System and method for determining occupant configuration in a space using sensors having single pixel thermopiles - Patents.com

Abstract

A monitoring system and method are disclosed. The system includes a sensor device having a single pixel thermopile (SPT) and a controller in communication with the sensor device. The controller is configured to determine or obtain training data for detecting a first configuration and a second configuration of one or more occupants in a space, receive a temperature signal from the SPT sensor, detect a step change in the temperature signal, determine a first probability and a second probability for at least one feature of a shape or pattern of the temperature signal, the first probability and the second probability corresponding to the first configuration and the second configuration of the one or more occupants in the space based on the training data, respectively, and identify a class label of the temperature signal based on the determined first and second probabilities. The class label corresponds to the first or second configuration of the one or more occupants in the space.

Description

[Technical field]

[0001] The present disclosure generally relates to systems and methods for determining occupant configurations in a space using sensors having single pixel thermopiles. [Background technology]

[0002] When an individual is admitted for a medical test or treatment, they are typically not allowed to walk to and / or from the test or treatment by themselves. Instead, the individual or patient is escorted by one or more authorized medical personnel in a wheelchair, stretcher, gurney, etc. Proper patient transport requires knowledge, skill, equipment, and communication, the absence of any of which can result in minor injuries, such as a cut on a finger caught in an unprotected rail, or catastrophic injuries, such as traumatic brain injury or death from a fall. Considering the number of hospitalized patients in the United States, if each patient is transported to a room, transported to and from the test, and transported from the room to the exit, there are at least 140 million opportunities for injury during patient transport. Furthermore, it is estimated that hospital accidents during patient transport are more likely to go unreported. Most healthcare providers invest significant resources in training, equipment upgrades, and capital investments to prevent these accidents from occurring.

[0003] Wheelchair assistance is also crucial in airports and other transportation hubs. For example, the number of people needing wheelchair assistance at airports is growing at a rate that is much higher than the growth in annual air passenger numbers. People who require wheelchair assistance are sometimes called Passengers with Reduced Mobility (PRM). The safe transport of PRM is crucial for an efficient and safe transportation system.

[0004] Additionally, airports, like medical facilities, have problems tracking their wheelchair fleets. These problems result in significant losses each year. Conventional systems and methods for tracking wheelchairs involve augmenting the wheelchairs or using additional equipment such as visible light communication (VLC) sensors or Bluetooth low-energy localization sensors. Unfortunately, such sensors can be expensive and / or vulnerable to privacy attacks. Summary of the Invention [Problem to be solved by the invention]

[0005] Thus, there is a need in the art for improved systems and methods for monitoring and tracking the transport of patients and PRMs through medical facilities and transportation hubs while respecting privacy concerns. [Means for solving the problem]

[0006] The present disclosure generally relates to an inventive system and method for monitoring patient and reduced mobility people (PRM) transport using a sensor device with a single pixel thermopile (SPT). In general, embodiments of the present disclosure relate to an improved system and method for determining the configuration of one or more occupants in a space using a connected lighting system with SPT sensors embedded in lighting fixtures and, optionally, at doorways. The inventive system and method includes detecting step changes in the SPT signal, classifying the detected step changes into two or more classes, calculating a confidence score for each classification, and analyzing events detected by the SPT sensor throughout the space or building to supplement the classification and / or update the wheelchair location. If an unacceptable occupant configuration is detected, the inventive system and method can provide notification of such configuration. Applicant recognizes and understands that features of the shape or pattern of the SPT signal are strongly correlated with the number of people and their configuration. Such features can be used to monitor the transport of patients and PRMs (including wheelchair-bound PRMs) without requiring augmentation for tracking and while respecting privacy concerns. The SPT sensors of the system and method also have the advantage of being cost-free.

[0007] In general, in one aspect, a system is provided for determining a configuration of one or more occupants in a space, the system including at least one sensor device having a single pixel thermopile in the space, the single pixel thermopile configured to capture a sensor signal related to the one or more occupants in the space, and a controller in communication with the at least one sensor device in the space. The controller of the system is configured to: determine or obtain training data for detecting a first configuration and a second configuration of one or more occupants in the space; receive a temperature signal from the single-pixel thermopile, the temperature signal corresponding to a detection area in the space within a field of view of the single-pixel thermopile over time; detect a change in the temperature signal, the change being equal to or greater than a predetermined step threshold; determine a first probability and a second probability for at least one feature of a shape or pattern of the temperature signal, the first probability and the second probability corresponding to the first configuration and the second configuration of the one or more occupants in the space based on the training data, respectively; and identify a class label of the temperature signal based on the determined first and second probabilities, the class label corresponding to the first configuration or the second configuration of the one or more occupants in the space.

[0008] In one embodiment, if the class label corresponds to the first configuration, the controller is further configured to determine whether a first probability for at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration exceeds a predetermined probability threshold.

[0009] In one embodiment, if the first probability for at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration exceeds a predetermined probability threshold, the controller is further configured to provide a notification including information of the first configuration of one or more occupants in the space.

[0010] In one embodiment, if the first probability for at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration does not exceed a predetermined probability threshold, the controller is further configured to send a signal including information about the at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration to another sensor device in the space or to another controller.

[0011] In one embodiment, if the first probability for at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration does not exceed a predetermined probability threshold, the controller is further configured to receive a signal from another sensor device in the space or another controller including information regarding the shape or pattern of another temperature signal corresponding to the first configuration.

[0012] In one embodiment, the controller is further configured to fuse a signal including information regarding at least one characteristic of a shape or pattern of a temperature signal corresponding to the first configuration with a signal including information regarding a shape or pattern of another temperature signal corresponding to the first configuration from another sensor device or another controller in the space for a subsequent determination regarding a first configuration of one or more occupants in the space.

