Fall detection device, method of detecting a fall of a subject, and computer program product for implementing the method
By integrating multiple fall detection algorithms and environmental sensor information into the fall detection device, the initial state of the object is analyzed and matched with the potential fall type, solving the sensor integration problem in the home environment and improving the reliability and accuracy of fall detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- KONINKLIJKE PHILIPS NV
- Filing Date
- 2019-06-24
- Publication Date
- 2026-06-02
Smart Images
Figure CN112400191B_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the detection of falls, and more specifically to a fall detection device, a method for detecting falls, and a computer program product for implementing a method capable of detecting several different types of falls. Background Technology
[0002] As we age, our physical functions decline. A person's mobility may be affected, and they may experience difficulty maintaining their independence. A large category of difficulties involves falls, which can have disastrous consequences for the health of the person who fell.
[0003] Falls affect millions of people every year and cause numerous injuries, especially among the elderly. In fact, falls are estimated to be one of the top three causes of death among older adults. A fall is defined as a sudden, uncontrolled, and non-intentional downward displacement of the body toward the ground, followed by impact, after which the body remains on the ground.
[0004] A Personal Emergency Response System (PERS) is a system that allows an individual to request help. Using a Personal Help Button (PHB), an individual can press the button to summon assistance in an emergency. However, if an individual suffers a severe fall (e.g., they become confused, or even worse, if they are knocked unconscious), they may be unable to press the button, potentially meaning they won't receive help for an extended period, especially if they live alone. The consequences of a fall can be more severe if the individual remains lying down for a prolonged period.
[0005] Therefore, a PHB can include one or more sensors, such as accelerometers (typically accelerometers that measure acceleration in three dimensions) and barometric pressure sensors (for measuring the PHB's height, height change, or absolute altitude), and can process the sensor outputs to determine if an object has experienced a fall. This processing can involve inferring the occurrence of a fall by processing a time series generated by the accelerometer and barometric pressure sensor. Typically, fall detection algorithms test one or more features, such as, but not limited to, impact, orientation, change of orientation, height change, and vertical velocity. Reliable fall detection is achieved when this set of calculated values for these features differs for a fall from those for other movements that are not falls. Upon detecting a fall, an alarm is triggered by the PHB without the object pressing a button.
[0006] Efforts are underway to develop robust classification methods or processing algorithms for accurate fall detection, as it is clearly important to correctly identify falls to provide assistance and to minimize (or even completely prevent) false alarms (FAs). Therefore, automatic fall detection algorithms are optimized to strike a trade-off between false alarms and the probability of fall detection.
[0007] However, the challenge in achieving reliable fall detection lies in the fact that not all falls are the same, and different types of falls can possess different characteristics. Typically, optimizing a fall detection algorithm means reliably detecting falls from a standing position (i.e., from a standing / upright posture), but this means it might miss falls from lower positions or those involving complex movements. Examples include falling from a chair, falling off a bed, falling while attempting to stand up, or falling while attempting to sit down. Falls can also be phased, as objects don't fall directly to the ground but rather, for example, slide down a wall, grab onto some furniture (e.g., a table, chair, bed, etc.), or fall onto furniture. These issues with reliable fall detection are particularly important for objects using wheelchairs, and there is an additional risk of falls when getting into or out of their wheelchairs. Summary of the Invention
[0008] A current trend is for home or care settings to incorporate a variety of sensors to monitor the home environment or specific objects within it. These sensors are becoming increasingly interconnected, as sensor measurements or analyses of those measurements can be transmitted via wired or wireless connections to other devices (e.g., remote servers, central home monitoring systems, smartphones, etc.) through local networks or the Internet. These connected sensors are often referred to as the Internet of Things (IoT) or the Internet of Medical Technology (IoMT). Because these sensors can monitor where objects are in the environment, what they are doing (e.g., which object an object is using), they can provide information useful for fall detection algorithms (which typically operate on measurements of object movement) to optimize fall detection decisions.
[0009] However, given the large number of different sensor types present in a home or care setting, integrating measurements from actual sensors in the environment into a fall detection algorithm implemented by a PHB or other dedicated fall detector will be challenging. One way to achieve integration is for the PHB or other dedicated fall detector to include discovery and communication protocols for connecting to any possible sensors available in the home or care setting. The PHB or other dedicated fall detector will need to understand all possible configurations, sensor types, formats, and protocols. In this architectural configuration, system maintenance and flexibility will be difficult, and users may face a disappointing experience where adding another sensor that could be used for fall detection in a home environment can be difficult, or even impossible, because it is not supported by their PHB / fall detector software version. Similarly, this type of system installation or setup will be difficult for elderly users (typical users of fall detectors).
[0010] Therefore, there is a need for an improved fall detection device, a method for detecting falls, and a computer program product for implementing the method, which can utilize information obtained from sensors in the object's environment to improve the reliability of fall detection, and in particular improve the reliability of different types of fall detection.
[0011] According to a first aspect, a fall detection device is provided, the fall detection device comprising one or more processing units configured to: obtain a first input indicating which of one or more of a plurality of fall detection algorithms has detected a potential fall of an object, wherein each of the plurality of fall detection algorithms is associated with a corresponding type of fall, and the associated type of potential fall is detected by analyzing a set of movement measurements for the object, wherein each corresponding type of fall has an associated initial state of the object; obtain a second input indicating the state of the object prior to the potential fall, wherein the state of the object is determined by analyzing a set of measurements from one or more sensors in the object's environment; compare the determined state of the object prior to the potential fall with the initial state of each type of fall associated with any potential fall indicated in the first input; and output an indication that the object has fallen if the determined state of the object matches the initial state of any corresponding type of fall associated with any potential fall indicated in the first input. Thus, the first aspect enables the determination, using information obtained from sensors in the object's environment, of whether a potential fall detected by one or more fall detection algorithms suitable for the corresponding type of fall is an actual fall. This improves the reliability of detecting different types of falls.
[0012] In some embodiments, the one or more processing units are further configured to determine that the object has not yet fallen if the determined state of the object does not match the initial state for any corresponding type of fall associated with any potential fall indicated in the first input. This means that potential falls can be identified without considering specific fall detection algorithms (as associated with that type of fall) even when the object is not in the correct initial state for a certain type of fall.
[0013] In some embodiments, the one or more processing units are further configured to: if the determined state of the object does not match the initial state for any corresponding type of fall associated with any potential fall indicated in the first input, then not output an indication that the object has fallen. This means that the fall is not alerted to care providers or other responders unless the object is determined to have fallen.
