reflected wave detector, and may optionally be more or mostly within said at least one fall detection region of interest, wherein the fall detection region of interest is a fall detection region of interest in which falls are detectable based on measurements by the ranging active reflected wave detector. The fall detection region of interest and / or the region that is observable by the ranging active reflected wave detector may be a fall detection region of interest for detecting one or more activities of a person and / or states of a person that may include non-fall states.

[0126] The sensor may be a pressure sensor. The signal from the sensor may be indicative that the person is on the surface by detecting that at least a minimum amount of force is upon sensor. The sensor may be in, on or under the resting surface, e.g. under a mattress, cushion, shikibuton, mat, or the like.

[0127] The sensor may be any other presence sensor, or a motion sensor. The sensor may be a bed sensor for detecting occupancy of a bed.

[0128] The processing system may be configured to determine that a person is on the resting surface based on the sensor detecting a presence of a person above the surface that is indicative of the person being on the surface. The processing system may be configured to detect the presence of a person above the surface using a proximity sensor that detects when a person is above surface but not when they are not above the surface. The processing system may be configured to detect the presence of a person above the surface using a motion sensor looking downward towards the surface with a field of view that does not extend laterally beyond the surface.

[0129] The processing system may be configured such that, even though the ranging active reflective wave detector may in some embodiments provide a signal that is indicative that a person has fallen, no fall detection is output or consequential action taken, optionally at least without qualification or further checks being performed. In other embodiments, the processing system may be configured such that operation of the active reflected wave detector may be inhibited (e.g. whereby it not does not provide a signal) in response to the signal from the sensor other than the ranging active reflective wave detector being indicative that the person is located on the object in the environment. In either case, when a person is lying or otherwise reclining on a resting surface of an object that may otherwise give rise to a false determination that the person has fallen when they are merely resting, no alarm may be raised.

[0130] A sixth example of the present disclosure provides a system comprising the processing system of the fifth example and a ranging active reflective wave detector, wherein the processing system is configured to identify falls in an environment using output of the ranging active reflective wave detector.

[0131] The processing system may be integrated into a single common device with the ranging active reflective wave detector. The processing system and the ranging active reflective wave detector may be contained within a common housing and / or physically coupled together.

[0132] According to another example of the present disclosure there is provided a computer-readable storage medium comprising instructions which, when executed by a processor of a device to cause the processor to perform any of the methods described herein, including that of at least one of: the first example and / or the fourth example.

[0133] The instructions may be provided on one or more carriers. For example there may be one or more non-transient memories, e.g. a EEPROM (e.g. a flash memory) a disk, CD- or DVD-ROM, programmed memory such as read-only memory (e.g. for Firmware), one or more transient memories (e.g. RAM), and / or a data carrier(s) such as an optical or electrical signal carrier. The memory / memories may be integrated into a corresponding processing chip and / or separate to the chip. Code (and / or data) to implement embodiments of the present disclosure may comprise source, object or executable code in a conventional programming language (interpreted or compiled) such as C, or assembly code, code for setting up or controlling an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array), or code for a hardware description language. The computer readable storage medium may be a non-transient and / or physical computer readable storage medium.

[0134] The individual features and / or combinations of features defined above in accordance with any aspect of the present invention or below in relation to any specific embodiment of the invention may be utilised, either separately and individually, alone or in combination with any other defined feature, in any other aspect or embodiment of the invention. Furthermore, the present invention is intended to cover apparatus configured to perform any feature described herein in relation to a method and / or a method of using or producing, using or manufacturing any apparatus feature described herein.

[0135] These and other aspects will be apparent from the embodiments described in the following. The scope of the present disclosure is not intended to be limited by this summary nor to implementations that necessarily solve any or all of the disadvantages noted.

[0136] BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0137] For a better understanding of the present disclosure and to show how embodiments may be put into effect, reference is made to the accompanying drawings in which:

[0138] Figure 1 illustrates an example arrangement in which a local control hub or remote server is in communication with an active reflected wave detector and a sensor;

[0139] Figure 2 illustrates an example arrangement in which a local control hub or remote server is in communication with a monitoring device that is in communication with a sensor;

[0140] Figure 3 illustrates another example arrangement of a monitoring device;

[0141] Figure 4 illustrates an example of a monitoring device that doesn’t use an external sensor;

[0142] Figure 5 illustrates a human body with indications of reflections measured by a reflective wave detector when the person is in a standing state;

[0143] Figure 6 shows an example of a room being monitored for falls utilising sensors to monitor whether a person is on a resting surface;

[0144] Figure 7 shows an example of a sensor for monitoring whether a person is on a resting surface;

[0145] Figure 8A is a flowchart illustrating a method of operation for a fall detection system that uses one or more regions of interest;

[0146] Figure 8B is a flowchart illustrating an alternative method of operation for a fall detection system that uses one or more regions of interest;

[0147] Figure 9A shows an example of a room being monitored using a fall detection region of interest that excludes a single exclusion region;

[0148] Figure 9B is a plan view of a room showing exclusion zone and a region of interest;

[0149] Figure 9C is a plan view of a room showing exclusion zone and a region of interest;

[0150] Figure 10A shows an example of a room being monitored using a fall detection region of interest that excludes a multiple exclusion regions;

[0151] Figure 10B is a plan view of a room showing multiple exclusion zones and a region of interest; Figures 11 A and 1 IB are plan views of a room showing exclusion zone and a region of interest;

[0152] Figures 12A and 12B are plan views of a room showing multiple regions of interest;

[0153] Figure 13 A is a flowchart illustrating another method of operation for a fall detection system with exclusion zones which may be adaptive;

[0154] Figure 13B is a flowchart illustrating an alternative method of operation for a fall detection system with exclusion zones which may be adaptive;

[0155] Figure 14 is a detailed example of the method of Figure 13A; and

[0156] Figure 15 is illustration of clustering to determine whether a person is in a region associated with false fall detections.

[0157] DETAILED DESCRIPTION

[0158] In the following detailed description, reference is made to the accompanying drawings that form a part hereof, and in which is shown by way of illustration specific embodiments in which the inventive subject matter may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice them, and it is to be understood that other embodiments may be utilized, and that structural, logical, and electrical changes may be made without departing from the scope of the inventive subject matter. Such embodiments of the inventive subject matter may be referred to, individually and / or collectively, herein by the term “invention” merely for convenience and without intending to voluntarily limit the scope of this application to any single invention or inventive concept if more than one is in fact disclosed.

[0159] The following description is, therefore, not to be taken in a limited sense, and the scope of the inventive subject matter is defined by the appended claims and their equivalents.

[0160] In the following embodiments, like components are labelled with like reference numerals.

[0161] In the following embodiments, the term data store or memory is intended to encompass any computer readable storage medium and / or device (or collection of data storage mediums and / or devices). Examples of data stores include, but are not limited to, optical disks (e.g., CD-ROM, DVD-ROM, etc.), magnetic disks (e.g., hard disks, floppy disks, etc.), memory circuits (e.g., EEPROM, solid state drives, random-access memory (RAM), etc.), and / or the like.

[0162] As used herein, except wherein the context requires otherwise, the terms “comprises”, “includes”, “has” and grammatical variants of these terms, are not intended to be exhaustive. They are intended to allow for the possibility of further additives, components, integers or steps. The functions or algorithms described herein are implemented in hardware, software or a combination of software and hardware in one or more embodiments. The software comprises computer executable instructions stored on computer readable carrier media such as memory or other type of storage devices. Further, described functions may correspond to modules, which may be software, hardware, firmware, or any combination thereof. Multiple functions are performed in one or more modules as desired, and the embodiments described are merely examples. The software is executed on a digital signal processor, ASIC, microprocessor, microcontroller, or other type of processor.

[0163] Specific embodiments will now be described with reference to the drawings.

[0164] Figure 1 illustrates one example arrangement in which a system 102 comprising a local control hub 101 and / or remote server 103 is in communication with monitoring device 200 comprising a ranging active reflective wave detector 106 and an operator device or system 110. The system 102 is external to the monitoring device and may therefore also be referred to herein as “external system 102”. The operator device or system 110 may be a user device which may be a portable device such as a mobile phone, tablet, smartwatch or the like (e.g. for personal use), or may be a computer system of a monitoring agency, nursing station or the like having or interfacing with one or more user interfaces or terminals. The monitoring device 200 that comprises the ranging active reflected wave detector 106 is configured to determine that a person has fallen and flag the fall to the external system 102, which in turn may notify the operator device or system 110 for further action. In some embodiments the monitoring device 200 may offload some of processing of the output of the ranging active reflected wave detector 106 to the external system 102 such that the external system 102 may also be involved in the determining that a person has fallen. In some embodiments, the external system 102 may additionally or alternatively be involved in determining that a person has fallen based on other input data, e.g. from a sensor 108, the function of which will be described herein. Although the remote server 103 is depicted as a single server, it will be appreciated that it may be a distributed system that may optionally include or utilize any number of cloud based services.

[0165] Optionally, the monitoring device 200 comprises a trigger sensor 112 for detecting presence or motion of a person and wherein the ranging active reflected wave detector 106 is triggerable from an inactive or low power state into an activated state (e.g. a higher power consumption operating mode) responsive to detection of motion or presence of a person by the trigger sensor 112. The trigger sensor 112 comprises a sensor that has lower power consumption than the ranging active reflected wave detector 106, and could comprise, for example, a passive infra-red (PIR) motion detector, an audio detector, an imaging sensor and / or the like. In this way, the ranging active reflected wave detector 106 is only active for fall detection when the trigger sensor 112 that has lower power requirements than the ranging active reflected wave detector 106 determines that a person is present. This may save power, which is particularly important for battery powered monitoring devices.