[0013] In one embodiment, at least one sensor device including a single pixel thermopile is integrated into a lighting fixture, and the lighting fixture is part of a plurality of connected lighting devices in a space.

[0014] In one embodiment, at least one sensor device including a single pixel thermopile is provided at a doorway of the space.

[0015] In general, in another aspect, a method is provided for determining a configuration of one or more occupants in a space, the method including: (a) providing at least one sensor device in the space, the sensor device including a single pixel thermopile configured to capture sensor signals related to one or more occupants in the space; (b) providing a controller in communication with the at least one sensor device in the space; (c) determining or acquiring, by the controller, training data for detecting a first configuration and a second configuration of the one or more occupants in the space; (d) measuring, by the single pixel thermopile, a temperature signal of a detection area in the space within a field of view of the single pixel thermopile over time; (e) detecting, by the controller, a change in the temperature signal, the change being equal to or greater than a predetermined step threshold; (f) determining, by the controller, a first probability and a second probability for at least one characteristic of a shape or pattern of the temperature signal, the first probability and the second probability corresponding to the first configuration and the second configuration, respectively, of the one or more occupants in the space based on the training data; and (g) determining, by the controller, a first probability and a second probability for at least one characteristic of a shape or pattern of the temperature signal, the first probability and the second probability corresponding to the first configuration and the second configuration, respectively, of the one or more occupants in the space based on the training data. and identifying, at the controller, a class label of the temperature signal based on the determined first and second probabilities, the class label corresponding to a first configuration or a second configuration of one or more occupants in the space.

[0016] In one embodiment, if the class label corresponds to a first configuration of one or more occupants in the space, the method further includes determining, at the controller, whether a first probability for at least one characteristic of a shape or pattern of the temperature signal corresponding to the first configuration exceeds a predetermined probability threshold.

[0017] In one embodiment, if the first probability for at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration exceeds a predetermined probability threshold, the method further includes providing, at the controller, a notification including information of the first configuration of one or more occupants in the space.

[0018] In one embodiment, if the first probability for at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration does not exceed a predetermined probability threshold, the method further includes transmitting, at the controller, a signal including information regarding the at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration to another sensor device in the space or to another controller.

[0019] In one embodiment, the method further includes fusing the signal including information regarding at least one characteristic of a shape or pattern of the temperature signal corresponding to the first configuration with at least one other signal from another sensor device for a subsequent determination regarding a first configuration of one or more occupants in the space.

[0020] In one embodiment, if the class label corresponds to a second configuration of one or more occupants in the space, the method further includes repeating steps (d), (e), (f), and (g) with the controller.

[0021] In one embodiment, at least one sensor device including a single pixel thermopile is integrated into a lighting fixture, and the lighting fixture is part of a plurality of connected lighting devices in a space.

[0022] It should be understood that all combinations of the above-mentioned concepts, and additional concepts discussed in more detail below (provided such concepts are not mutually inconsistent), are contemplated as part of the inventive subject matter disclosed herein, and in particular, all combinations of claimed subject matter described at the end of this disclosure are contemplated as part of the inventive subject matter disclosed herein. [Brief description of the drawings]

[0023] In the drawings, like reference characters generally refer to the same parts throughout the different views. Also, the drawings are not necessarily to scale, emphasis instead generally being placed upon illustrating the principles of the disclosure. [Figure 1]1 illustrates a schematic diagram of a space including at least one sensor device having an embedded single pixel thermopile sensor, according to an embodiment of the present disclosure. [Figure 2A] FIG. 1 shows a schematic diagram of a connected lighting system including at least one sensor device having a single pixel thermopile, a controller, and one or more lighting devices, according to an embodiment of the present disclosure. [Figure 2B] 1 shows a schematic diagram of a connected lighting system including at least one lighting device having a controller and a sensor device including a single pixel thermopile, according to an embodiment of the present disclosure. [Diagram 3] 1 illustrates an example schematic temperature signal from a single pixel thermopile, according to an aspect of the present disclosure. [Figure 4] 1 illustrates one embodiment of a system and method for monitoring the entry and exit of one or more wheelchairs in a healthcare facility, according to aspects of the present disclosure. [Diagram 5] 1 illustrates an exemplary method for monitoring the entry and exit of one or more wheelchairs in a space, according to an aspect of the present disclosure. [Figure 6] 1 illustrates an exemplary method for determining a configuration of one or more occupants in a space, according to an aspect of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0024] This disclosure describes various embodiments of improved systems and methods for determining the configuration of one or more occupants in a space using a connected lighting system having an embedded sensor device including a single pixel thermopile (SPT). Applicant recognizes and understands that features of the shape or pattern of an SPT signal are strongly correlated with the number of people and their configuration. Furthermore, Applicant recognizes and understands that it is beneficial to monitor the transport of patients and reduced mobility persons (PRM) based on features of the shape or pattern of an SPT signal as described herein. The systems and methods can be trained with data corresponding to two or more configurations of one or more occupants in a space. After setup and training, the systems and methods can detect step events in one or more measured SPT signals, calculate statistical features for the shape or pattern of one or more measured SPT signals, calculate posterior probabilities based on the calculated statistical features and training data, and classify one or more SPT signals as corresponding to at least one of two or more configurations of one or more occupants in a space. In an exemplary embodiment, Bayesian classification is used to enable the system. In addition to monitoring the transport of patients in medical facilities and PRMs in transportation hubs such as airports, the systems and methods described herein can be used to detect tailgating in offices and enable any suitable safety-related use case.