[0014] In some embodiments, the initial state of the object associated with a type of fall includes any one or more of the following: (i) a standing posture, (ii) a sitting posture, and (iii) a lying posture.
[0015] In some embodiments, the corresponding types of falls associated with various fall detection algorithms include any one or more of the following: (i) falling from a standing position, (ii) falling from a sitting position, (iii) falling from a lying position, (iv) falling while moving from a sitting position to a standing position, (v) falling while moving from a standing position to a sitting position, (vi) falling from a standing position onto furniture, and (vii) an object sliding down a wall from a standing position.
[0016] In some embodiments, the one or more processing units are configured to obtain the first input by: analyzing a set of motion measurements of an object using multiple fall detection algorithms to detect whether the object has experienced a potential fall of a corresponding type associated with each fall detection algorithm; and forming the first input based on the results of analyzing the set of motion measurements using multiple fall detection algorithms. This has the advantage of enabling the fall detection algorithms and comparisons with the object's state to be performed in the same device, thus eliminating the need for a separate fall detection device. In these embodiments, the one or more processing units can also be configured to receive a set of motion measurements of the object from one or more sensors carried or worn by the object.
[0017] In these embodiments, the set of motion measurements can relate to a first time period, and the one or more processing units are configured to analyze the set of motion measurements using the multiple fall detection algorithms to detect whether the object has experienced a potential fall of a related type during the first time period. This means that all the fall detection algorithms operate on the same motion measurements to identify the associated type of fall; that is, each set of motion measurements is evaluated for each type of fall among different types.
[0018] In some embodiments, each of the plurality of fall detection algorithms may include a first fall detection algorithm having a corresponding threshold or set of thresholds for detecting potential falls of the associated type. In these embodiments, the first fall detection algorithm may include a log-likelihood ratio (LLR) table. In these embodiments, each of the plurality of fall detection algorithms may correspond to a corresponding point in the receiver operating characteristic (ROC) for the first fall detection algorithm. In alternative embodiments, each of the plurality of fall detection algorithms may include a corresponding set of parameters to be analyzed from the set of motion measurements.
[0019] In alternative embodiments, the one or more processing units are configured to receive a first input from a fall detection device carried or worn by the object. These embodiments offer the advantage that the fall detection device can operate alongside existing fall detection equipment.
[0020] In some embodiments, the indication is a fall alert, and the indication is output to a call center or care provider device.
[0021] In some embodiments, the one or more processing units are configured to obtain a second input by analyzing a set of measurements from one or more sensors in the object's environment to determine the state of the object prior to a potential fall; and to form the second input based on the results of the analysis of the set of measurements from one or more sensors in the object's environment. This has the advantage of enabling state determination and comparison with the outputs of multiple fall detection algorithms to be performed in the same device, thus eliminating the need for a separate monitoring system.
[0022] In alternative embodiments, the one or more processing units are configured to obtain a second input from a monitoring system of one or more sensors in the environment including the object. These embodiments have the advantage that the fall detection device can be used with existing monitoring systems.
[0023] In some embodiments, the one or more sensors in the environment of the object include one or more of the following: (i) a sensor for measuring whether the object is using a piece of furniture; (ii) a sensor for measuring whether the object is using a wheelchair; (iii) a sensor for measuring whether the object is in a room; and (iv) a sensor for measuring whether the object is using an object in the environment.
[0024] In some embodiments, the state of the object includes any one or more of the following: (i) sitting in a chair or bed, (ii) lying in bed, (iii) walking or standing, (iv) sitting in a wheelchair, (v) about to enter a wheelchair.
[0025] According to a second aspect, a method for detecting falls is provided, the method comprising: obtaining a first input indicating which one or more fall detection algorithms among a plurality of fall detection algorithms has detected a potential fall of the object, wherein each of the plurality of fall detection algorithms is associated with a corresponding type of fall, and detecting the associated type of potential fall by analyzing a set of movement measurements for the object, wherein each corresponding type of fall has an associated initial state of the object; obtaining a second input indicating the state of the object prior to a potential fall, wherein the state of the object is determined by analyzing a set of measurements from one or more sensors in the object's environment; comparing the determined state of the object prior to a potential fall with the initial state for each type of fall associated with any potential fall indicated in the first input; and outputting an indication that the object has fallen if the determined state of the object matches the initial state of any corresponding type of fall associated with any potential fall indicated in the first input. Thus, the second aspect enables the use of information obtained from sensors in the object's environment to determine whether a potential fall detected by one or more fall detection algorithms suitable for the corresponding type of fall is an actual fall. This improves the reliability of detecting different types of falls.
[0026] In some embodiments, the method further includes determining that the object has not fallen if the determined state of the object does not match the initial state for any corresponding type of fall associated with any potential fall indicated in the first input. This means that, even if the object is not in the correct initial state in which a certain type of fall has already occurred, potential falls identified by a specific fall detection algorithm (as associated with that type of fall) can be disregarded.
[0027] In some embodiments, if the determined state of the object does not match the initial state of any corresponding type of fall associated with any potential fall indicated in the first input, no indication that the object has fallen is output. This means that the fall is not alerted to care providers or other responders unless the object has already been determined to have fallen.
[0028] In some embodiments, the initial state of the object associated with a type of fall includes any one or more of the following: (i) a standing posture, (ii) a sitting posture, and (iii) a lying posture.
[0029] In some embodiments, the corresponding types of falls associated with the various fall detection algorithms include any one or more of the following: (i) falling from a standing position, (ii) falling from a sitting position, (iii) falling from a lying position, (iv) falling while moving from a sitting position to a standing position, (v) falling while moving from a standing position to a sitting position, (vi) falling from a standing position onto furniture, and (vii) an object sliding down a wall from a standing position.
[0030] In some embodiments, obtaining the first input includes: analyzing a set of motion measurements of the object using the multiple fall detection algorithms to detect whether the object has experienced a potential fall of a corresponding type associated with each fall detection algorithm; and forming the first input based on the results of the analysis of the set of motion measurements using the multiple fall detection algorithms. This has the advantage of enabling the fall detection algorithms and comparisons with the object's state to be performed in the same device, thus eliminating the need for a separate fall detection device. In these embodiments, the method may also include receiving a set of motion measurements of the object from one or more sensors carried or worn by the object.
[0031] In these embodiments, the set of motion measurements can relate to a first time period, and the analysis step includes using the multiple fall detection algorithms to analyze the set of motion measurements to detect whether the object has experienced a potential fall of a related type during the first time period. This means that all the fall detection algorithms operate on the same motion measurements to identify the associated type of fall; that is, each set of motion measurements is evaluated for each type of fall of a different type.