[0166] The monitoring device 200 can transmit data to, and receive data from, the external system 102 via a wired and / or wireless connection. Similarly, the operator device or system 110 can receive data from the external system 102 via a wired and / or wireless connection. The external system 102 comprises a communications interface to facilitate communication with other devices such as the operator device or system 110 or a plurality thereof.

[0167] In embodiments in which a local control hub 101 is used, the local control hub 101 may be a control hub of a monitoring system and / or home automation system. The local control hub 101 may for example be a wall or table mounted control hub. The control hub 101 may be “local” in that it is located in the same premises as the ranging active reflected wave detector 106 and hence the monitoring device 200. In such embodiments, the control hub 102 and the monitoring device 200 may thus be part of a common Local Area Network (LAN), Personal Area Network (PAN) or home area network (HAN), but may more specifically be part of a common Wireless Local Area Network (WLAN), Wireless Personal Area Network (WPAN) or wireless home area network (WHAN). Communications between the control hub 102 and the monitoring device 200, respectively, may for example employ a protocol in accordance with or similar to IEEE 802.15.4, a long-range wireless communication protocol (e.g. Amazon Sidewalk™), Bluetooth™, or WiFi™. Thus optionally the control hub 101 may be considered as “local” to the ranging active reflected wave detector 106 based on them being in the same LAN, PAN or HAN.

[0168] In embodiments in which a remote server 103 is used, the remote server 103 is not located in the same premises as the ranging active reflected wave detector 106 and hence the monitoring device 200. For such embodiments, the remote server 103 and the operator device or system 110 and the monitoring device 200 (and / or the control hub 101, if also present) may be part of Wide Area Network (WAN), e.g. the Internet, and may each have different WAN addresses for use in communications between the remote server 103 and the monitoring device 200 (and / or the control hub 101, if also present) and the different device 110. Such communications over the WAN may be facilitated by the monitoring device 200 (and / or the control hub 101, if also present) and the different device 110, each having a cellular modem, but optionally may additionally or alternatively have a Wi-Fi™ and / or Ethernet interface, for example.

[0169] In the embodiments exemplified herein, regardless of whether the local control hub 101 is included in the external system 102 or not, the external system 102 includes the remote server 103. However, it will be appreciated that the use of the remote server 103 may not necessarily be essential if the local control hub 101 is used.

[0170] A processing module 202 may be used to control the activation of the ranging active reflected wave detector 106 in accordance with embodiments of the present disclosure. The processing module 202 may, in some embodiments, be provided by the local control hub 101 (as illustrated in Figure 1) and / or by the remote server 103, but in other embodiments it can additionally or alternatively be provided by the monitoring device 200 (e.g. by a CPU of the monitoring device 200). Thus, processing module 202 may be provided in a single processing chip or a distributing processing system.

[0171] In an activated state (e.g. a higher power consumption operating mode) the ranging active reflected wave detector 106 operates to measure wave reflections from an environment. The ranging active reflected wave detector 106 is a multidimensional (ideally 3 dimensional) ranging detector (and may thus determine coordinates corresponding to locations of reflection). That is, in contrast with Doppler-only detectors, the ranging active reflected wave detector 106 may determine the location of an object (e.g. a person) in its field of view. This enables a person in the environment to be tracked and enables determinations to be made about a state and / or activity of a person in the environment using the ranging active reflected wave detector 106 and in some embodiments additional hardware e.g. the processing module 202. The environment may for example be an indoor space such as a room of a home, a nursing home, a public building or other indoor space. The ranging active reflected wave detector 106 may operate in accordance with one of various reflected wave technologies.

[0172] Preferably, the ranging ranging active reflected wave detector 106 is a radar sensor. The radar sensor 106 may use millimeter wave (mmWave) sensing technology. The radar is, in some embodiments, a continuous-wave radar, such as frequency modulated continuous wave (FMCW) technology. Such a chip with such technology may be, for example, Texas Instruments Inc. part number IWR6843. The radar may operate in microwave frequencies, e.g. in some embodiments a carrier wave in the range of l-100GHz (76-8 IGhz or 57-64GHz in some embodiments), and / or radio waves in the 300MHz to 300GHz range, and / or millimeter waves in the 30GHz to 300GHz range. In some embodiments, the radar has a bandwidth of at least 1 GHz. The ranging active reflected wave detector 106 may comprise antennas for both emitting waves and for receiving reflections of the emitted waves, and in some embodiment different antennas may be used for the emitting compared with the receiving.

[0173] The ranging active reflected wave detector 106 is able to observe a region (the detector’s observable region) 601 (see Figures 6, 9 and 10). The observable region 601 is a region in the environment that the ranging active reflected wave detector 106 is able to observe, when in its position for use and operational in that position, and may be defined by distance limitations (e.g. the maximum distance that is observable by the radar) and its field of view. Optionally the observable region may be software configurable one or both of: on installation and / or during use.

[0174] The active reflected detector 106 may monitor a fall detection region of interest 600 (see Figures 6, 9B, 9C, 10B, 12A, 12B, 13A and 13B). The fall detection region of interest 600 is software configurable one or both of: on installation and / or during use. The fall detection region of interest

[0175] 600 can be the same or preferably less than the observable region 601 in cases where the observable region is also software configurable on installation and / or during use. However, wherein the observable region 601 is not tailored on installation (e.g. it is factory configured), the fall detection region of interest 600 is less than the observable region 601. In particular, the fall detection region of interest 600 may be the same as the observable region 601 or may consist of one or more subregions of the observable region 601that make up less than an entirety of the observable region 601.

[0176] Each fall detection region of interest 600 is optionally determined by two-dimensional limitations (e.g. limitations to horizontal dimensions) or could be determined by three-dimensional limitations. In either case, each fall detection region 600 of interest can optionally be defined as covering a horizontal area. In the case of having three-dimensional limitations each fall detection region of interest 600 covers a volume. As described above, a total volume or total area (e.g. horizontal area) of the at least one fall detection region of interest 600 (i.e. for a single fall detection region of interest 600, the volume or area that it covers; for a plurality of fall detection regions of interest 600, a volume or area covered by the fall detection regions of interest 600 collectively) is the same or preferably less than a volume or area (e.g. horizontal area) of the observable region

[0177] 601 that is observable by the ranging active reflective wave detector 106. The fall detection region of interest 600 is optionally a subset of the region that is observable by the ranging active reflective wave detector. The monitoring device 200 can be configured to store a spatial characterization of a single fall detection region of interest 600 that is smaller than the observable region 601. The monitoring device 200 is optionally configured to store a spatial characterization of a plurality of fall detection regions of interest 600, wherein the total volume or area (e.g. horizontal area) of the fall detection regions of interest 600 is smaller than the volume or area (e.g. horizontal area) of the observable region 601.

[0178] In some examples, areas outside the fall detection region of interest 600 are not monitored using the ranging active reflected wave detector 106, even though the observable region 601 is greater than the fall detection region of interest 600. In other examples, any data from the ranging active reflected wave detector 106 relating to the region that is within the observable region 601 but outwith the fall detection region of interest 600 is simply disregarded or at least not used for fall detection. For example, in some embodiments, the monitoring may comprise generating reflective wave measurements (e.g. radar measurements) for all of the observable region 601 and subsequently reducing the set of measurements to be confined to the at least one smaller fall detection region of interest 600 that is under surveillance. However, in some embodiments, measurements from zones (“exclusion zones” such as exclusion zone 606 shown in Figures 9B, 9C, 10B, 12A, 12B, 13A, and 13B) excluded from the fall detection region of interest 600 are instead analyzed for other purposes, e.g. to detect a state or activity of a person when they are outside of the fall detection fall detection region of interest 600. Thus the observable region 601 may comprise a monitoring region that includes not only one or more fall detection regions of interest 600 but also one or more exclusion zones outside the fall detection region(s) of interest 600.

[0179] The monitoring by the ranging active reflected wave detector 106 (e.g. a radar) may be, or include, any one or more of observing, checking, processing or keeping a continuous record of reflective wave measurements (e.g. radar measurements).

[0180] As will be appreciated the ranging active reflected wave detector 106 is an “active” detector in the sense of it relying on delivery of waves from an integrated source in order to receive reflections of the waves. The ranging active reflected wave detector 106 is not limited to being a radar sensor, and in other embodiments alternative active reflected wave detectors may be used, for example the ranging active reflected wave detector 106 may be a lidar sensor, or a sonar sensor. The ranging active reflected wave detector 106 being a radar sensor is advantageous over other reflected wave technologies in that radar signals can be effective through some materials, e.g. wood or plastic, but not others - notably water which is important because humans are mostly water. This means that the radar, unlike sonar or lidar, may be able to detect a person in the environment 100 even if they are behind solid furnishings consisting of a material(s) transmissive to radar signals.

[0181] The local control hub or remote server 102 is in communication with a sensor 108. The sensor 108 can transmit data to, and receive data from, the local control hub or remote server 102 via a wired and / or wireless connection. The sensor 108 may be configured to detect presence of a person on a resting surface 602, as illustrated in Figure 7, within the fall detection region of interest 600 of the active reflected wave detector 106.