[0025] This disclosure describes various embodiments of systems and methods for providing a distributed network of single pixel thermopile sensors by utilizing illumination devices that may already be deployed in a multi-grid and connected architecture (e.g., connected lighting infrastructure). Such existing infrastructure can be used as a backbone for the additional detection and notification functions described herein. Signify's SlimBlend® pendant luminaire is an example of a suitable illumination device with built-in IoT sensors, such as microphones, cameras, and thermopile infrared sensors, as described herein. In an embodiment, the illumination device includes a USB-type connector slot for receivers, sensors, and the like. Illumination devices that include a sensor ready interface are particularly suitable, already providing power, Digital Addressable Lighting Interface (DALI) connectivity to the functionality of the luminaire, and standardized slot geometry. It should be understood that any illumination device that is connected or connectable and sensor-enabled is contemplated, including recessed or surface mounted luminaires, pendant luminaires, wall mounted luminaires, floor free standing luminaires, etc. A pendant or floor free standing luminaire that includes a thermopile infrared sensor may be advantageous as the sensor is placed closer to the occupants in the space and can detect higher temperatures of people. Furthermore, the resolution of the thermopile sensor may be lower than that of a thermopile sensor mounted in a recessed or surface mounted luminaire that is mounted at a ceiling height of about 3m.

[0026] As used herein, the term "luminaire" refers to a device that includes one or more light sources of the same or different types. A given luminaire may have any of a variety of mounting configurations for the light source(s), a variety of configurations and shapes of enclosures / housings, and / or a variety of configurations of electrical and mechanical connections. Additionally, a given luminaire may optionally be associated with (e.g., include, be coupled to, and / or be packaged together with) a variety of other components (e.g., control circuitry) related to the operation of the light source(s). It should also be understood that light sources may be configured for a variety of applications, including, but not limited to, indicating, displaying, and / or illumination.

[0027] Referring to FIG. 1, a schematic diagram of a space 10 including a sensor device SD according to an aspect of the disclosure is shown. The space 10 can be a medical facility such as a hospital or an independent living facility, an airport, an office space, etc. The sensor device SD includes a single pixel thermopile SPT and a controller 102 as shown in FIG. 2A. FIG. 2A shows a schematic diagram of a connected lighting system 100 according to an aspect of the disclosure. The connected lighting system 100 includes at least one sensor device SD having a single pixel thermopile SPT, a controller 102, and one or more lighting devices 104A, 104B, etc. The single pixel thermopile SPT and the controller 102 are housed and operably coupled within the same housing of the sensor device SD as shown in FIG. 2A. In an alternative embodiment, the single pixel thermopile SPT and the controller 102 are housed separately. In the system 100 shown in FIG. 2A, one or more lighting devices 104A and 104B are operatively coupled to a sensor device SD.

[0028] In an embodiment, the sensor device SD is embedded in at least one lighting device 204A as shown in the connected lighting system 200 of FIG. 2B. In other words, in an embodiment, the lighting device 204A may include a sensor device SD with a single pixel thermopile SPT. In an exemplary embodiment, the sensor device SD is part of a lighting fixture or part of a sensor-bundle embodied in or connectable to a lighting fixture. The at least one lighting device 204A is configured to illuminate the space 20 including the detection area 106. The at least one lighting device 204A can also be configured to provide a notification as described herein as a result of the controller 202 determining a configuration of one or more occupants in the detection area 106 in the space 20.

[0029] Each lighting device or luminaire includes one or more light sources, which may include light emitting diodes (LEDs) arranged on a printed circuit board. The LEDs can be configured to be driven by one or more light source drivers to emit light of specific characteristics (i.e., color intensity and color temperature). The LEDs may be active (i.e., on), inactive (i.e., off), or dimmed by a factor d (0≦d≦1). A value d=0 means that the LED is turned off, and d=1 represents that the LED is at its maximum illumination.

[0030] The controller 102 includes a network interface 120, a memory and / or storage device 122, and one or more processors 124. The network interface 120 can be embodied as a wireless transceiver or any other device that allows the connected lighting fixtures to wirelessly communicate with other devices, including mobile devices, and each other in the connected lighting system 100 utilizing the same wireless protocol standard and / or monitor network activity, and allows the controller 102 to receive data from the SPT sensor SPT. In an embodiment, the network interface 120 may use a wired communication link. The SPT sensor is configured to transmit data to the one or more processors 124 of the controller 102 via any suitable wired / wireless network communication channel. In an embodiment, the data can be transmitted directly to the one or more processors 124 of the controller 102 without going through a network. The data can be stored in the memory 122 of the controller 102 via the wired / wireless communication channel. The memory 122 and one or more processors 124 of the controller 102 may take any suitable form in the art for controlling, monitoring, and / or assisting in the operation of the lighting devices 104A, 104B, 204A, the SPT sensors, and performing other functions of the controller 102 described herein. The one or more processors 124 of the controller 102 may also execute instructions stored in the memory 122 or process data, for example, to perform one or more steps of the methods described herein. The one or more processors 124 of the controller 102 may include one or more modules, such as one or more modules for capturing data and detecting step events in the captured data, one or more modules for classifying step events, one or more modules for analyzing step event classifications based on associated confidence scores, and one or more training modules, as described herein.Although the following description details an embodiment employing Bayesian classification techniques, the disclosure should not be so limited. Any suitable form of classification is envisioned. For example, logistic regression or any other suitable alternative could be used.

[0031] In the embodiment shown in FIG. 1, the sensor device SD is mounted in a doorway facing down towards the floor of the area 10. In an alternative embodiment, the sensor device SD is mounted in the doorway at an angle towards the floor of the area 10. The sensor device can also be mounted on the ceiling and facing straight down or diagonally towards the floor of the area 10. The single pixel thermopile SPT converts thermal energy into electrical energy in a detection area 106 of the space 10. The detection area 106 is the area to be monitored by the single pixel thermopile SPT. The detection area 106 is embodied as a volumetric cone with a centerline CL as shown in FIG. 2A and FIG. 2B. The centerline CL emanates from the center of the single pixel thermopile SPT in FIG. 2A and FIG. 2B. The volumetric cone may be considered as the field of view of the single pixel thermopile SPT. Thus, the detection area 106 corresponds to the single-pixel resolution of the single-pixel thermopile SPT. In FIG. 1, due to the mounting of the sensor device SD in the space 10, only about half of the volumetric cone of the detection area 106 is visible. In FIGS. 2A and 2B, the detection area 106 represents a 90 degree field of view, but it should be understood that the field of view can be increased or decreased depending on the application or type of single-pixel thermopile SPT. The detection area 106 can also be considered as a surface area, e.g., a two-dimensional footprint of the volumetric cone on the floor of the space.