[0032] In some embodiments, each of the plurality of fall detection algorithms may include a first fall detection algorithm having a corresponding threshold or set of thresholds for detecting potential falls of the associated type. In these embodiments, the first fall detection algorithm may include a log-likelihood ratio (LLR) table. In these embodiments, each of the plurality of fall detection algorithms may correspond to a corresponding point in the receiver operating characteristic (ROC) for the first fall detection algorithm. In alternative embodiments, each of the plurality of fall detection algorithms may include a corresponding set of parameters to be analyzed from the set of motion measurements.
[0033] In alternative embodiments, obtaining the first input includes obtaining the first input from a fall detection device carried or worn by the object. These embodiments have the advantage that the method can operate with existing fall detection devices.
[0034] In some embodiments, the indication is a fall alert, and the indication is output to a call center or care provider device.
[0035] In some embodiments, the step of obtaining the second input includes: analyzing a set of measurements from one or more sensors in the environment of the object to determine the state of the object prior to a potential fall; and forming the second input based on the results of the analysis of the set of measurements from one or more sensors in the environment of the object. This has the advantage of enabling state determination and comparison with the outputs of multiple fall detection algorithms to be performed in the same device, thus eliminating the need for a separate monitoring system.
[0036] In an alternative embodiment, obtaining the second input includes obtaining the second input from a monitoring system of one or more sensors in the environment including the object. These embodiments have the advantage that the method can be used with existing monitoring systems.
[0037] In some embodiments, the one or more sensors in the environment of the object include one or more of the following: (i) a sensor for measuring whether the object is using a piece of furniture; (ii) a sensor for measuring whether the object is using a wheelchair; (iii) a sensor for measuring whether the object is in a room; and (iv) a sensor for measuring whether an object in the environment is being used.
[0038] In some embodiments, the state of the object includes any one or more of the following: (i) sitting in a chair or bed, (ii) lying in bed, (iii) walking or standing, (iv) sitting in a wheelchair, (v) about to enter a wheelchair.
[0039] According to a third aspect, a computer program product including a computer-readable medium is provided, wherein computer-readable code is embedded therein, the computer-readable code being configured to cause the computer or processor, when executed by a suitable computer or processor, to perform the method according to the second aspect or any embodiment thereof.
[0040] According to a fourth aspect, a fall detection device is provided, comprising: one or more motion sensors for measuring the movement of an object; one or more processing units configured to receive a set of motion measurement results of the object from the one or more motion sensors; analyzing the set of motion measurement results using multiple fall detection algorithms to detect whether the object has experienced a potential fall of a corresponding type associated with each fall detection algorithm, wherein each corresponding type of fall has an associated initial state of the object; and forming a first input based on the result of the analysis of the set of motion measurement results using the multiple fall detection algorithms; and a fall detection apparatus according to the first aspect above. Therefore, in this aspect, the fall detection apparatus, or the function defined in the first aspect, is part of or implemented by the fall detection device.
[0041] According to a fifth aspect, a monitoring system is provided, comprising one or more processing units configured to: receive a set of measurements from one or more sensors in the environment of an object; analyze the set of measurements to determine the state of the object prior to a potential fall; and form a second input based on the analysis of the set of measurements; and according to the fall detection device of the first aspect above. Therefore, in this aspect, the fall detection device, or the function defined in the first aspect, is part of or implemented by the monitoring system.
[0042] These and other aspects will become apparent and be illustrated with reference to one or more embodiments described below. Attached Figure Description
[0043] Exemplary embodiments will now be described by way of example only with reference to the following accompanying drawings, in which:
[0044] Figure 1 This is a block diagram of a library device according to an exemplary embodiment; and
[0045] Figure 2 This is a flowchart illustrating a method according to an exemplary embodiment. Detailed Implementation
[0046] As described above, the present invention aims to improve the reliability of fall detection by utilizing information obtained from sensors in the environment of the object, and in particular to improve the reliability of detection for different types of falls, while minimizing the occurrence of false alarms.
[0047] Fall detection algorithms can be optimized to detect different types of falls, but this means that other types of falls may not be reliably detected by the algorithm. For example, an algorithm optimized to reliably detect falls from a standing position (including while walking) may not reliably detect falls when getting up from a chair, because the characteristics of a fall from a standing position may not be present in the movement measurements corresponding to a fall when attempting to stand up, and vice versa.
[0048] Therefore, several different fall detection algorithms, each optimized for the corresponding type of fall (e.g., falling from a standing position, falling while attempting to stand up, etc.), can be used to evaluate the motion measurement results of an object, and each algorithm can provide an output indicating whether a fall is likely to be detected in the motion measurement results. It is possible that, depending on the specific configuration of the algorithm and the specific motion measurement results, more than one fall detection algorithm can indicate a fall at a given time.
[0049] One way to implement different fall detection algorithms is to use the same set of features / parameters (e.g., impact, height change, orientation change, etc.) and the same log-likelihood ratio (LLR) table, but each algorithm can use different decision thresholds for the total LLR value depending on the type of fall. In other words, different operating points on the receiver operating characteristic (ROC) curve can be used for each fall detection algorithm / fall type. As is well known, the reliability of a classification method can be visualized using an ROC curve, where the detection probability is plotted relative to the false alarm rate, and the operating point of the algorithm on the ROC curve can be selected to achieve the desired detection probability or false alarm rate. As is known from detection theory, the optimal detector is found by testing the so-called likelihood ratio. This ratio expresses the probability of a given feature value (e.g., the size of the impact) in a fall situation divided by the probability of that given feature value in a non-fall situation (i.e., any movement that produces the same value but is not a fall). The larger the ratio, the more likely the observed event (in this example, an impact) is due to a fall. Comparison with (by design) setting a threshold allows the detector to conclude whether the event is a fall. Likelihood ratios for a range of eigenvalues (impact size in this example) are typically stored in a table. For ease of calculation, the logarithm of the ratio is stored instead of the ratio itself.