[0182] The resting surface 602 is a surface of an object for a person 604 to rest on top of, e.g. for lying, sitting or otherwise reclining on top of. The resting surface 602 is a surface elevated from a floor that, in examples, is a safe location, e.g. it is padded or impact absorbing such that a fall onto the resting surface is much less likely to be an issue relative to a fall onto a hard floor or other surface. Examples of a resting surface 602 could be or comprise a resting surface of one of: a mat, bed, mattress, futon, shikibuton, chair, seat, sofa, couch, chaise lounge, bench, stool, beanbag, cushion, floor cushion, pouffe, or the like. If a person 604 is resting on the resting surface 602, then they may be in a similar position to a fallen person 604c (see Figure 6), particularly if the resting surface 602 is close to the floor, as might be expected for a low bed, futon, couch, shikibuton or other floor mattress, bean bag, floor cushion, chaise lounge or the like. As such, the output of the ranging active reflected wave detector 106 based on reflections from the person 604 resting on the resting surface may meet the criteria for a fall, even though the person 604 on the resting surface is simply resting rather than having fallen.

[0183] In the arrangement of Figure 1, the ranging active reflected wave detector 106 is configured for use in detecting a fallen person 604c within a fall detection region of interest 600 and the sensor 108 is configured to identify if the fallen person is on the resting surface.

[0184] As such, by employing the sensor 108 to provide an indication that the person 604 that is detected by the ranging active reflected wave detector 106 is on a resting surface 602, then mitigating action against false alarms can be taken, e.g. automatically disregarding fall detection arising from the person 604a, 604b on the resting surface 602a, 602b. Possible mitigating actions are not limited to disregarding fall detections but could include, for example, the ranging active reflected wave detector 106 remaining or being switched into the inactive or low power state or otherwise disabled, such that fall detection is not operated.

[0185] The arrangement shown in Figure 1 has the ranging active reflected wave detector 106 as part of a monitoring device 200 that can be mounted at a premises, and the sensor 108 as a separate component, with both in communication with the local control hub or remote server 102. The output of the ranging active reflected wave detector 106 is processed, in the illustrated topology, by a processing module 202 of the external system 102 to determine a fall. The processing module 202 of the external system 102 is also arranged to receive the output of the sensor 108 to determine if the person giving rise to the detected fall is on the resting surface.

[0186] Although in the example of Figure 1, the processing is depicted as being performed by the processing module 202 on the external system 102, in some embodiments, such as those illustrated in Figures 2 to 4, the processing module 202 that performs the controlling and processing functionality described above is on-board the monitoring device 200 that comprises the ranging active reflected wave detector 106. For example, at least aspects of the processing module 202 that control the ranging active reflected wave detector may be provided by a CPU of the monitoring device 200. The CPU may also provide some or all of the processing functionality described above. Optionally some or all of the processing functionality described above may be performed using processing hardware that is shared by the ranging active reflected wave detector 106, but in other embodiments the processing module 202 and ranging active reflected wave detector 106 represent wholly distinct hardware components. In the examples of Figures 3 and 4, like the example of Figure 2 the processing module 202 is on-board the monitoring device 200 but there is no local control hub or remote server 102, and the monitoring device 200 communicates any fall alert directly to the operator device or system 110.

[0187] In more detail, Figure 2 illustrates a further example arrangement in which a local control hub or remote server 102 is in communication with a monitoring device 200.

[0188] The monitoring device 200 comprises the processing module 202 and the ranging active reflected wave detector 106, and may optionally further comprise the sensor 108, but in the embodiment illustrated, the sensor 108 is separate from, but in communication with, the monitoring device 200. The monitoring device 200 may use a communications module 203 (optionally a wired and / or wireless communications module, but in some embodiments at least a wireless communications module) to communicate with the operator device or system 110 via the local control hub 101 and / or remote server 103, but it is the processing module 202 of the monitoring device 200 which processes the signal from the active reflected wave detector 106 to detect any falls that occur in the fall detection region of interest 600 of the ranging active reflected wave detector 106.

[0189] The monitoring device 200 comprises a communications interface to facilitate communication with the local control hub or remote server 102 via a wired and / or wireless connection, which can in turn signal a fall alert to the operator device or system 110.

[0190] Figure 3 illustrates another example arrangement in which there is no local control hub interfacing with the monitoring device 200, and in which the monitoring device 200 communicates directly with the operator device or system 110. Again, in this arrangement it is the processing module 202 of the monitoring device 200 which controls the activation of the active reflected wave detector 106 in accordance with embodiments of the present disclosure.

[0191] In the example of Figure 3, the monitoring device 200 comprises a communications interface to facilitate communication with other devices such as the operator device or system 110 or a plurality thereof. Like in the case of Figure 2, the monitoring device 200 comprises the processing module 202 and the ranging active reflected wave detector 106, and may optionally further comprise the sensor 108, but in the embodiment illustrated, the sensor 108 is separate from, but in communication with, the monitoring device 200.

[0192] In each of the examples of Figures 1 and 2, the remote server 103 may optionally be provided to manage and handle messages from a plurality of monitoring devices 200 and / or active wave reflectors 106 installed at different locations and / or different premises.

[0193] The functionality of the processing module 202 described herein may be implemented in code (software) stored on a memory comprising one or more storage media, and arranged for execution on a processor comprising one or more processing units, which optionally may be distributed amongst different devices of the system disclosed herein, e.g. the active reflected wave detector 106 and / or the external system 102. Further, the processing module 202 may be at least partly implemented using processing resources that also form part of the active reflected wave detector 106 and / or the sensing device 108. The storage media may be integrated into and / or separate from the processing module 202. The code is configured so as when fetched from the memory and executed on the processor (e.g. one or more microprocessors, microcontrollers or any other type of code -reading processor, or combination thereof) to perform operations in line with embodiments discussed herein. Alternatively it is not excluded that some or all of the functionality of the processing module 202 or processing unit(s) thereof is implemented in dedicated hardware circuitry, or configurable hardware circuitry, e.g. an FPGA, and / or a combination of discrete analog and / or digital circuits and components.

[0194] However, although the sensor 108 is useful in determining if a person 604 is on the resting surface, the sensor 108 is not essential and other mechanisms could be used to take into account the resting surface, e.g. based on the output of the ranging active reflected wave detector 106 and / or on other data available to the processing module 202. Figure 4 shows such an example, in which the monitoring device 200 comprises the processing module 202 and the ranging active reflected wave detector 106, such that fall detection is performed on the monitoring device 200 but one or more exclusion regions (606a, 606b, see Figures 9 and 10) are set up to exclude the resting surfaces and the areas within a set or pre-set distance from it from the fall detection region of interest 600a, 600b (see Figures 9, 10, 12, and 13). The monitoring device 200 can communicate fall alerts to the operator device or system 110 upon determination of a fall. However, it will be appreciated that arrangements equivalent to those of Figures 1 and 2, but without the sensor 108, and instead alternative mechanisms for determining whether a person is resting on the resting surface could be used.

[0195] For each of the embodiments of Figures 1 to 4, in operation, the active reflected wave detector 106 performs one or more reflected wave measurements at a given moment of time, and over time these reflected wave measurements can be analysed by the processing module 202 to determine the presence of a person 604 and / or a state and / or activity of a present person and / or a condition of a present person. In other embodiments, some or all of such analysis may instead be performed upstream, for example by the local control hub 101 or remote server 103. In either case, the active reflected wave detector 106, in some embodiments, more specifically forms part of a fall detector.

[0196] In the context of the present disclosure, the state of the person 604 may be a characterization of the person 604 based on a momentary assessment. For example, a classification based on their position (e.g. in a location in respect to the floor and in a configuration which are consistent or inconsistent with having fallen) and / or their kinematics (e.g. whether they have a velocity that is consistent or inconsistent with them having fallen, or having fallen possibly being immobile). In the context of the present disclosure, the condition of the person may comprise a determination of an aspect of the person’s health or physical predicament, for example whether they are in a fall condition whereby they have fallen and are substantially immobile, such that they may not be able (physically and / or emotionally) to get to a phone to call for help. In some embodiments this involves an assessment of the person’s status over time, such as in the order or 30-60 seconds. However, the condition of the person may in some contexts be synonymous with the status of the person. For example, by determining that the person is in a safe supported state or a standing state, it may be concluded that the person is not currently in a fall condition, whereby they are on the floor and potentially unable to seek help. It may additionally or alternatively be concluded that they are in a resting condition because of their status being determined to be in a safe supported state, e.g. lying on a bed. In another example their condition may be classified as active and / or mobile based on a determination of a walking status.

[0197] The determining herein of whether there is a fallen person may comprise determining that a person is in a condition in which they are assessed, based at least on an output of the ranging active reflected wave detector, to have fallen. Examples of suitable techniques for assessing whether a person is in a fall position (i.e. a position that is consistent with them haven fallen) or a non-fall position (indicative that they are, at least temporarily, in a safe state) are described in International patent application publication WO / 2021 / 137220 and WO / 2021 / 245673 in the name of the present applicants, the contents of which are incorporated by reference in their entirety as if set out in full herein. However, the present disclosure is not limited to this and other suitable techniques for determining whether a person has fallen using the ranging active wave detector 106 could be used. When it is determined that the person has fallen, then at least one fall detection response action is taken responsive to the determination that the person is in a fall position. The at least one fall detection response action may comprise generating a fall detection alert locally (e.g. via audio and / or visual output device on the monitoring device 200) and / or generating a notification that is transmitted to another device. The fall detection alert could be transmitted, for example, directly or indirectly from the monitoring device 200 to the external system 102 and / or to the operator device or system 110. Whether the alert is provided locally or remotely, the alert may require a response that involves a human. This response may for example comprise at least an acknowledgement of the alert. Where an alert is provided locally this may comprise playing of a prestored audio enquiry to ask if the person has fallen and if they are alright before listening for a verbal response (optionally using a speaker and microphone, respectively, incorporated on the monitoring device 200).