[0032] The conversion of thermal energy into electrical energy by the single pixel thermopile SPT generates an SPT sensor signal, which may also be referred to as a temperature signal, heat signal, or enthalpy signal. The temperature signal may also be considered an object infrared (IR) measurement signal. The single pixel thermopile SPT generates a single temperature value due to the single pixel resolution. In FIG. 3, an exemplary temperature signal 300 is shown over a period of time on the horizontal axis (X). The exemplary signal 300 shows drastic changes over time as one or more persons P enter or leave the detection area 106 or field of view. Thus, when the temperature signal 300 is measured over a period of time, the temperature signal 300 includes the transient temperature response (or thermal response) of the complete detection area 106.

[0033] As shown in FIG. 3, signal 300 includes portions A, B, and C in some embodiments. In some embodiments, other signals may include additional portions. Portion A of signal 300 represents the temperature of the detection area 106 without the presence of persons P at the beginning of a period of time. After portion A, one or more persons P enter the detection area 106, causing a step change B in the IR signal. A step change refers to an abrupt change, step, jump, or shift in the average temperature level of the signal over time. After step change B, as one or more persons P continue to move further within the detection area 106 over the period of time, the IR signal gradually decreases during portion C.

[0034] Applicant recognizes and understands that the magnitude of the step change B along with other statistical characteristics such as rise time and overshoot are strongly correlated with the number of people detected within the detection area 106 and / or their configuration within the detection area 106. In FIG. 3, the size of the step change B is shown at 302. The term "size" with respect to a step change refers to the amount of temperature change between a particular low temperature value or range of low temperature values ​​and a particular high temperature value or range of high temperature values. The change in the temperature signal at 302 must be equal to or greater than a predefined step threshold to avoid detecting a minimal variation in the temperature signal that is not indicative of a step event such as at least one person entering / exiting the detection area.

[0035] The rise time of step change B in FIG. 3 is shown at 304. The term "rise time" refers to the amount of time it takes for a signal to change from a particular low temperature value or range of low temperature values ​​to a particular high temperature value or range of high temperature values. The overshoot of step change B in FIG. 3 is shown at 306. The term "overshoot" refers to a transient temperature value that exceeds its steady state or final value during a transition from a low temperature value or range of low temperature values ​​to a high temperature value or range of high temperature values. Step size, rise time, and overshoot are example signal features that may be strongly correlated with the number of people detected within the detection area 106 and / or their configuration within the detection area 106.

[0036] For example, the step size for two people moving into or out of the SPT's field of view is larger than the step size for one person moving into or out of the SPT's field of view, such as size 302 in Figure 3. A positive step size corresponds to a person moving into the field of view. A negative step size corresponds to a person moving out of the field of view.

[0037] The step size can also be analyzed as a function of distance from the SPT. Generally, the step size when a person is walking into the field of view is larger than the step size when a person is sitting into the field of view. The step size can be analyzed within 1 m of the SPT, but any suitable distance can be used depending on the application. The step size when a person is facing the SPT is generally larger than the step size when the person is facing away from the SPT. Thus, a person standing facing the sensor generally has the largest step size. A person sitting facing the sensor generally has the second largest step size. A person standing with their back to the sensor generally has the next largest step size. A person sitting with their back to the sensor generally has the smallest step size compared to the three other configurations. The step size is larger when a person is facing the SPT because more heat is emitted from the face than from the back of the head.

[0038] The statistical characteristics of the two-person field of view are generally characterized as a superposition of the individual effects of the people. Thus, when two people stand in the field of view, the step size is approximately twice the size of the step detected when one person stands in the field of view. The same is true for the three-person field of view. Thus, when three people stand in the field of view, the step size is approximately three times the size of the step detected when one person stands in the field of view.

[0039] As described further herein, the single pixel thermopile SPT and / or the controller 102 can be trained to detect the configuration of occupants in a space based on the shape or pattern of the detected temperature signal. As described herein, the training can be performed with a supervised learning approach, which requires a significant amount of labeled data. Alternatively, the training can be performed with a representative learning approach, which requires a large amount of unlabelled data after the system is installed in the space. In an embodiment, the controller 102 acquires the temperature signal transmitted from the single pixel thermopile SPT. As shown in FIG. 4, an embodiment of a system and method for monitoring wheelchair entry and exit in a facility is shown. The system and method 400 of FIG. 4 includes a system and method 402 for detecting a step change in a temperature signal, a system and method 404 for classifying the detected step change, and a system and method 406 for analyzing the classified detected step change. Advantageously, the systems and methods described herein can track wheelchair entry and exit at a facility without the need to augment the wheelchair with a tracking device.

[0040] The system and method 402 for detecting a step change in a temperature signal includes comparing the temperature change in the detected signal to a predefined step threshold, and if the temperature change is equal to or greater than the predefined step threshold, a step event can be determined to have occurred.