[0050] Another way to implement different fall detection algorithms is to use different sets of features / parameters, for example, for one or more fall detection algorithms suitable for the type of fall to be detected. For example, the set of parameters used by a fall detection algorithm to detect a fall when an object is approaching or sitting in a chair (including a wheelchair) may differ from the set of parameters used by a fall detection algorithm to detect a fall when the object is walking. Exemplary features / parameters that can be used include a time window for calculating height changes, the required height change at the event, and a decision threshold for the overall probability between a fall and no fall. Alternatively or additionally, the LLR tables used by each algorithm can also be different, wherein the LLR tables are fitted to a distribution corresponding to the associated fall type. For example, an LLR table for height changes when falling from a chair may have its maximum probability at lower height changes compared to an LLR table used for a fall from a standing position. Similarly, impact and / or orientation LLR tables can reflect different log-likelihood values. It is also possible or alternatively that the methods for calculating features / parameters differ between different algorithms, for example, by using different signal processing techniques.
[0051] As described above, it is desirable to utilize information available from one or more sensors in the home environment, such as sensors that are part of a home environment system. Therefore, it is possible to "filter" or "verify" the output of any fall detection algorithm indicating a potential fall may have occurred using the state of an object derived from measurements from one or more environmental sensors. For example, based on a set of movement measurements, a fall detection algorithm optimized for detecting falls out of bed might indicate that an object may have fallen (wherein, a fall detection algorithm optimized for other types of falls does not indicate a potential fall), but the state of the object derived from one or more environmental sensors might indicate that the object is walking around the house (and that the object is not in bed when the potential fall is indicated). In that case, the potential fall indicated by the fall detection algorithm optimized for falls out of bed can be disregarded or ignored because it is inconsistent with the current state of the object provided by one or more environmental sensors. On the other hand, if one or more environmental sensors indicate that the object was in bed when (and / or before) a potential fall was detected, then the potential fall is consistent with the state of the object, and a fall can be detected with certainty (and an alarm is triggered and / or an alert is sent).
[0052] In a particular embodiment of the invention, a fall detection device carried or worn by an object (e.g., a personal help button (PHB) including one or more motion sensors) can use a range of fall detection algorithms to evaluate motion measurement results, wherein each algorithm determines, for a given (triggered) event (i.e., the set of motion measurement results meets a certain triggering condition), whether the event is a fall under a hypothetical specific condition (e.g., a fall from a standing position, a fall from a chair, a fall from a bed, etc.). The algorithms can share computing components, i.e., the algorithms can be evaluated by the same processing unit in the fall detection device.
[0053] In some embodiments, the first part of the analysis of motion measurement results can be common to all fall detection algorithms, where an individual fall detection algorithm is used if a trigger condition is met. Alternatively, the first part of the analysis can be different for different fall detection algorithms. In either case, motion measurement results (e.g., acceleration, air pressure, etc.) are received and tests can be run on the measurement results to determine whether a trigger condition has been met. For example, it can be tested whether the air pressure has increased by an amount greater than the air pressure equivalent to a predetermined height change (e.g., 50 cm) relative to an earlier time period (e.g., 2 seconds). Accelerometer-based trigger conditions can be observed by observing orientation changes or impacts (e.g., the norm of the accelerometer signal exceeds a certain threshold). If a trigger occurs in this way (i.e., the trigger condition is met), a segment of motion measurement results (i.e., a segment of motion measurement signal) near the time the trigger condition was met is forwarded for further processing. In this way, the (potentially continuous) sensor signal / measurement results are converted into a series of (discrete) events using the trigger condition. The trigger condition should require low complexity and low power consumption for evaluation. It should pass through all “real” falls and through as few “non-falls” as possible (although it should be recognized that the main task of non-fall suppression is the task of subsequent fall detection algorithms, but the ratio of these non-fall events sets the call rate of the fall detection device).
[0054] If one or more algorithms determine that the event is a fall, each positive decision (i.e., a fall is detected) can be transmitted (e.g., sent) to a central control console (hereinafter referred to as the fall detection device) in the home or care environment. Each positive decision can be labeled using the type of algorithm / situation that generated the positive decision (i.e., a fall from a standing position, a fall from a chair, a fall from a bed, etc.).
[0055] The central control console can be connected to a pre-existing home or care environment monitoring system (e.g., a burglary monitoring system, fire / smoke detection system, and / or activities of daily living (ADL) monitoring system) (or at least be able to receive information from it). The monitoring system implements and processes the detection and communication with any environmental sensors in the home or care environment (thus avoiding any need for fall detection devices or a central control console to do so). The monitoring system is also capable of implementing and executing algorithms that analyze the environmental sensor measurements to determine the status of an object in the home or care environment. This status is then provided to the central control console.
[0056] The environmental sensors can include sensors that can be placed on or associated with furniture, such as chairs, benches, beds, cupboards, showers, bedside tables, etc. These sensors can be used to measure whether an object is using a particular piece of furniture and / or is in the vicinity of a particular piece of furniture.
[0057] When the central control console receives an indication that a fall has been detected by the fall detection device, along with one or more associated fall type tags, the console tests whether the fall type matches the current state inferred by the monitoring system. If they match, an alert that the object has fallen is forwarded to a call center or other assistance provider (e.g., emergency services). In some implementations, if a fall detection algorithm used to detect a fall from a standing position (i.e., standing) detects a potential fall, an alert or warning can always be triggered (e.g., it can be excluded from the test relative to the current state, or the mismatch with the current state can be ignored).
[0058] In another specific embodiment of the invention (which can be used in conjunction with or independently of the home monitoring system used in the above specific embodiments), environmental sensors can be provided to detect when a subject is in and / or about to sit in a wheelchair (i.e., the sensors can be used to detect whether the subject is standing in front of the wheelchair). Examples of such sensors include passive infrared (PIR) sensors, ultrasonic (US) sensors, radar-based sensors, near field communication (NFC) sensors, pressure sensors (i.e., for detecting pressure or force applied to parts of the wheelchair, such as the seat and / or handles / armrests), and light sensors (e.g., photodiodes) for sensing light from sources such as lasers or light-emitting diodes (LEDs). A fall detection algorithm can be provided or used to assess whether a fall has occurred from the wheelchair (either a fall from the wheelchair or a fall while attempting to sit down and / or get up from the wheelchair). A positive fall indication from the fall detection algorithm can be compared with measurements from environmental sensors associated with the wheelchair, and a fall is detected if the subject is sitting in or near the wheelchair at a time corresponding to the time the algorithm detects the fall.
[0059] In some embodiments, if the wheelchair is an electric wheelchair and / or otherwise has electrically actuated brakes (for preventing movement of the wheelchair), the brakes can be automatically actuated to prevent movement of the wheelchair if an environmental sensor detects that an object is standing in front of the wheelchair. The brakes can be released if the sensor (or another) detects that the object is sitting in the wheelchair (unless manually applied by the object).