[0198] Figure 5 illustrates a technique for detecting a fall using the ranging active reflected wave detector 106. However, there are various techniques for fall detection using the ranging active reflected wave detector 106 that may be understood by a skilled person based on the disclosure of the present application and normal skill in the art, and the present disclosure is not limited to the example of Figure 5.

[0199] Figure 5 illustrates a free-standing human body 104 with indications of reflective wave reflections therefrom in accordance with some embodiments.

[0200] For each reflected wave measurement, for a specific time in a series of time-spaced reflected wave measurements, the reflected wave measurement may include a set of one or more measurement points that make up a “point cloud”, the measurement points representing reflections from respective reflection points from the environment (e.g. from the ranging active reflected wave detector’s observable region 601 or fall detection region of interest 600a, 600b within the environment). In embodiments, the active reflected wave detector 106 provides an output to the processing module 202 for each captured frame as a point cloud for that frame. Each point 302 in the point cloud may be defined by a 3-dimensional spatial position from which a reflection was received, and defining a peak reflection value, and a Doppler value from that spatial position. Thus, a measurement received from a reflective object may be defined by a single point, or a cluster of points from different positions on the object, depending on its size.

[0201] In some embodiments, such as in the examples described herein, the point cloud represents only reflections from moving points of reflection, for example based on reflections from a moving target. That is, the measurement points that make up the point cloud represent reflections from respective moving reflection points in the environment. This may be achieved for example by the active reflected wave detector 106 using moving target indication (MTI). Thus, in these embodiments there must be a moving object in order for there to be reflected wave measurements from the active reflected wave detector (i.e. measured wave reflection data), other than noise Alternatively, the processing module 202 receives a point cloud from the active reflected wave detector 106 for each frame, where the point cloud has not had pre-filtering out of reflections from moving points. Preferably for such embodiments, the processing module 202 filters the received point cloud to remove points having Doppler frequencies below a threshold to thereby obtain a point cloud representing reflections only from moving reflection points. In both of these implementations, the processing module 202 accrues measured wave reflection data which corresponds to point clouds for each frame whereby each point cloud represents reflections only from moving reflection points in the environment. In some embodiments, measured wave reflection data may comprise signals received from an array of transducers (e.g. antennas) and / or may be represented by analog or digital signals that precede a digital signal processing (dsp) component of the apparatus. For example, even in embodiments that generate a point cloud, the measured wave reflection data may be data that precedes calculation of the point cloud by the dsp component.

[0202] The region for which the active reflected wave detector is capable of identifying reflections of waves emitted from the active reflected wave detector is a way to define the observable region 601 that is observable by the ranging active reflected wave detector 106. This is consistent with the above description of the observable region 601 being a region in the environment that the ranging active reflected wave detector 106 is able to observe, when in its position for use and operational in that position, and may be defined by distance limitations (e.g. the maximum distance that is observable by the radar) and its field of view. Optionally the observable region may be software configurable one or both of: on installation and / or during use.

[0203] In other embodiments, no moving target indication (or any filtering) is used. In these implementations, the processing module 202 accrues measured wave reflection data which corresponds to point clouds for each frame whereby each point cloud can represent reflections from both static and moving reflection points in the environment.

[0204] Figure 5 illustrates a map of reflections. The size of the point represents the intensity (magnitude) of energy level of the radar reflections (see larger point 306). Different parts or portions of the body reflect the emitted signal (e.g. radar) differently. For example, generally, reflections from areas of the torso 304 are stronger than reflections from the limbs. Each point represents coordinates within a bounding shape for each portion of the body. Each portion can be separately considered and have separate boundaries, e.g. the torso and the head may be designated as different portions. The point cloud can be used as the basis for a calculation of a reference parameter or set of parameters which can be stored instead of or in conjunction with the point cloud data for a reference object (human) for comparison with a parameter or set of parameters derived or calculated from a point cloud for radar detections from an object (human).

[0205] When a cluster of measurement points are received from an object in the environment, a location of a particular part / point on the object or a portion of the object, e.g. its centre, may be determined by the processing module 202 from the cluster of measurement point positions having regard to the intensity or magnitude of the reflections (e.g. a centre location comprising an average of the locations of the reflections weighted by their intensity or magnitude). As illustrated in Figure 5, the reference body 104 has a point cloud from which its centre has been calculated and represented by the location 308, represented by the star shape. In this embodiment, the torso 304 of the body is separately identified from the body and the centre of that portion of the body is indicated. In alternative embodiments, the body can be treated as a whole or a centre can be determined for each of more than one body part e.g. the torso and the head, for separate comparisons with centres of corresponding portions of a scanned body.

[0206] In one or more embodiments, the object’s centre or portion’s centre is in some embodiments a weighted centre of the measurement points. The locations may be weighted according to a Radar Cross Section (RCS) estimate of each measurement point, where for each measurement point the RCS estimate may be calculated as a constant (which may be determined empirically for the reflected wave detector 106) multiplied by the signal to noise ratio for the measurement divided by R4, where R is the distance from the reflected wave detector 106 antenna configuration to the position corresponding to the measurement point. In other embodiments, the RCS may be calculated as a constant multiplied by the signal for the measurement divided by R4. This may be the case, for example, if the noise is constant or may be treated as though it were constant. Regardless, the received radar reflections in the exemplary embodiments described herein may be considered as an intensity value, such as an absolute value of the amplitude of a received radar signal.

[0207] In any case, the weighted centre, WC, of the measurement points for an object may be calculated for each dimension as:

[0208] Where:

[0209] N is the number of measurement points for the object;

[0210] Wnis the RCS estimate for the nthmeasurement point; and

[0211] / A is the location (e.g. its coordinate) for the nthmeasurement point in that dimension.

[0212] The location of the weighted centre can be indicative of a fall, e.g. if it is within a threshold distance of the floor. Analysis of the distribution of the point cloud may also provide information indicative of a fall, e.g. whether the person is moving or stationary or whether the person is in a regular or other predetermined position or in an irregular position or other position indicative of having fallen. Although a point cloud comprising many points is shown in Figure 5, in many instances, the point cloud can be considerably more sparse. This may result in errors in fall detection. In some embodiments locating a person may involve a tracking algorithm across a plurality of time sequential measurement frames, thereby filtering out noise detections, and / or providing more accurate detection and locating of a person.

[0213] Systems that rely, at least in part, on determining how close a person is to the floor in order to determine a fall can be subject to false positive fall detections due to a person merely resting on a resting surface that is low to the ground. Such a low resting surface could be, but not limited to, one of: a couch, a futon, a shikibuton or other floor mattress, a beanbag, a chaise lounge, a stool, a low bed, or the like.

[0214] Another potential cause of errors is a person sitting, e.g. on a chair, stool, couch or the like. In this case, the feet of the person could be moving but their body is motionless, or it could be that their feet are in the observable region 601 of the ranging active wave detector 106 but their body is outside of the observable region 601. In these cases, and others, most of the reflections of the active wave emitted by the ranging active wave detector 106 and reflected by a person may be from the person’s feet that are proximate to the ground. This may give rise to false fall detection events.

[0215] These are just a few examples of situations that could lead to false alarms in certain zones of the observable region 601. There could be wide range of situations or other factors that could give rise to false detections in one or more particular zones. As such, functionality for dynamically determining and reconfiguring regions of interest 600a, 600b and / or exclusion zones 606 based on the location of determined or inferred false detections may be beneficial.

[0216] Figure 6 illustrates the operation of a system such as those of any of Figures 1 to 3 that comprises the ranging active reflected wave detector 106 that is configured to monitor an area of interest 600 within an environment. The processing module 202 (see Figures 1 to 4) is configured to process a signal from the ranging active reflected wave detector 106 to determine if the signal from the ranging active reflected wave detector 106 is indicative of a fallen person. The environment comprises a first resting surface 602a monitored by a first sensor 108a and a second resting surface 602b monitored by a second sensor 108b. The first and second resting surfaces 602a, 602b are also within the area of interest 600 of the ranging active reflected wave detector 106. The sensors 108a, 108b are configured to determine if a person 604a, 604b is resting on the respective resting surface 602a, 602b. The processing module 202 (see Figures 1 to 4) may be configured to not operate fall detection, to abort fall detection, or may disregard the fall detection by not generating an alert based on fall detection, if one of the sensors 108a, 108b indicates that the person 604a, 604b is present on the respective resting surface 602a, 602b.

[0217] In examples, each resting surface 602a 602b has its own dedicated sensor 108a, 108b that are provided as separate devices in separate housings, which may or may not be the same type of sensor.