[0041] The system and method 404 for classifying the detected step change includes determining whether one or more signal features of the detected step change are sufficiently similar to signal features associated with one or more classes. The classes can include classes that identify an abnormal or unacceptable number and / or configuration of occupants in the space and one or more classes that identify a normal or acceptable number and / or configuration of occupants in the space. In one embodiment, an abnormal or unacceptable scenario or class is one that identifies only one person sitting in a wheelchair without anyone accompanying them. In this case, the system and method can detect that the person in the wheelchair is unaccompanied according to the facility's policies and procedures. Normal or acceptable scenarios or classes can include those that identify one person walking with no one sitting, those that identify one person walking with another person sitting, those that identify two or more people walking with no one sitting, and other scenarios or classes that are different from the others. One or more of the normal or acceptable scenarios or classes can be combined in any combination. Additional normal or acceptable scenarios or classes not listed above are also envisioned.

[0042] One particular advantage of the systems and methods described herein is that wheelchairs detected in a classified configuration can be tracked among multiple SPT sensors in a space without requiring additional tracking enhancements. Another advantage of the systems and methods described herein is that one or more unique SPT signals can be learned for individuals that are routinely present in a space.

[0043] The abnormal or unacceptable scenario or class may be CLASS A as shown in FIG. 4, where one or more features of the detected temperature signal indicate that only one person, sitting in a wheelchair, is entering or exiting the detection area 106.

[0044] The normal or acceptable scenarios or classes listed above are shown in FIG. 4 as CLASS B, CLASS C, CLASS D, and CLASS E. If one or more features of the detected temperature signal indicate a person walking with no person sitting, the signal can be classified as CLASS B. Since no person is sitting, it is unlikely that a person in a wheelchair is entering or leaving the detection area 106 unaccompanied, as indicated by a signal associated with CLASS A. If one or more features of the detected temperature signal indicate a person walking with another person sitting, the signal can be classified as CLASS C. This configuration represents a situation where a person in a wheelchair is entering or leaving the detection area with an escort, as desired in some embodiments. If one or more features of the detected temperature signal indicate two or more people walking with no person sitting, the signal can be classified as CLASS D. Since there is no seated person (similar to CLASS B above), it is unlikely that a person in a wheelchair is entering or exiting the detection area 106 unaccompanied. If one or more features of the detected temperature signal indicate a configuration of occupants in the space different from those associated with CLASS A, B, C, and D, the signal may be classified as CLASS E. Again, since CLASS E does not indicate the presence of at least one person in a wheelchair in the detection area 106 unaccompanied (as shown in CLASS A above), the configuration may still be considered normal or acceptable.

[0045] The systems and methods 406 for analyzing the classified detected step changes 406 include combining information from multiple SPT sensors across the space 10 to determine if the classification is sufficiently reliable or trustworthy and / or making subsequent classifications based on one or more individual weak decisions. The systems and methods 406 for analyzing the classified detected step changes may also include updating the location of the wheelchair depending on the detected configuration.

[0046] In an embodiment using a supervised learning approach and a large amount of labeled data, statistical features for each of the five classes (CLASSE A, B, C, D, and E) are characterized by the labeled data, and the system is trained by estimating likelihood densities based on the labeled data. These likelihood densities can be learned from data related to step sizes and how they correlate to the number and / or configuration of occupants. For example, likelihood densities for step sizes for CLASS A, B, C, D, and E can be of the form P(step_size|A), P(step_size|B), P(step_size|C), P(step_size|D), and P(step_size|E). Similar densities can be learned for other statistical features, such as rise time and overshoot, from the labeled data. With sufficient labeled data for CLASS A, B, C, D, and E, a time-series deep learning model can be used to learn features and make classifications for A-E. Classes A to E have different step sizes at the input and output, as well as amplitudes and ripples that occur during transmission.

[0047] After at least one sensor device SD having a single pixel thermopile SPT is installed in the space 10, the controller 102 is installed in the space 10, and the controller 102 is operatively coupled to the at least one sensor device, the controller 102 obtains or determines training data for detecting the configuration of occupants in the space. The training data can be based on the labeled data as described above, or any suitable alternative. The training data can also be based on the configuration of the at least one sensor device in the space 10. For example, the height of the at least one sensor device SD, or the angle at which the at least one sensor device SD is mounted, can play a role in the shape or pattern of the signal features of CLASS A, B, C, D, and E. One suitable alternative for generating training data includes obtaining a large amount of unlabeled data after the system 100 is installed in the space 10. Such an alternative includes applying predictive coding for representation learning to the large amount of unlabeled data to learn SPT features due to the number of people, posture, walking speed, etc., and then fine-tune with a small set of labeled data. The learned SPT features can then be incorporated into a classification framework to detect events A-E.

[0048] When the system 100 is operational to determine the configuration of one or more occupants in the space 10, after training as described herein, the single pixel thermopile SPT measures a temperature signal in the detection area 106 and communicates the temperature signal to the controller 102. The controller 102 detects a step change in the temperature signal by comparing the change in the signal to at least one predetermined step threshold.