[0060] It will be appreciated that in some implementations, the environmental sensor is capable of operating continuously or periodically to monitor the environment / object, in which case the state of the object can be determined continuously or periodically. Alternatively, the environmental sensor is capable of operating continuously or periodically to monitor the environment / object, but can only perform processing to determine the state of the object when needed (e.g., after receiving a positive fall indication from one or more fall detection algorithms). As another alternative, the environmental sensor can measure the environment / object only when requested to do so (e.g., after receiving a positive fall indication from one or more fall detection algorithms). This alternative reduces the system's energy consumption.
[0061] Figure 1 An exemplary fall detection device 2, which can be used to implement various embodiments of the invention, is illustrated. Device 2 is shown as part of a system 4, which includes one or more motion sensors 6 provided to measure the movement of an object and one or more environmental sensors 8 provided to measure aspects of the object's environment. The fall detection device 2 is provided to detect whether an object has fallen by comparing the state of the object prior to a potential fall (as determined based on measurements from the environmental sensors 8) with an initial state of a type of fall associated with any fall detection algorithm for a potential fall of the object (determined based on measurements from the motion sensors 6), and outputting an indication that the object has fallen when a match exists between the state and the initial state. Thus, the fall detection device 2 can also be referred to as a fall decision device 2 because it makes a final decision on whether a fall has occurred and whether an alarm or alert should be triggered.
[0062] In some embodiments, measurements from one or more motion sensors 6 are provided to a fall detection device 2, and the fall detection device 2 analyzes the motion measurements using multiple fall detection algorithms to detect a potential fall of the subject. In other embodiments, one or more motion sensors 6 can be integrated with the fall detection device 2. In this case, the fall detection device 2 can be worn or carried by the subject and can be in the form of a watch, bracelet, necklace, chest strap, etc. In other embodiments, one or more motion sensors 6 are part of a separate fall detection device 10 (indicated by the dashed box 10 around one or more motion sensors 6), and the fall detection device 10 applies fall detection algorithms to the motion measurements to detect a potential fall of the subject. The fall detection device 10 can be carried or worn by the subject and can include, for example, a PHB (Prognostics and Hygiene Device). The fall detection device 10 can be in the form of a watch, bracelet, necklace, chest strap, etc. It will be appreciated that, when present, the fall detection device 10 merely provides input to the fall detection device 2, indicating the results of the analysis of the motion measurements by multiple fall detection algorithms. The fall detection device 2 determines whether a fall alert should be issued based on a comparison between the results of a fall detection algorithm and the state of the object determined from one or more environmental sensors 8. In some alternative embodiments, the function of the fall detection device 2 described herein is part of or implemented by the fall detection device 10. In these embodiments, the fall detection device 2 can be worn or carried by the object and can be in the form of a watch, bracelet, necklace, chest strap, etc., and can include or be connected to one or more motion sensors 6.
[0063] In some embodiments, measurements from one or more motion sensors 8 are provided to the fall detection device 2, and the fall detection device 2 analyzes the measurements to determine the state of the object. In other embodiments, one or more of the environmental sensors 8 can be integrated with the fall detection device 2 (wherein, optionally, one or more other environmental sensors 8 are independent of the fall detection device 2). In other embodiments, the environmental sensors 8 are part of the monitoring system 12 (indicated by the dashed boxes 12 surrounding the environmental sensors 8). In some alternative embodiments, the fall detection device 2 described herein functions as part of or implemented by the monitoring system 12.
[0064] It will be appreciated that various combinations of the embodiments in the first two paragraphs are possible. For example, the fall detection device 2 may perform all processing on the sensor measurement results (e.g., analysis of motion measurement results received from one or more motion sensors 6 using multiple fall detection algorithms, and analysis of environmental sensor measurement results received from one or more environmental sensors 8 (where one or more motion sensors 6 and one or more environmental sensors 8 may be integrated with the fall detection device 2) to determine the state of the object), perform no processing on the sensor measurement results (e.g., the fall detection device 2 receives the results of fall detection algorithm analysis from the fall detection device 10 and receives the state of the object from the monitoring system 12), or perform processing on one set of sensor measurement results while receiving the result of processing another set of sensor measurement results. In any of the embodiments above, one or more motion sensors 6 are carried or worn by the object, and one or more environmental sensors 8 are located in the object's environment (i.e., they are worn or carried by the object).
[0065] The fall detection device 2 includes a processing unit 14 that controls the operation of the fall detection device 2 and can be configured to run or perform the methods described herein. The processing unit 14 can be implemented in numerous ways using software and / or hardware to perform the various functions described herein. The processing unit 14 may include one or more microprocessors or digital signal processors (DSPs) that can be programmed using software or computer program code to perform desired functions and / or control the components of the processing unit 14 to achieve desired functions. The processing unit 14 may be implemented as a combination of dedicated hardware to perform some functions (e.g., amplifiers, preamplifiers, analog-to-digital converters (ADCs) and / or digital-to-analog converters (DACs)) and processors (e.g., one or more programmed microprocessors, controllers, DSPs, and associated circuitry) to perform other functions. Examples of components that may be employed in the various embodiments of this disclosure include, but are not limited to, conventional microprocessors, DSPs, application-specific integrated circuits (ASICs), and field-programmable gate arrays (FPGAs).
[0066] Processing unit 14 is connected to memory unit 16, which is capable of storing data, information, and / or signals for use by processing unit 14 in controlling the operation of fall detection device 2 and / or in running or performing the methods described herein. In some embodiments, memory unit 16 stores computer-readable code that can be executed by processing unit 14, causing processing unit 14 to perform one or more functions, including the methods described herein. Memory unit 16 can include any type of non-transient machine-readable medium, such as cache or system memory, including volatile and non-volatile memory, such as random access memory (RAM), static RAM (SRAM), dynamic RAM (DRAM), read-only memory (ROM), programmable ROM (PROM), erasable PROM (EPROM), and electrically erasable PROM (EEPROM), implemented as memory chips, optical discs (such as compact discs (CD), digital versatile optical discs (DVD), or Blu-ray discs), hard disks, magnetic tape storage solutions, or solid-state devices, including memory sticks, solid-state drives (SSDs), memory cards, etc.