[0218] The first resting surface 602a in this example is in the form of a low bed monitored by the first sensor 108a. The second resting surface 602b in this example is in the form of a shikibuton or floor mattress monitored by the second sensor 108b. In this example, the first sensor 108a monitoring the first resting surface 602a is a pressure sensor configured to detect pressure above a threshold amount associated with a person 604a upon the first resting surface 602a. The first sensor 108a and first resting surface 602a are shown in greater detail in Figure 7, in which the weight of a person 604a on the first resting surface 602a applies a pressure above a threshold on the pressure sensor 108a, resulting in a signal being output from the pressure sensor 108a indicative of the person 604a being on the first resting surface 602a. In this example, the second sensor 108b monitoring the second resting surface 602b is a different type of sensor to the first sensor 108a, for example, another active reflected wave detector or a PIR motion detector, or the like, focussed on the resting surface 602b. However, each sensor 108a, 108b regardless of which type is used, provides a signal from which it can be determined whether or not a person 604a, 604b is on the respective resting surface 602a, 602b. The present disclosure is not limited to the above arrangement and other types of sensors for determining whether or not a person is on a resting surface can be used. In this way, the output of the sensors 108a, 108b can be used to determine if a fall detection is due to a person 604a, 604b simply resting on one of the resting surfaces 602a, 602b or a person 604c who has actually fallen.

[0219] A method 1000 of operation of a fall detection system, such as those of Figures 1 to 4, that utilise sensors 108a, 108b to detect the presence of a person 604a, 604b on one or more resting surfaces 602a, 602b, such as those illustrated in Figures 6 and 7, is shown in Figure 8A with reference to Figure 6.

[0220] In step S 1002 the ranging active reflective wave detector 106 of the monitoring system 200 is used to monitor an area of interest 600 that covers at least part of an environment for a fall. In step S1003, the output of the ranging active reflective wave detector 106 of the monitoring system 200 is processed by the processing module 202 in order to determine if a person 604a, 604b, 604c has fallen. For example, the processing module 202 may determine a height from a floor of the person 604a, 604b, 604c, e.g. the height from the floor of a weighted centre of the person 604a, 604b, 604c or a height of the highest point of the person 604a, 604b, 604c. A fall condition used by the processing module 202 to determine that the person 604a, 604b, 604c has fallen may comprise that the height from the floor of the person 604a, 604b, 604c is below a threshold height for more than a threshold period of time. If no fallen person is detected, the process returns to step S 1002 and the ranging active reflective wave detector 106 continues to monitor the fall detection region of interest 600 in the environment.

[0221] If a fallen person is detected in step S1004 then, in step S1006, the processing module 202 checks if a signal indicative of a person 604a, 604b resting on one of the resting surfaces 602a, 602b has been received from one of the sensors 108a, 108b. A signal from one of the sensors 108a, 108b indicative of a person 604a, 604b resting on one of the resting surfaces 602a, 602b is an indication that the fall detection in S1004 may be false due to detecting a person 604a, 604b simply lying on one of the resting surfaces 602a, 602b. Even if the fallen person detected by the ranging active reflective wave detector 106 is not the person 604a, 604b resting on the resting surface 602a, 602b, then at least there are other people in the vicinity of the fallen person who can assist and / or raise an alarm as required.

[0222] If it is determined that the person 604a, 604b is resting on one of the resting surfaces 602a, 602b, then the fall detection is disregarded in step S1008 and the process returns to monitoring the area of interest 600 in step S 1002. If no signal from the sensors 108a, 108b indicative of a person 604a, 604b resting on the resting surfaces 602a, 602b is received, then in step S1010 the processing module 202 makes a fall detection response action, comprising raising a fall alert, e.g. by sending an alert raising signal to the operator device or system 110 or to the external system 102.

[0223] In an alternative to step S1008, if it is determined that the person 604a, 604b is resting on one of the resting surfaces 602a, 602b, then automated confirmation may be sought before the response action is taken to generate an alert. The alert may comprise notifying the person 604a (e.g. using a speaker on monitoring device 200) that a fall has been detected. A user interface may be provided in which the person is invited to cancel the fall detection within a certain period and, if they do not manually cancel the fall detection, then a fall alert is also issued to the external system 102 and / or the operator device / system 110. In an example, the alert notifying the person 604a may comprise an automated voice may query the person to see if they have fallen and if they are ok and to employ an audio sensor and associated audio processing, e.g. speech processing, to determine from the response to the query (or lack of response) if a fall has occurred. For the automated confirmation, if employed, may take the form of a fall detection redundancy, i.e. an additional sensor may be utilised to make a further determination if a fall has or has not taken place, whereby a fall alert is only raised if both sensors agree. In other examples, even the outcome of confirmation actions may be disregarded or not performed if a person is detected by the sensor 108.

[0224] In the above way, false positive fall detections due to a person 604a, 604b resting on a resting surface 602a, 602b, particularly a resting surface 602a, 602b close to the floor, may be reduced.

[0225] Figure 8B illustrates an alternative process 2000 to that of Figure 8A. In Figure 8B, rather than not actioning a fall detection event in the event of one of the sensors 108 a, 108b indicating that someone is on one of the resting surfaces 602a, 602b, the process of operating the ranging active wave detector 106 to determine a fall is prevented or supressed entirely or at least for the region associated with the occupied resting surface 602a, 602b.

[0226] In step S2002, it is determined if a signal from a sensor 108a, 108b indicative that a person is located on a resting surface has been received. If so, then the ranging active wave detector 106 is kept or switched into a sleep or low power non-emitting mode in step S2004. If no signal from a sensor 108a, 108b indicative that a person is located on a resting surface has been received, then the ranging active wave detector 106 is operated to monitor at least part of an environment for falls in step S2006. If a fall is detected using the ranging active wave detector 106 in step S2008, then one or more fall detection response actions, such as those discussed above, are taken. If no fall is detected, then the process loops back around to step S2002.

[0227] The process 2000 of Figure 8B may be more energy efficient than that of Figure 8A, as operation of the ranging active wave detector 106, which can be power intensive, is prevented if the resting surface 602a or 602b is occupied, which may otherwise give rise to false fall detections.

[0228] Figures 9A to 9C illustrate an arrangement in which the area of interest 600 is controlled, optionally dynamically, to avoid areas around locations that are likely to give rise to false positive detections, such as but not limited to those around resting surfaces 602a, 602b. This arrangement could be implemented using the system of Figure 4, but is not limited to this. Selected examples of alternative systems are discussed above in relation to Figure 4. The approach illustrated in Figures 9A to 9C can also be used in conjunction with the systems shown in Figures 1 to 3, the arrangement of Figure 6 and the method of Figure 8 A or 8B, e.g. instead of one or more of the sensors 108a, 108b.

[0229] Figure 9A illustrates an environment in which the monitoring device 200, including the ranging active wave detector 106, may be operable to monitor for falls. The processing module 202 is configured to process a signal from the ranging active reflected wave detector 106 to determine if the signal from the ranging active reflected wave detector 106 is indicative of a fallen person. The environment comprises a first resting surface 602a and a second resting surface 602b. The processing module 202 is configured to determine locations in the environment that result in false fall detections (in some case at least a predetermined amount of false fall detections), such as but not limited to regions around the resting surfaces 602a, 602b, and use these to create and / or adapt one or more exclusion zones 606 and to exclude the one or more exclusion zones 606 from the area of interest 600. The exclusion zones 606 can be stored in one or more spatial maps. The area of interest 600 monitored by the ranging active reflected wave detector 106 that can give rise to a fall alert is configured to exclude the one or more exclusion zones 606 around locations tending to giving rise to false fall detections. The exclusion zones 606 are usually specific to the particular environment in which the monitoring device 200 is installed. In examples, the exclusion zones 606 can be set at installation of the monitoring device or subsequently, e.g. by selecting an exclusion region setup mode, or the like. However, preferably the exclusion zones can be dynamically and / or adaptively determined using the process outlined in Figure 11 and described below. The one or more exclusion zones 606, 606a, 606b can optionally take the form of a bounding region (illustrated as a bounding box in Figure 11) extending by a set or pre-set distance from locations determined to give rise to false fall detections, such as the resting surface or surfaces 602a, 602b.

[0230] Figure 9A illustrates a perspective view of the environment and Figure 9B illustrates a plan view of an example of how a region of interest is used.

[0231] In the example of Figures 9A to 9C, a single exclusion zone 606 is used to cover all of the locations determined to give rise to false fall detections, such as the resting surfaces 602a, 602b. In Figure 9B, an exclusion zone 606 around the locations determined to give rise to false fall detections (in this example, the resting surfaces 602a, 602b) is determined. A region of interest 600a that is monitored for falls may be determined by subtracting or otherwise removing the exclusion region 606 from the observable region 601 of the ranging active wave detector 106.

[0232] Figure 9C illustrates an alternative example, in which the region of interest 600a is positively defined and is defined in such a way that it excludes the region (i.e. the exclusion zone 606) around the locations determined to give rise to false fall detections (e.g. around the resting surfaces 602a, 602b). In addition, in this example, the region of interest 600a is positively defined to include only a part that is less than all of the observable region 601 of the ranging active wave detector 106 that doesn’t correspond to the exclusion zone 606.

[0233] An alternative arrangement is shown in Figures 10A and 10B, which are similar to the arrangements of Figures 9A to 9C, other than a plurality of exclusion zones 606a, 606b are set-up, with each exclusion zone 606a, 606b being used to encompass a different resting surface 602a, 602b or other location determined to give rise to false fall detections. In this case, the plurality of exclusion zones 606a, 606b are excluded from the observable region 601 of the ranging active wave detector 106 to form the area of interest 600 that is monitored for falls by the ranging active wave detector 106. The sum of the volumes of the exclusion zones 606a, 606b, is less than the observable region 601 of the ranging active reflected wave detector 106. Although the example of Figures 10A and 10B corresponds to the arrangement of Figure 9B, the approaches discussed above in relation to Figure 9C could be used in other examples.