[0049] To perform step classification, the controller 102 determines a posterior probability of at least one feature of the shape or pattern of the detected step change that meets at least a predefined step threshold. For an example using five classes (CLASS A, B, C, D, and E), the controller 102 determines the posterior probability of the statistical features of the detected step change (e.g., step size, rise time, and / or overshoot) that are being used depending on the application. An example posterior probability of detected step change based on step size, rise time, and overshoot for CLASS A features can be of the form P(A|step_size, overshoot, rise_time) ∝ P(step_size|A).P(overshoot|A).P(rise_time|A).P(A), where the first three terms on the right side of the proportionality are the likelihoods learned during training, and P(A) is the prior bias of a true event being CLASS A. The prior probability can be adjusted based on different types of areas in the healthcare facility or for different types of healthcare facilities. In one embodiment, the relationship between posterior probability and likelihood can also accommodate for dependencies, e.g., dependencies on step size and overshoot. The posterior probability of a detected step change based on step size, rise time, and overshoot for CLASS B features can be of the form P(B|step_size, overshoot, rise_time) ∝ P(step_size|B).P(overshoot|B).P(rise_time|B).P(B), where the first three terms on the right side of the proportionality are the likelihoods learned during training, and P(B) is the prior bias for true events to have been CLASS B.The posterior probability of a detected step change based on step size, rise time, and overshoot for CLASS C features can be of the form P(C|step_size, overshoot, rise_time) ∝ P(step_size|C).P(overshoot|C).P(rise_time|C).P(C), where the first three terms on the right side of the proportionality are the likelihoods learned during training, and P(C) is the prior bias of a true event being CLASS C. The posterior probability of a detected step change based on step size, rise time, and overshoot for CLASS D features can be of the form P(D|step_size, overshoot, rise_time) ∝ P(step_size|D).P(overshoot|D).P(rise_time|D).P(D), where the first three terms on the right side of the proportionality are the likelihoods learned during training, and P(D) is the prior bias of a true event being CLASS D. The posterior probability of a detected step change based on step size, rise time, and overshoot for CLASS E features can be of the form P(E|step_size, overshoot, rise_time) ∝ P(step_size|E).P(overshoot|E).P(rise_time|E).P(E), where the first three terms on the right side of the proportionality are the likelihoods learned during training, and P(E) is the prior bias for true events to have been CLASS E.

[0050] After the posterior probability is calculated for the detected step change based on the step size, rise time, and overshoot for classes A, B, C, D, and E, the controller 102 identifies a class label for the detected temperature signal. The identified class label corresponds to at least one of the following scenarios: "one person sitting" (CLASS A), "one person walking" (CLASS B), "one person walking and one person sitting" (CLASS C), "more than one person walking" (CLASS D), and "other" (CLASS E). In one embodiment, the class label is identified by the controller 102 by determining a maximum a posteriori probability estimate (MAP).

[0051] In one embodiment, if the class corresponds to "one person walking" (CLASS B), "one person sitting and other standing" (CLASS C), ">1 person walking" (CLASS D), or "other" (CLASS E), no action is taken by the system other than continuing to monitor the detection area 106. On the other hand, if the class corresponds to "one person sitting" (CLASS A), where only one person in a sitting position is entering or leaving the detection area 106, the system can be configured to issue an alarm, but only if the posterior probability exceeds some predefined probability threshold. The predefined probability threshold is used in one embodiment to reduce the occurrence of false positives. The probability threshold can be adjusted in one embodiment using a receiver operating characteristic (ROC) curve or any suitable alternative. In such a curve, the true positive rate can be plotted against the false positive rate at various threshold settings. As shown in FIG. 2A, if the controller 102 determines that the detected temperature signal corresponds to “one person sitting” (CLASS A) as described herein, the controller 102 can provide a notification 110 including information about the CLASS A configuration. The notification can be a visual alert on the sensor device SD, such as a solid or flashing red light. Additionally or alternatively, the visual alert can come from one or more lighting devices 104A, 104B, 204A, etc. connected to the sensor device SD. In an embodiment, the notification can be an audible alert. The audible alert can come from the sensor device SD or a sound system connected to the sensor device SD or one or more lighting devices 104A, 104B, 204A.Additionally or alternatively, the notification may include information sent to a remote device, such as a computing or monitoring device at a nurse's station, or any suitable alternative. The controller 102 may output an output signal including information regarding the Class A configuration, the output signal configured to control one or more light sources on the sensor device SD, or one or more light sources on one or more lighting devices 104A, 104B, 204A. Additionally or alternatively, the output signal from the controller 102 may be configured to control one or more speakers on or in the sensor device SD, or one or more speakers on or in the one or more lighting devices 104A, 104B, 204A.

[0052] If the predetermined probability threshold is not met, the information can be broadcast to other sensors or controllers in the space 10 or building so that a joint decision can be made by fusing the individual weak decisions. As shown in FIG. 2A, if the controller 102 identifies a class label but determines that the posterior probability of the signal does not meet the predetermined probability threshold, the controller 102 can transmit information 112 about the signal to another sensor device or another controller 114 in the space 10. Similarly, the controller 102 can receive information 112 about another signal from another sensor or another controller 114 in the space 10.

[0053] FIG. 5 shows an exemplary method 500 for monitoring the entry and exit of one or more wheelchairs in a space. In step S501, a step event is detected in a temperature signal from an SPT sensor. In step S502, a statistical feature of the step event is calculated. In step S503, a posterior probability is calculated based on the calculated statistical feature and training data, and a maximum a posteriori probability estimation (MAP) is used to classify the signal with a class label. In step S504, depending on the class label, it is determined whether the posterior probability associated with the signal exceeds a predetermined probability threshold. As mentioned above, if the class label is one of the acceptable or normal scenarios, there is no need to compare the posterior probability associated with the signal with the predetermined probability threshold. Instead, the system can continue the monitoring step. On the other hand, if the class label is associated with an abnormal or unacceptable scenario (e.g., an unaccompanied patient), it is determined whether the posterior probability associated with the signal exceeds a predetermined probability threshold. In step S505A, if the posterior probability exceeds a predetermined probability threshold, the patient or PRM's status may be updated based on the configuration associated with the class label, and / or any other appropriate action may be taken, such as notification as described above. On the other hand, in step S505B, if the posterior probability does not exceed the predetermined probability threshold, information from one or more SPT sensors of the connected lighting system may be broadcast for subsequent determination of the configuration. The predetermined probability threshold is set at a level that represents a sufficient confidence level associated with the classification.

[0054] 6 is an exemplary method 600 for determining a configuration of one or more occupants in a space. In step 602, at least one sensor device SD is provided in the space 10, the at least one sensor device having a single pixel thermopile SPT configured to capture a sensor signal related to one or more occupants in the space. As described herein, the single pixel thermopile SPT converts thermal energy into electrical energy. In step 604, a controller 102 is provided in the space, the controller communicating with the at least one sensor device in the space.