[0067] The fall detection device 2 also includes interface circuitry 18 for establishing data connections to and / or exchanging data with other devices, including any one or more of servers, databases, user equipment, and sensors. The connection can be direct or indirect (e.g., via the Internet), and therefore interface circuitry 18 can connect the fall detection device 2 to a network such as the Internet via any desired wired or wireless communication protocol. For example, interface circuitry 18 can operate using WiFi, Bluetooth, Zigbee, or any cellular communication protocol (including, but not limited to, GSM, UMTS, LTE, LTE-Advanced, etc.). In the case of a wireless connection, interface circuitry 18 (and therefore the fall detection device 2) may include one or more suitable antennas for transmitting / receiving over a transmission medium (e.g., air). Alternatively, in the case of a wireless connection, interface circuitry 18 may include modules (e.g., connectors or plugs) to allow interface circuitry 18 to be connected to one or more suitable antennas external to the fall detection device 2 for transmitting / receiving over a transmission medium (e.g., air). The interface circuit 18 is connected to the processing unit 14.
[0068] Interface circuit 18 can be used to receive motion measurement results from one or more motion sensors 6, or, if one or more motion sensors 6 are part of fall detection device 10, interface circuit 18 can be used to receive the results of analysis of motion measurement results by multiple fall detection algorithms. Interface circuit 18 can also be used to receive measurement results from one or more environmental sensors 8, or, if one or more environmental sensors 8 are part of monitoring system 12, interface circuit 18 can be used to receive the determined state of the object.
[0069] Interface circuit 18 can also be used to output an indication that an object has fallen. In that case, interface circuit 18 can transmit the indication to a call center or emergency services and / or to the user equipment of a physician or care provider.
[0070] In some embodiments, the fall detection device 2 includes a user interface 20, which includes one or more components that enable a user of the fall detection device 2 (e.g., an object or a care provider for the object) to input information, data, and / or commands into the fall detection device 2, and / or enable the fall detection device 2 to output information or data to the user. The output may be an audible alarm or alert that the object has fallen. The user interface 20 may include any suitable input(s) components, including, but not limited to: a keyboard, keypad, one or more buttons, switches or dial pads, a mouse, a tracking pad, a touchscreen, a stylus, a camera, a microphone, etc., and the user interface 20 may include any suitable output(s) components, including, but not limited to: a display screen, one or more lights or light elements, one or more speakers, a vibration element, etc.
[0071] The fall detection device 2 can be any type of electronic or computing device. For example, the fall detection device 2 can be a server, computer, laptop computer, tablet computer, smartphone, smartwatch, or a part thereof.
[0072] It will be recognized that the actual implementation of the fall detection device 2 can include in Figure 1 Other components besides those shown. For example, the fall detection device 2 may also include a power source, such as a battery, or a component for enabling the fall detection device 2 to be connected to an AC power source.
[0073] In embodiments where motion sensors (one or more) are part of fall detection device 10, fall detection device 10 may include a processing unit (shown by dashed box 22) for analyzing motion measurement results using various fall detection algorithms and determining whether an object may have experienced a fall. Fall detection device 10 may also include interface circuitry (shown by dashed box 24) for enabling the transmission of the analysis results of the motion measurement results to fall detection device 2. Processing unit 22 and / or interface circuitry 24 may be implemented in a manner similar to processing unit 14 and / or interface circuitry 18 in fall detection device 2.
[0074] In embodiments where one or more environmental sensors 8 are part of monitoring system 12, monitoring system 12 may include a processing unit (shown by dashed box 26) for analyzing environmental sensor measurements and determining the state of an object. Monitoring system 12 may also include interface circuitry (shown by dashed box 28) for enabling the determined state to be transmitted to fall detection device 2. Processing unit 26 and / or interface circuitry 28 may be implemented in a manner similar to processing unit 14 and / or interface circuitry 18 in fall detection device 2.
[0075] One or more motion sensors 6 may include one or more of any type of sensor for measuring the movement of an object or for providing a measurement result indicating the movement of the object. For example, one or more motion sensors 6 may include any one or more of the following: accelerometer, magnetometer, satellite positioning system receiver (e.g., GPS receiver, GLONASS receiver, Galileo positioning system receiver), gyroscope, and barometric pressure sensor (which can provide a measurement result indicating the altitude of the object or the height / altitude change of the object).
[0076] One or more environmental sensors 8 may include sensors for monitoring one aspect of the environment or one aspect of objects in the environment. For example, one or more environmental sensors 8 may include one or more sensors 8 for detecting whether an object is using a piece of furniture, one or more sensors 8 for measuring or detecting whether an object is using a wheelchair, one or more sensors 8 for measuring whether an object is in a particular room, and / or one or more sensors 8 for measuring whether an object in the environment is being used. One or more environmental sensors 8 may be or include any one or more of the following: accelerometer, gyroscope, PIR sensor, US sensor, radar-based sensor, light-based sensor, radio frequency (RF) signal-based sensor (e.g., using WiFi, Bluetooth, Zigbee, etc.) from which signal strength measurements can be obtained, NFC sensor, pressure sensor (i.e., for detecting pressure or force applied to a part of an object), camera, etc.
[0077] In some embodiments, in addition to one or more motion sensors 6, one or more physiological characteristic sensors may be provided for monitoring or measuring the physiological characteristics of the object, and these physiological characteristics may be evaluated as part of one or more fall detection algorithms. For example, after a fall, physiological characteristics such as heart rate, skin conductivity, respiratory rate, blood pressure, and / or body temperature may change, and therefore the evaluation of these measurements can provide useful information for determining whether the object has fallen. The one or more physiological characteristic sensors may include photoplethysmography (PPG) sensors, skin conductivity sensors, blood pressure monitors, thermometers, etc., capable of measuring heart rate, heart rate-related characteristics, and respiratory rate.
[0078] It will be appreciated that, in the case where environmental sensor 8 is used to monitor a specific object (e.g., a specific piece of furniture), one or more environmental sensors 8 may include corresponding environmental sensors 8 for monitoring the respective pieces of furniture (e.g., providing corresponding pressure sensors on each chair in the environment). Similarly, in the case where environmental sensor 8 is used to monitor the presence of an object in a specific room, one or more environmental sensors 8 may include corresponding environmental sensors 8 for monitoring the corresponding room (e.g., providing corresponding PIR sensors in bedrooms, kitchens, bathrooms, etc.).
[0079] Figure 2 The flowchart illustrates an exemplary method according to the techniques described herein. One or more steps of the method can be performed by the processing unit 14 in device 2, in combination with any one or more of the memory unit 16, interface circuitry 18, and user interface 20, as appropriate. The processing unit 14 can perform one or more steps in response to executing computer program code, which can be stored on a computer-readable medium, such as, for example, the memory unit 16.