[0234] The examples of Figures 9 and 10 give examples of a resting surface 602a, 602b such as a low- lying bed as a location associated with a high potential for false fall detections. Figures 11 A and 11B give other examples of a chair as a location associated with a high potential for false fall detections. As discussed above, the feet of a person sitting on a chair or stool or the like may give rise to detections that are similar to those of someone who has fallen to the ground. In the example of Figure 10A, an exclusion zone 606c is set up around the whole of the chair and the rest of the observable region 601 of the ranging active wave detector 106, excluding the exclusion zone 606c, forms the observable region 600a that is monitored for falls using the ranging active wave detector 106. In the example of Figure 10A, the location or even the presence of the chair may not be known to monitoring device 200, but it may be inferred based on the locations of false detections. It will be appreciated then that even if the intention is to generate an exclusion zone that includes all the chair, it may be that some of the chair (for example a part(s) further from where the false detections occurred) remains outside the exclusion zone. In the example of Figure 10B, an exclusion zone 606c is set up around only the feet of the person sat on the chair and the rest of the observable region 601 of the ranging active wave detector 106, excluding the exclusion zone 606c, forms the observable region 600a that is monitored for falls using the ranging active wave detector 106.

[0235] Figures 12A and 12B show examples where multiple regions of interest 600a, 600b are monitored for falls. In the example of Figure 12A the observable region 601 of the ranging active wave detector 106 is further limited to a monitoring region 618 that is a subset, i.e. less than, the observable region 601 of the ranging active wave detector 106. This may be done manually, for example, e.g. by inputting coordinates of the monitoring region 618, by moving a special object whose location can be detected such as a reflective or emitting tool around a perimeter of the monitoring region 618, or the like. In this example, there is a region the regions of interest 600a and 600b that is determined to be likely to give rise to false fall detections, and therefore is defined as an exclusion region 606b. Two regions of interest 600a and 600b are positively defined on either side of, and separated by, the exclusion zone. In this example, the total volume or area of the regions of interest 600a and 600b is less than the volume or area of the monitoring area 618 after the exclusion zone 606b has been excluded. It may for example that an installer defines portion of the monitoring region 618 for fall detection monitoring, the portion encompasses regions the regions of interest 600a, 600b and the exclusion zone 606b before it is yet known that false detections tend to occur in that region 606b. An example of such an arrangement is depicted in Figure 12B. It may therefore be determined, e.g. dynamically during use, that false detections occur in that region 606b, which may thereafter may be treated as an exclusion zone. Optionally the part of the monitoring region that is outside the regions of interest 600a and 600b and outside the exclusion zone 606b may be used to monitor for other activities or events, but not for falls. The regions of interest 600a and 600b may optionally be used to monitor for those other activities or events in addition to monitoring for falls.

[0236] Figure 13 shows a flowchart of a method 900 of configuring the ranging active reflected wave detector 106, such as but not limited to that shown in Figures 4, 9 and 10, to determine the regions of interest 600 by determining and excluding one or more exclusion zones. Such determinations may be performed dynamically and adaptively.

[0237] In step S902, measurements made using the ranging active reflected wave detector 106 are used to identify at least one first location where a respective fall detection event is determined to be a false fall detection. In some examples, this could comprise identifying at least one first event as a respective fall detection event from measurements from the ranging active reflective wave detector 106. Thereafter, at least one event location is determined from measurements from the ranging active reflective wave detector, wherein the at least one event location corresponds to the at least one first event. It is then determined whether the at least one first event is a false fall detection and in an event that the at least one first event is determined to be a false fall detection, the at least one event location is identified as being said at least one first location where a respective fall detection event is determined to be a false fall detection.

[0238] There are various ways of determining whether a fall detection event is a false fall detection.

[0239] In examples, false fall detections are based on a received response that is indicative that there has been a fall, or indicative that there has not been a fall. The response may be a manually provided response, such as an audible response, to a prompt, such as a voice prompt, that is output responsive to a fall event being detected. The response could alternatively be received by a button push on an input device, such as a wearable input device, or the like. In some implementations, however, the response may come from the operator system 110 (optionally via the external system 102) after the operator system 110 receives (optionally via the external system 102) an alert in response to the fall detection and speaks with the relevant person at the premises or an on-site carer of the person, optionally via the monitoring device 200. In these examples, operator-monitored (i.e. using operating system 110) fall detection may be provided for a person immediately after installation, and any relevant exclusion zones determined dynamically during use.

[0240] In other examples, false fall detections are based on an assumption that no falls occurred during a period of time. During a period where it is assumed or confirmed that no falls have occurred, event locations giving rise to at least one first event, particularly to a plurality of first events or clusters of first events, may be indicative of something at the one or more of the at least one first locations gives rise to false fall detection events, such that subsequent events detected at those event locations are more likely to be fall detection events falsely detecting a fall. As such, the above approach may be a convenient way of identifying whether the at least one first event is a false fall detection regardless of the mechanism that causes the false detection. In these examples, operator- monitored fall detection may be delayed until after the period of time, but with the advantage that when operator-monitored fall detection then commences it is less likely to involve false detection alerts being received by the operator system 110.

[0241] Although examples of detecting false fall detections are given above, the present disclosure is not limited to these and other suitable techniques could be used. In step S904, an exclusion zone 606 associated with the identified at least one first location is determined.

[0242] In step S906, subsequent to identifying the at least one first location, the ranging active reflective wave detector 106 is used to monitor for falls.

[0243] Using the ranging active reflective wave detector 106 to monitor for falls comprises determining whether a second location of a detected fall derived from measurements made by the ranging active reflected wave detector 106 is in a region of interest that is outside the exclusion zone 606 associated with the at least one first location determined in step S904.

[0244] In step S908, it is determined if a fall has been detected. If no fall is detected, then the system continues to monitor for falls using the ranging active reflective wave detector 106, returning to step S906.

[0245] If a fall is detected in step S908, then in step S910 it is determined if a second location (of a person) is in a region of interest that is outside of the exclusion zone or zones determined in step S904.

[0246] While it may be that spatial bounds of the / each exclusion zone may be determined in order to determine whether the second location is inside the exclusion zone, this need not necessarily be the case. That is, in some embodiments it may be determined whether a second location is within an exclusion zone without determining the bounds of the exclusion zone. This may be achieved for example by determining whether the second location is part a cluster of locations comprising first locations, e.g. as described herein, which infers that the second location is within an exclusion zone.

[0247] Thus, it will be appreciated that step S904 need not exist (e.g. step S910 rather step S906 may follow directly from step S902) and instead step S910 may more generally be determining whether a second location derived from ranging active reflected wave detector measurements is in a region of interest that is outside an exclusion zone, wherein the exclusion zone is associated with the at least one first location.

[0248] If a fall is detected in a region of interest that is outside the exclusion zone 606, then at least one fall detection response action is performed at step S910 which comprises generating a fall alert. The generating of the fall alert may comprise local alerts and / or alert notification transmitted to other devices to provide a remote alert response. The remote alert response may involve a verbal conversation between the person 604c and remote conversation partner, which may comprise a human and / or artificial intelligence.

[0249] In the example of Figure 13A, the ranging active reflective wave detector 106 is used to monitor for falls and, only once a fall has been detected, is it determined if the fall occurred in a region of interest. However, this need not be the case, and the monitoring for falls may be conditional on, or at least occurs after, it being determined that an entity such as a person is present in the area of interest. An example of this is illustrated in Figure 13B. Initial steps S902 and S904 of the method of Figure 13B are the same as those described above in relation to Figure 13A in which at least one first location where a respective fall detection event, determined from ranging active reflected wave detector measurements, is determined to be a false fall detection in step S902 and an exclusion zone associated with the identified at least one first location is determined in step S904.

[0250] However, rather than proceeding to monitor for falls using the ranging active reflected wave detector and then determining the second location once a fall has been detected, as in the method of Figure 13A, the method of Figure 13B then monitors for entities, e.g. people, prior to monitoring falls. The monitoring for entities may use a sensor other than the ranging active reflected wave detector, such as the pressure sensor 108a shown in Figure 6 or a PIR motion detection sensor such as sensor 108b shown in Figure 6, or the like. Alternatively or additionally, the ranging active reflected wave detector 106 could be used to monitor for entities.

[0251] If an entity is identified by the monitoring in step S914, it is determined in step S910’ if the second entity is present in the region or regions of interest. In this case, the second location is the location of the entity (e.g. person) but the person at this stage has not yet fallen. If it is determined in step S910’ that none of the detected entities are in the region or regions of interest, then the process returns to step S914 and monitoring continues. However, if it is determined in step S910’ that there is at least one entity in one of the regions of interest, then in step S906’, conditional on there being an entity in a region of interest, the fall monitoring is performed. The fall monitoring comprises operating ranging active reflective wave detector to generate ranging active reflected wave measurement data sufficient to identify a fall, which may in some embodiments involve more data than only identifying and / or tracking the location of an entity. The fall monitoring also comprises analyzing the measurement data (e.g. using a fall classifier) to identifier when the measurement data is representative of a fall apparently having occurred. The analysis, and optionally the operating ranging active reflective wave detector to generate the ranging active reflected wave measurement data, may be avoided if it is determined that a person is in an exclusion zone.