[0055] At step 606, the controller 102 determines or obtains training data for detecting at least a first configuration and a second configuration of one or more occupants in the space. In some embodiments, the controller 102 determines or obtains training data for detecting at least one configuration corresponding to an abnormal or unacceptable configuration and at least one configuration corresponding to a normal or acceptable configuration. In some embodiments, the controller 102 determines or obtains training data for detecting two or more normal or acceptable configurations.

[0056] In step 608, the single pixel thermopile SPT measures or acquires a temperature signal of a detection area 106 in space over time. The detection area is within the field of view of the single pixel thermopile. In one embodiment, the field of view is about 90 degrees.

[0057] In step 610, the controller 102 detects a change in the temperature signal (eg, step change B), which is greater than or equal to a predetermined step threshold.

[0058] In step 612, the controller 102 determines a first probability and a second probability for at least one characteristic of the temperature signal shape or pattern, the first probability and the second probability corresponding to a first configuration and a second configuration, respectively, based on the determined or acquired training data. The first probability for the at least one characteristic of the temperature signal shape or pattern corresponds to the first configuration. The second probability for the at least one characteristic of the temperature signal shape or pattern corresponds to the second configuration.

[0059] In step 614, the controller 102 identifies a class label for the temperature signal based on the determined probability. The class label corresponds to a first configuration or a second configuration of one or more occupants in the space.

[0060] In one embodiment, the first configuration represents "one person sitting" (CLASS A) and the second configuration represents "one person walking" (CLASS B), "one person walking and one person sitting" (CLASS C), "two or more people walking" (CLASS D), or "other" (CLASS E). In such an embodiment, the method 600 may further include determining, by the controller 102, whether a posterior probability for at least one feature of the shape or pattern of the temperature signal corresponding to the first configuration exceeds a predetermined probability threshold. Such an embodiment is applicable where the controller 102 identifies a class label corresponding to the first configuration that may represent "one person sitting" (CLASS A). If the posterior probability exceeds the predetermined probability threshold, the method 600 may further include providing a notification including information of the first configuration of one or more occupants in the space, as described herein.

[0061] In another embodiment where the posterior probability does not exceed a predetermined probability threshold, the method 600 further includes transmitting a signal including information regarding at least one characteristic of a shape or pattern of a temperature signal corresponding to the first configuration. The transmitted signal is sent to another sensor device or another controller in the space. If the posterior probability does not exceed the predetermined probability threshold, it can be inferred that the class label had a low or unreliable associated confidence score. Alternatively, the method 600 further includes receiving a signal including information regarding a shape or pattern of another temperature signal corresponding to the first configuration. The received information can be from another sensor device or another controller in the space. The information regarding at least one characteristic of a shape or pattern of a temperature signal corresponding to the first configuration in the controller 102 can then be fused with information from another sensor device or another controller in the space for subsequent determination of the configuration of one or more occupants in the space. Combining data from multiple SPT sensors in the space can generate a signal with a higher associated confidence score that is more reliable. Combining data from multiple SPT sensors can also compensate for noise in the signal. Additionally, additional signal processing can be used in any of the systems and methods described herein to identify wheelchair users using unique SPT sensor signals.

[0062] It should also be understood that, unless expressly indicated to the contrary, in any method claimed herein that includes two or more steps or actions, the order of the method steps or actions is not necessarily limited to the order in which the method steps or actions are recited.

[0063] As defined and used herein, all definitions should be understood to govern dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

[0064] The indefinite articles "a" and "an," as used in the specification and claims, unless clearly indicated otherwise, should be understood to mean "at least one."

[0065] The term "and / or," as used in the specification and claims, should be understood to mean "either or both" of the elements so conjoined, i.e., elements that are conjunctive in some cases and disjunctive in other cases. Multiple elements listed with "and / or" should be construed in the same manner, i.e., "one or more" of the elements so conjoined. Other elements than the elements specifically identified by the "and / or" clause may be optionally present, whether related or unrelated to those elements specifically identified.

[0066] As used in the specification and claims, "or" should be understood to have the same meaning as "and / or" as defined above. For example, when separating items in a list, "or" or "and / or" shall be interpreted as inclusive, i.e., including at least one, but also including more than one of an element or list of elements, and optionally additional unlisted items. Only terms such as "only one of" or "exactly one of," or when used in the claims, "consisting of," where the contrary is clearly indicated, refer to the inclusion of exactly one of an element or list of elements. In general, the term "or," as used herein, shall be interpreted as indicating exclusive alternatives (i.e., "one or the other, but not both") only when preceding terms of exclusivity, such as "any of," "one of," "only one of," or "exactly one of."

[0067] As used in this specification and claims, the phrase "at least one" referring to a list of one or more elements should be understood to mean at least one selected from any one or more of the elements in the list of elements, but not necessarily including at least one of each element specifically listed in the list of elements, and not excluding any combinations of elements in the list of elements. This definition also allows for elements other than the elements specifically identified in the list of elements to which the phrase "at least one" refers may optionally be present, whether related or unrelated to those elements specifically identified.

[0068] In the claims and in the above specification, all transitional phrases such as "comprising," "including," "carrying," "having," "containing," "involving," "holding," "consisting of," and the like, are to be understood to be open-ended, i.e., to mean including but not limited to. Only transitional phrases such as "consisting of" and "consisting essentially of" are closed or semi-closed transitional phrases, respectively.