[0080] In the first step, step 101, processing unit 14 receives input (referred to as "first" input for clarity) indicating which of one or more of a variety of fall detection algorithms has detected a potential fall on the object. Each fall detection algorithm is associated with a corresponding type of fall and detects a potential fall of the associated type by analyzing a set of movement measurements for the object. Each corresponding type of fall has an associated initial state of the object, i.e., the posture or state of the object immediately preceding the fall.
[0081] Some exemplary types of falls that can be detected using the corresponding fall detection algorithm and its corresponding initial state include (but are not limited to) any one or more of the following: falling from a standing posture, including falling while walking, jogging or running (where the initial state is standing), falling from a sitting posture (where the initial state is sitting), falling from a lying posture (where the initial state is lying), falling while moving from a sitting posture to a standing posture (where the initial state is sitting), falling while moving from a standing posture to a sitting posture (where the initial state is standing), falling from a standing posture onto furniture (where the initial state is standing), and falling down a wall from a standing posture (where the initial state is standing).
[0082] In some embodiments, step 101 includes obtaining a first input from a fall detection device 10 carried or worn by the object.
[0083] In an alternative embodiment, step 101 includes a processing unit 14 that determines a first input by analyzing a set of motion measurements from one or more motion sensors 6 using multiple fall detection algorithms to detect whether the object has experienced a potential fall of a corresponding type associated with each fall detection algorithm. The first input can be formed based on the results of the analysis of the set of motion measurements using multiple fall detection algorithms. In this embodiment, the processing unit 14 is capable of receiving the set of motion measurements obtained using one or more motion sensors 6.
[0084] In any embodiment of step 101, the set of movement measurement results pertains to a first time period. The fall detection device 10 or processing unit 14 uses the multiple fall detection algorithms to analyze the set of movement measurement results to detect whether the object has experienced a potential fall of a related type during the first time period. That is, the multiple fall detection algorithms are used to evaluate measurement results over the same time period to identify potential falls.
[0085] In some embodiments, each of the multiple fall detection algorithms uses the same (shared) fall detection algorithm (e.g., the extracted feature set), but has a corresponding threshold or set of thresholds for detecting potential falls of the associated type. The shared fall detection algorithm can include an LLR table. Each of the multiple fall detection algorithms can correspond to a corresponding point in the ROC for the shared fall detection algorithm.
[0086] Alternatively, each of the multiple fall detection algorithms may include a corresponding set of parameters or features to be analyzed or extracted from the set of movement measurements.
[0087] It will be recognized that each fall detection algorithm can be trained or configured based on known types of falls. For example, it is possible to train the parameters, features, LLR table, and / or thresholds of a fall detection algorithm for detecting falls from a lying position based on movement measurements from known falls from a bed.
[0088] Next, in step 103, processing unit 14 receives an input (referred to as the "second" input for clarity) indicating the state of the object prior to a potential fall. The state of the object is determined based on an analysis of a set of measurements from one or more environmental sensors 8 in the object's environment.
[0089] Step 103 can include obtaining a second input from a monitoring system 12, which includes one or more sensors 8 in the environment of the object.
[0090] Alternatively, step 103 may include a processing unit 14 that receives a set of measurements from one or more sensors 8 in the object's environment, analyzes the set of measurements from the one or more sensors 8 to determine the state of the object prior to a potential fall, and forms a second input based on the results of the analysis of the set of measurements from the one or more sensors in the object's environment.
[0091] The state of the object indicated in the second input can include any one or more of the following: sitting in a chair or bed, lying in bed, walking (including jogging or running) or standing, sitting in a wheelchair, and about to enter a wheelchair.
[0092] Techniques for analyzing environmental sensor measurements to determine the current state of an object are known in the prior art and, without detail herein, the current state of an object includes, for example, its location (the room it is in), the object it is using (e.g., sitting in a chair, pouring water from a teapot). Such processing techniques are known, for instance, in determining an object's activities of daily living (ADL). In any case, it will be appreciated that many of these processing techniques are simple and straightforward to implement. For example, if a pressure sensor on a chair indicates that a person is sitting in a chair, it can be inferred that the object is sitting in the chair associated with that pressure sensor. In a similar example, several pressure sensors can be provided at different locations on a bed, and high pressure measured by one sensor can indicate that the object is sitting on the bed, while high pressure measured by several sensors can indicate that the object is lying on the bed. If an object is detected in the living room and the television is on, it can be inferred that the object is sitting.
[0093] In step 105, the determined state of the object prior to the potential fall (from the second input) is compared with the initial state for each type of fall associated with any potential fall indicated in the first input. That is, for any type of fall indicated in the first input, the initial state is compared with the determined state of the object.
[0094] Then, in step 107, if the determined state of the object matches the initial state of any corresponding type of fall associated with any potential fall indicated in the first input, a fall is detected, and the fall detection device 2 outputs an indication that a fall has occurred. This indication can be a fall alert. For example, if the first input indicates two potential falls, one from a fall detection algorithm evaluating a fall from a standing position to a sitting position, and the other from a fall detection algorithm evaluating a fall from a sitting position, and the determined state prior to the potential fall is that the object is sitting in a chair, then a match occurs, and a fall from a sitting position is identified.
[0095] The indication may be output in the form of an audible alarm, a visible message or light, or a signal transmitted to a care provider's device, physician's device, call center, or emergency services.
[0096] If the determined state of the object in step 105 does not match the initial state of any corresponding type of fall associated with any potential fall indicated in the first input, then processing unit 14 determines that the object has not fallen. In this case, no indication that the object has fallen is output. In the example above, if the determined state prior to the potential fall is that the object is lying in bed, there is no match, and no fall is detected.
[0097] Therefore, the above methods offer several improvements to the reliability of fall detection. First, different fall detection algorithms are optimized for detecting specific types of falls (e.g., falling from a standing position, falling while attempting to stand, etc.), increasing the chances of successfully detecting specific types of falls. However, recognizing that these optimized fall detection algorithms have a higher false alarm rate in environments where the object is not in an appropriate initial state (e.g., the output of a fall detection algorithm from a standing position would be less reliable if the object is lying down rather than standing), sensors in the object's environment are used to determine the object's state and to check the credibility of any indicated potential fall. Therefore, in the case where one of several fall detection algorithms has detected a potential fall, the object's state prior to the potential fall is checked relative to the initial state associated with the fall detection algorithm that detected the potential fall, to confirm that the potential fall was credible in the object's given state prior to detection. Thus, using the object's state as a check relative to a positive detection of a potential fall improves the reliability of fall detection. In the case of multiple indications of potential falls in various fall detection algorithms, the object's state prior to the potential fall is examined relative to the initial state associated with each of the multiple fall detection algorithms that detected the potential fall. This determines whether any potential fall is credible given the object's state prior to the detected potential fall. Therefore, the object's state serves as a check for a positive detection of a potential fall relative to multiple fall detection algorithms, and if a fall has occurred, allows for more reliable detection of that type of fall. In any of the above examples, if the object's state does not match the detection of a potential fall by a particular fall detection algorithm, then that potential fall can be disregarded as a false alarm because it is inconsistent with the object's initial state (and may be triggered by that fall detection algorithm optimized for different types of falls with different initial states).