[0252] If a fall in the region or regions of interest is detected in step S908’, then a fall detection event is generated in response to the detected fall in step S912. Conversely, if no fall in the region or regions of interest is detected, then the process continues by monitoring for falls in the region or regions of interest. This may occur by returning to step S910’ to determine whether the entity is still in the region or regions of interest. For example, an updated second location for the entity (e.g. as next determined from tracking the entity) may be assessed, e.g. to determine whether it is inside an exclusion zone.

[0253] This process of Figure 13B may be more energy efficient, as the ranging active reflective wave detector may have a higher power requirement than the other sensor, such as sensors 108a, 108b, or the like, or it may consume more power to operate the ranging active reflective wave detector in a manner for generating sufficient data for fall detection compared with only identifying and / or tracking the location of an entity . According to this process in which monitoring for falls using the ranging active reflective wave detector selectively occurs responsive to determining that there is an entity in the region or regions of interest, the time in which the ranging active reflective wave detector is operational or fully operational or whose output is processed, may be reduced relative to the process of Figure 13 A. Optionally, fall detection may proceed in full even if the person is determined to be in an exclusion zone but the device 200 may merely not action a response to the fall detection, e.g. by disregarding the fall detection.

[0254] Figure 14 is a more specific example of the process shown in Figure 13A. However, the specific implementation details of the process of Figure 14 may also be applicable to a process such as that shown in Figure 13B in which the location of the object may be detected before detecting a fall. Specifically, applied to the process of Figure 13B the presence and location of an entity, such as the presence of an entity in the region or regions of interest, can be detected and the monitoring for fall detection using the ranging active reflective wave detector may be selectively carried out responsive to the determination that there is at least one entity in a region of interest.

[0255] In step 3002, at least part of an environment covered by the observable region of the ranging active reflected wave detector 106 is monitored for events. In step 3004, an event presenting as a fall detection based on the output from the ranging active reflected wave detector 106 is detected. In step 3006, the location of the detected event, i.e. a second location, is determined from the output of the ranging active reflected wave detector 106. The ranging active reflected wave detector 106 is particularly suited to identifying locations of objects. It will be understood however, that the location of the object may be detected before or as part of detecting an event or whether the event presents as a fall detection event. For example, in an alternative to step 3004, the event could simply be identifying the presence of an entity, e.g. person, with respect to one or more exclusion zones without that person having yet fallen. In an alternative to step 3006, the location of the entity may be determined before they have fallen.

[0256] In a process generally labelled 3008, it is determined if the event detected in step 3004 is outside of an exclusion zone 606 based on the locations of previous fall detections that have been determined to be false fall detections.

[0257] In this example, the process 3008 of determining if the event detected in step 3004 is outside of an exclusion zone 606 comprises running a clustering algorithm at step 3010 on the location of the event detected in step 3004 and any first locations of previously detected events that have been determined to be false fall detections.

[0258] The running of the clustering algorithm may exclude first locations of previously detected events that are older that a threshold age, such as older than 2 months, and have not been part of an exclusion zone within a threshold period of time, e.g. immediately past two months. In this way, if the cause of the false detections is removed, e.g. due to a change of furniture configuration in the environment, then monitoring device 200 will eventually adapt to start or again monitor for falls in that region that had hitherto been an exclusion zone and take any needed actions in response to detected falls. Disregarding old first locations, even if such first locations never were determined to be part of an exclusion zone (e.g. because they were one-off false detections remote from any other false detections), may have the advantage of mitigating against the possibility of forming exclusion zones where there is not truly an elevated likelihood of false fall detections.

[0259] Examples of suitable clustering algorithms include a DBscan clustering algorithm, but are not limited to that. In step 3012 of the determination 3008 of whether the event detected in step 3004 is outside of an exclusion zone 606, it is determined using the clustering algorithm if the location of the event detected in step 3004 (i.e. a second location) is in a cluster having at least N points, where one of the N points is the location of the event detected in step 3004 and the other N-l points are locations of previous fall detections that have been identified as being false fall detections (and have not been disregarded based on their age). In this example, N is 3 and the location of the event detected in step 3004 would need to form a cluster with at least two locations of false fall detections by the clustering algorithm in order for it to be determined that it is within an exclusion zone 606 associated with the N-l locations of previous fall detections that have been identified as being false fall detections. If the event detected in step 3004 does not form a cluster with at least N-l locations of previous fall detections that have been identified as being false fall detections, then it is determined that the event detected in step 3004 is outwith any exclusion zone 606 associated with locations of previous fall detections that have been identified as being false fall detections.

[0260] If the location of the event detected in step 3004 does not form a cluster with at least N-l locations of previous fall detections that have been identified as being false fall detections then, in step 3013, all points in the cluster (i.e. the location of the event detected in step 3004 and the at least N-l locations of previous fall detections that have been identified as being false fall detections that form a cluster) are flagged as being an active exclusion zone, with a time stamp of the latest detected event.

[0261] In step 3014 the event may be recorded to be used as a prior false fall detection event in a future iteration of the process. The process optionally comprises a recency criterion on the prior false fall detection events. In step 3016, it is determined if a number M of points in the cluster of prior false fall detections to which the event detected in step 3004 was determined to belong has more than a maximum X number of locations. If so, then in step 3018, the oldest location in the cluster is deleted and replaced by the location of the event detected in step 3004. Any remaining locations of prior false fall detections are made available for use in the next iteration of the process at 3008. Disregarding the oldest location in the cluster in this way results in a sliding window of false detection locations that enables the location, size and / or shape of the exclusion zone to adapt with time. This may be particularly useful for example if an item of furniture that is contributing to the false detections has its position shifted slightly.

[0262] If, in step 3012, it is determined that the event detected in step 3004 does not form a cluster with at least N-l locations of previous fall detections that have been identified as being false fall detections, i.e. it is determined that the event detected in step 3004 is outwith any exclusion zone 606 associated with locations of previous fall detections that have been identified as being false fall detections, then it is concluded that a fall has been detected and an associated detected fall action (i.e. an action in response to a detected fall) is performed at step 3020. In embodiments where it is determined if an entity is present and if the entity is located in an exclusion zone (e.g. whether the entity forms a cluster with a plurality of first locations of prior false fall detections) before performing monitoring for falls, then there may be additional steps between 3012 and 3020 of selectively configuring the ranging active reflective wave detector to monitor for falls to detect if the entity not in an exclusion zone falls.

[0263] It will appreciated that identifying whether the event detected in step 3004 is or is not part of such a cluster, especially if using a clustering algorithm like DBscan, will not generally result in the exclusion zone being or comprising a neat shape such as circle or a rectangle as or other parallelogram, like illustrated in the Figures 9A-C, 10A-B, 11A-B, or 12A-B. Rather the exclusion zone may me an irregular shape. Further it need not be necessary for the monitoring device 200 to determine the shape of the exclusion. It need merely determine whether the event detected in step 3004 occurred at a location that is within or outwith an exclusion zone, e.g. by virtue of a determined proximity, or lack thereof, to a number of prior locations of false fall detections, etc.

[0264] An illustration of a clustering approach is shown in Figure 15, in which each point (e.g. first or second locations) is associated with a threshold distance 8 for a point to be considered a near neighbour. In the example of Figure 15, the threshold distance 8 is shown as a circle around each point. In examples, 8 could beneficially be a value in a range from 40cm to 80cm, but is not limited to this and could take other values depending on a required application and tolerance. Another parameter of the clustering algorithm might be a minimum number of near neighbour points to be considered a cluster. In this example, the minimum number of near neighbour points is a value in a range from 2 to 4, e.g. three points, i.e. one second location forming a cluster with two or more near neighbour first locations. In an example, for there to be a cluster, there may need to exist a core point which has the minimum number of points MinPts, minus one, within the distance threshold 8 of it. The cluster may be extended to include one or more further points if the further point is within the distance threshold 8 to another point of the cluster.

[0265] In the example of Figure 15, four second locations 1500, 1501, 1502, 1503 of a person are shown, and six first locations 1504 of false fall detections are shown. For the sake of this example only one of the second locations 1500, 1501, 1502 and 1503 is considered to exist at a time, and MinPts has a value of 4. It can be seen that second locations 1501 and 1502 are part of clusters with near neighbour first locations 1504. First location 1502 is a core point in a cluster that includes first locations 1504a-d. In this cluster point 1504d is also a core point of that cluster because points 1502, 1504e and 1504f are within the threshold distance 8 of it. Second location 1501 on the other hand would form a cluster, as a core point, with points 1504d-f which are within the distance a of it. This cluster does not include points 1504a-c because none of them are within the threshold a to one of the other cluster points 15Od-f (nor to the second location 1501). However, second location 1503 does not form a cluster with any first locations as only one first location is within the distance a, such that the number of points in this cluster is less than the threshold number required to form a cluster, and the point 1504c to which the second location 1503 is within the distance a to, is not itself a part of a cluster of first locations. Similarly, second location 1500 does not form a cluster with any first locations as only one first location is within the distance a, such that the number of points in this cluster is less than the threshold number required to form a cluster, and the point 1504f to which the second location 1503 is within the distance a to, is not itself a part of a cluster of first locations. Therefore, second location 1500 does not form a cluster with first locations. As such, a fall detection action may be raised in relation to a fall detected at location 1503 and 1500, whereas falls at locations 1501 and 1502 may be considered likely to be false as they form part of an excluded zone by virtue of forming a cluster with locations of known false fall detections.