[0069] While several inventive embodiments have been described and illustrated herein, those skilled in the art will readily envision various other means and / or structures for performing the functions described herein and / or obtaining one or more of the results and / or advantages thereof, and such variations and / or modifications are deemed to be within the scope of the inventive embodiments described herein. More generally, those skilled in the art will readily appreciate that all parameters, dimensions, materials, and configurations described herein are intended to be exemplary, and that the actual parameters, dimensions, materials, and / or configurations will vary depending on the particular application in which the teachings of the present invention are used. Those skilled in the art will recognize, or be able to ascertain, using no more than routine experimentation, many equivalents to the specific inventive embodiments described herein. Thus, the above-described embodiments are presented by way of example only, and it will be understood that, within the scope of the appended claims and their equivalents, inventive embodiments may be practiced other than as specifically described and claimed. Inventive embodiments of the present disclosure are directed to each individual feature, system, article, material, kit, and / or method described herein. Furthermore, any combination of two or more such features, systems, articles, materials, kits, and / or methods is within the scope of the present disclosure, provided that such features, systems, articles, materials, kits, and / or methods are not mutually inconsistent.

Claims

1. A system for determining the configuration of one or more occupants in a space, the system comprising: at least one sensor device including a single pixel thermopile in the space, the single pixel thermopile being configured to capture a sensor signal regarding one or more occupants in the space; a controller in communication with the at least one sensor device in the space, the controller being configured to: determine or obtain training data for detecting a first configuration and a second configuration of one or more occupants in the space; receive a temperature signal from the single pixel thermopile, the temperature signal corresponding to a detection area in the space within the field of view of the single pixel thermopile over time; detect a change in the temperature signal, the change being equal to or greater than a predetermined step threshold; determine a first probability and a second probability for at least one feature of the shape or pattern of the temperature signal, the first probability and the second probability corresponding respectively to the first configuration and the second configuration of one or more occupants in the space based on the training data; and identify a class label of the temperature signal using the determined first and second probabilities and a posterior probability estimate, the class label corresponding to the first configuration or the second configuration of one or more occupants in the space, the posterior probability estimate indicating the first probability or the second probability; a controller configured as such; a system comprising the same.

2. The system according to claim 1, wherein when the class label corresponds to the first configuration, the controller is configured to determine whether the first probability for at least one feature of the shape or pattern of the temperature signal corresponding to the first configuration exceeds a predetermined probability threshold.

3. The system according to claim 2, wherein when the first probability for at least one feature of the shape or pattern of the temperature signal corresponding to the first configuration exceeds the predetermined probability threshold, the controller is configured to provide a notification including information on the first configuration of one or more occupants in the space.

4. If the first probability for at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration does not exceed the predetermined probability threshold, the controller is configured to transmit a signal including information about at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration to another sensor device or another controller in the space. The system according to claim 2.

5. If the first probability for at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration does not exceed the predetermined probability threshold, the controller is configured to receive a signal including information about the shape or pattern of another temperature signal corresponding to the first configuration from another sensor device or another controller in the space. The system according to claim 2.

6. The controller is configured to fuse a signal including information about at least one characteristic of the shape or pattern of the temperature signal corresponding to the first configuration for subsequent determination of the first configuration of one or more occupants in the space with the signal including information about the shape or pattern of the another temperature signal corresponding to the first configuration from the another sensor device or the another controller in the space. The system according to claim 5.

7. The at least one sensor device including the single pixel thermopile is built into a lighting fixture, and the lighting fixture is part of a plurality of connected lighting devices in the space. The system according to claim 1.

8. The at least one sensor device including the single pixel thermopile is provided at an entrance and exit of the space. The system according to claim 1.

9. A method for determining the configuration of one or more occupants in a space, the method comprising: (a) providing at least one sensor device including a single pixel thermopile in the space, the single pixel thermopile being configured to capture a sensor signal regarding one or more occupants in the space; (b) providing a controller that communicates with the at least one sensor device in the space; (c) determining or obtaining, by the controller, training data for detecting a first configuration and a second configuration of one or more occupants in the space; (d) measuring, by the single pixel thermopile, a temperature signal of a detection area in the space that is within the field of view of the single pixel thermopile over time; (e) detecting, by the controller, a change in the temperature signal, the change being equal to or greater than a predetermined step threshold; (f) determining, by the controller, a first probability and a second probability for at least one feature of the shape or pattern of the temperature signal, the first probability and the second probability corresponding to the first configuration and the second configuration of one or more occupants in the space based on the training data, respectively; (g) identifying, by the controller, a class label of the temperature signal using the determined first and second probabilities and a posterior probability estimation, the class label corresponding to the first configuration or the second configuration of one or more occupants in the space, the posterior probability estimation indicating the first probability or the second probability; A method comprising the above.

10. When the class label corresponds to the first configuration of one or more occupants in the space, the method includes determining, by the controller, whether the first probability for the at least one feature of the shape or pattern of the temperature signal corresponding to the first configuration exceeds a predetermined probability threshold. The method according to claim 9.

11. When the first probability for the at least one feature of the shape or pattern of the temperature signal corresponding to the first configuration exceeds the predetermined probability threshold, the method includes providing, by the controller, a notification including information on the first configuration of one or more occupants in the space. The method according to claim 10.

12. If the first probability for at least one feature of the shape or pattern of the temperature signal corresponding to the first configuration does not exceed the predetermined probability threshold, the method includes, by the controller, transmitting a signal including information regarding at least one feature of the shape or pattern of the temperature signal corresponding to the first configuration to another sensor device or another controller in the space. The method according to claim 10.

13. The method includes, for a subsequent determination by one or more occupants in the space regarding the first configuration, fusing the signal including information regarding at least one feature of the shape or pattern of the temperature signal corresponding to the first configuration with at least one other signal from the other sensor device. The method according to claim 12.

14. If the class label corresponds to a second configuration of one or more occupants in the space, the method includes repeating steps (d), (e), (f), and (g) by the controller. The method according to claim 9.

15. The at least one sensor device including the single pixel thermopile is built into a lighting fixture, and the lighting fixture is part of a plurality of connected lighting devices in the space. The method according to claim 9.

Images