[0098] Therefore, an improved technique for fall detection is provided, which can utilize information obtained from sensors in the object's environment to improve the reliability of fall detection, and improve the reliability of different types of fall detection.
[0099] Variations of the disclosed embodiments can be understood and implemented by those skilled in the art during practice of the principles and techniques described herein, through study of the accompanying drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite articles "a" or "an" do not exclude a plurality. A single processor or other unit can perform the functions of several items set forth in the claims. The simple fact that specific measures are mentioned in mutually different dependent claims does not indicate that combinations of these measures cannot be advantageously used. A computer program may be stored or distributed on a suitable medium, such as an optical storage medium or solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems. No reference numerals in the claims should be construed as limiting the scope.
Claims
1. A fall detection device, the fall detection device comprising one or more processing units, the one or more processing units being configured to: Obtain a first input, which indicates which one or more fall detection algorithms among a plurality of fall detection algorithms detects a potential fall of the object, wherein, Each of the multiple fall detection algorithms is associated with a corresponding type of fall, and detects potential falls of the associated type by analyzing a set of movement measurements for the object, wherein each corresponding type of fall has an associated initial state of the object; A second input is obtained, indicating the state of the object prior to the potential fall, wherein the state of the object is determined by analyzing a set of measurements from one or more sensors in the object's environment; The determined state of the object prior to the potential fall is compared with the initial state for each type of fall associated with any potential fall indicated in the first input; and If the determined state of the object matches the initial state of any corresponding type of fall associated with any potential fall indicated in the first input, an indication that the object has fallen is output.
2. The fall detection device according to claim 1, wherein, The initial state of the object associated with a type of fall includes any one or more of the following: (i) a standing posture, (ii) a sitting posture, and (iii) a lying posture.
3. The fall detection device according to claim 1 or 2, wherein, The corresponding types of falls associated with the various fall detection algorithms include any one or more of the following: (i) falling from a standing position, (ii) falling from a sitting position, (iii) falling from a lying position, (iv) falling while moving from a sitting position to a standing position, (v) falling while moving from a standing position to a sitting position, (vi) falling from a standing position onto furniture, and (vii) the object sliding down a wall from a standing position.
4. The fall detection device according to any one of claims 1-3, wherein, The one or more processing units are configured to obtain the first input by: The various fall detection algorithms are used to analyze a set of movement measurements of an object to detect whether the object has experienced a potential fall of the corresponding type associated with each fall detection algorithm; as well as The first input is formed based on the results of the analysis of the set of movement measurement results using the various fall detection algorithms.
5. The fall detection device according to any one of claims 1-3, wherein, The one or more processing units are configured to obtain the first input from a fall detection device carried or worn by the object.
6. The fall detection device according to any one of claims 1-5, wherein, The one or more processing units are configured to obtain the second input by: Analyze a set of measurements from one or more sensors in the environment of the object to determine the state of the object prior to a potential fall; and The second input is formed based on the analysis of the set of measurements from one or more sensors in the environment of the object.
7. The fall detection device according to any one of claims 1-5, wherein, The one or more processing units are configured to obtain the second input from a monitoring system of the one or more sensors in the environment that includes the object.
8. A fall detection device, comprising: One or more motion sensors are used to measure the movement of an object; One or more processing units are configured as follows: Receive a set of motion measurements for the object from the one or more motion sensors; Multiple fall detection algorithms are used to analyze the set of movement measurements to detect whether the object has experienced a potential fall of a corresponding type associated with each fall detection algorithm, wherein each corresponding type of fall has an associated initial state of the object; and The first input is formed based on the analysis of the set of movement measurement results using the various fall detection algorithms; and The fall detection device according to any one of claims 1, 2, 3, 6 or 7.
9. A monitoring system, comprising: One or more processing units are configured as follows: Receive a set of measurements from one or more sensors in the object's environment; Analyze the set of measurements to determine the state of the object prior to a potential fall; and The second input is formed based on the results of the analysis of the set of measurements; as well as The fall detection device according to any one of claims 1-5.
10. A method for detecting a fall, the method comprising: A first input is obtained, indicating which one or more fall detection algorithms among a plurality of fall detection algorithms detect a potential fall of the object, wherein each of the plurality of fall detection algorithms is associated with a corresponding type of fall, and the associated type of potential fall is detected by analyzing a set of movement measurements for the object, wherein each corresponding type of fall has an associated initial state of the object; A second input is obtained, indicating the state of the object prior to the potential fall, wherein the state of the object is determined by analyzing a set of measurements from one or more sensors in the object's environment; The determined state of the object prior to the potential fall is compared with the initial state for each type of fall associated with any potential fall indicated in the first input; and If the determined state of the object matches the initial state of any corresponding type of fall associated with any potential fall indicated in the first input, an indication that the object has fallen is output.
11. The method according to claim 10, wherein, The steps to obtain the first input include: The various fall detection algorithms are used to analyze a set of movement measurements of an object to detect whether the object has experienced a potential fall of the corresponding type associated with each fall detection algorithm; and The first input is formed based on the results of the analysis of the set of movement measurement results using the various fall detection algorithms.
12. The method according to claim 10 or 11, wherein, The step of obtaining the first input includes obtaining the first input from a fall detection device carried or worn by the object.
13. The method according to any one of claims 10-12, wherein, The steps to obtain the second input include: Analyze a set of measurements from one or more sensors in the object's environment to determine the object's state prior to a potential fall; and The second input is formed based on the analysis of the set of measurements from one or more sensors in the environment of the object.
14. The method according to any one of claims 10-12, wherein, The step of obtaining the second input includes obtaining the second input from a monitoring system of one or more sensors in the environment including the object.
15. A computer program product comprising a computer-readable medium, wherein computer-readable code is embedded therein, the computer-readable code being configured to cause the computer or processor, when run by a suitable computer or processor, to perform the method according to any one of claims 10-14.