[0266] Were MinPts set to 3 for example, as is the case in another example, the point 1504f would form a cluster with points 1504d and 1504e, with each being a core point. Second location 1500 would then be a part of that cluster because it is within the threshold distance a to one of the cluster points 1504f, and would therefore be considered part of excluded zone. As a result a fall detection action would not be raised in relation to a fall detected at location 1500.

[0267] In another clustering example, to be a member of the cluster a given point must be a core point of a cluster (such as defined above) or be within the threshold distance a of a core point of a cluster. Similarly in this case the second location 1500 would then be a part of a cluster, since point 1504f is a core point.

[0268] In an exemplary embodiment the clustering algorithm is a DBscan algorithm based on DBscan parameters MinPts and a, with MinPts equal to at least 3 (equal to 3 in an embdiment) and / or with a in the range of 40 to 80cm (e.g. 60cm).

[0269] The detected fall action comprises generating a fall detection alert, such as described herein.

[0270] In the example shown, the monitoring device 200 awaits a response to the alert to verify whether a fall has in fact taken place. In Figure 14 the response is stated to be received from a call center. For example it may be determined at the call center, based on a conversation with the person at site of the detected fall (e.g. with the person though to have fallen or a carer of the person), whether the fall detection was genuine or a false detection. The monitoring device 200 may receive that response at step 3022. In embodiments where an alert is locally generated, the response to the alert may be received by any other means, e.g. via microphone or button on the monitoring device 200 or via a signal received from a local device.

[0271] The response from the verification is analysed in step 3024. If the response from the verification defines the fall as a false alarm or a false detection indicating that that the person did not fall, then the location of the event detected in step 3004 is stored for future use in the clustering steps of 3008 / 3010. If the response from the verification defines the fall as genuine, then the location of the event detected in step 3004 is not stored for future use in the clustering steps of 3008 / 3010.

[0272] The approaches outlines in Figures 13A, 13B and 14 ae beneficial in that they allow exclusion zones 606 and thereby regions of interest 600 to be created and modified, optionally dynamically in use, and may be independent of the mechanism causing the false fall events, i.e. it may be agnostic to any reason for the false fall events.

[0273] Though determining that a person is in a region of interest that is outside an exclusion zone may involve determining a boundary of a region of interest in such a manner that it excludes a region of interest, determining that a person is in a region of interest that is outside an exclusion zone may comprise or consist of identifying a location of a person and determining that that location is outside an exclusion zone.

[0274] Although the subject matter has been described in language specific to structural features and / or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.

Claims

WHAT IS CLAIMED IS:

1. A computer implemented method of identifying falls in an environment using a ranging active reflective wave detector, the method comprising: identifying at least one first location of a respective fall detection event, determined from ranging active reflected wave detector measurements, that is determined to be a false fall detection; subsequent to identifying the at least one first location, using the ranging active reflective wave detector to monitor for falls and generating a fall detection action in response to a fall detected using the ranging active reflective wave detector; and determining whether a second location derived from ranging active reflected wave detector measurements is in a region of interest that is outside an exclusion zone, the exclusion zone being associated with the at least one first location; wherein at least one of the generating of a fall detection action based on a detected fall and / or detecting a fall using the ranging active reflective wave detector is conditional upon the second location being in the region of interest.

2. The method of claim 1, wherein determining whether the fall detection event is a false fall detection comprises accruing a plurality of first events over a period of time and inferring, after the period of time, one or more of the at least one first locations from respective event locations of the plurality of first events.

3. The method of claim 2 wherein first events during the period of time are treated as false fall detection events.

4. The method of claim 2 or claim 3, wherein the period of time is a set-up or learning period during which fall detection alerts in response to the respective first events are at least one of: not generated or generated but not treated as an alert, or not as real alert, by a device receiving the alert.

5. The method of any of claims 2 to 4, wherein the plurality of first events are detected prior to the monitoring of the at least one region of interest to detect a fall.

6. The method of any of claims 2 to 5, wherein the period of time is a period in which there is an assumption that no falls occurred during the period of time.

7. The method of any of claims 2 to 6, comprising automatically inferring the exclusion zone based on the plurality of first events being detected prior to the monitoring of the at least one region of interest to detect a fall.

8. The method of claim 7, wherein the inferring of the exclusion zone based on the plurality of first events comprises basing the exclusion zone on a cluster of the first locations associated with the plurality of first events and / or inferring the exclusion zone from the plurality of first events using one or more learning algorithms.

9. The method of any of the claims 1-8, wherein a set of first locations of a respective fall detection event determined to be a false fall detection that are associated with an exclusion zone or form a cluster is limited to a maximum size.

10. The method of claim 9, wherein the set of first locations of a respective fall detection event determined to be a false fall detection that are associated with an exclusion zone or form a cluster is limited to a set or preset number of most recently identified first locations.

11. The method of any of the claims 1-10, in which event locations are disqualified from being a first location if more than a predetermined amount of time has elapsed since the event location was identified.

12. The method of any of the claims 1-11, wherein determining whether the fall detection event is a false fall detection is not based on a response for a person to which the fall detection event relates.

13. The method of any of the claims 1-12, wherein the ranging active reflected wave detector measurements used to determine the fall detection event from the at least one first location are from said active reflected wave detector.

14. The method of any of the claims 1-13, wherein the identifying at least one first location comprises: identifying at least one first event as a respective fall detection event from measurements from the ranging active reflective wave detector;determining, from measurements from the ranging active reflective wave detector, at least one event location, the at least one event location respectively corresponding to the at least one first event, determining whether the at least one first event is a false fall detection; and in an event that the at least one first event is determined to be a false fall detection, identifying the at least one event location as being said at least one first location.

15. The method of any of the claims 1-14, wherein the determining that the second location is in a region of interest that is outside an exclusion zone comprises determining that the second location is not proximate to the at least one first location.

16. The method of claim 15, wherein the determining that the second location is not proximate to the at least one first location comprises determining that the second location is not part of a cluster of locations comprising or consisting of the second location and at least a set or preset minimum number of first locations.

17. The method of claim 16, wherein the set or preset minimum number of first locations is two or more first locations.

18. The method of any of the claims 1-17, comprising identifying the second location, wherein the second location represents a location of a person and the monitoring for falls comprises performing fall detection process based on the ranging active reflected wave detector measurements conditional at least upon a person being determined to be in the region of interest that is outside the exclusion zone.

19. The method of any of the claims 1-18 comprising implementing a plurality of exclusion zones, each exclusion zone being associated with at least one different first location.

20. A processing system for identifying falls in an environment using output of a ranging active reflective wave detector, the processing system comprising a processor configured to: identify at least one first location where a respective fall detection event, determined from ranging active reflected wave detector measurements, is determined to be a false fall detection;subsequent to identifying the at least one first location, use the ranging active reflective wave detector to monitor for falls and generate a fall detection action in response to a fall detected using the ranging active reflective wave detector; and determine whether a second location derived from ranging active reflected wave detector measurements is in a region of interest that is outside an exclusion zone, the exclusion zone being associated with the at least one first location; wherein at least one of the generating of a fall detection action based on the detected fall and / or detecting a fall using the ranging active reflective wave detector is conditional upon the second location being in the region of interest.

21. A system comprising the processing system of claim 20 and a ranging active reflective wave detector, wherein the processing system is configured to identify falls in an environment using output of the ranging active reflective wave detector.

22. The system of claim 21, wherein the processing system is integrated into a single common device with the ranging active reflective wave detector.

23. A computer implemented method of operating a device comprising a ranging active reflective wave detector to detect a fall in an environment in which the device is installed, the method comprising: receiving a signal from a sensor other than the ranging active reflective wave detector, wherein the receiving of the signal from the sensor is indicative that the person is located on an object in the environment; using the ranging active reflective wave detector to monitor for falls; generating a fall detection alert in response to a fall detected using the ranging active reflective wave detector; wherein at least one of the generating of a fall detection action based on a fall detected using the ranging active reflective wave detector and / or detecting a fall using the ranging active reflective wave detector is conditional upon the sensor not indicating that a person is located on the object.

24. The method of claim 23, wherein the ranging active reflective wave detector comprises one of: radar, lidar, or sonar and the sensor is a different type of device to the ranging active reflective wave detector.

25. The method of any of claims 23 or 24, wherein a total volume or total area of the at least one region of interest is the same or less than a volume or area of a region that is observable by the ranging active reflective wave detector.

26. A processing system for identifying falls in an environment using output of a ranging active reflected wave detector, the processing system comprising a processor configured to: receive a signal from a sensor other than the ranging active reflective wave detector, wherein the receiving of the signal from the sensor is indicative that the person is located on an object in the environment; use the ranging active reflective wave detector to monitor for falls; generate a fall detection action in response to a fall detected using the ranging active reflective wave detector; wherein at least one of the generating of a fall detection action based on a fall detected using the ranging active reflective wave detector and / or detecting a fall using the ranging active reflective wave detector is conditional upon the sensor not indicating that a person is located on the object.

27. A system comprising the processing system of claim 26 and a ranging active reflective wave detector, wherein the processing system is configured to identify falls in an environment using output of the ranging active reflective wave detector.

28. A computer-readable storage medium comprising instructions which, when executed by a processor of a device to cause the processor to perform the methods of any of claim 1 to 19 or claims 23 to 25.

Images