Patent Application
20140330535
The invention relates to a method and apparatus that can provide an indication of the location of a user of the apparatus, and in particular relates to a method and apparatus that can detect motion of the user in an elevator and thus provide an indication of the elevation of the user.
Satellite positioning systems, such as GPS, provide an indication of the location of the user using signals received from a number of satellites. Although these types of systems work well in outdoor environments, they do not work very well, or at all, in indoor environments since it can be difficult to receive the signals from the satellites.
Therefore, attempts have been made to provide indoor tracking and navigation systems. Some systems use motion detection systems (for example requiring infrastructure to be installed in the indoor environment and/or through the use of a user-worn device that contains motion sensors), that can optionally make use of maps of the indoor environment or other context information.
Accelerometers can be used in these user-worn devices that measure acceleration, including gravitational fields, to which they are subjected. Tri-axial (3D) accelerometers have this capability irrespective of their own orientation, and can provide a relative direction of the acceleration along with the magnitude of the acceleration. This makes them useful for a wide variety of motion related applications.
One important aspect of indoor tracking and navigation systems is determining which floor of the building the user is on. However, there are currently no reliable solutions for determining the floor that a user is on based on accelerometer measurements.
Therefore, there is a need for an improved method and apparatus for detecting the motion of the user in an elevator and for determining the elevation of the user from that motion, for example in terms of the floor or level of a building that the user is on.
It has been found that elevators of different types exhibit very typical motion patterns that can be observed using just an accelerometer. A typical journey in an elevator starts with an upward or downward acceleration from a static situation (i.e. zero acceleration apart from gravity), followed by a (usually) steady upward or downward motion, and then a deceleration until the elevator returns to a static situation. The amount of upward or downward displacement of the elevator relates to the total of acceleration, steady motion and deceleration. Moreover, the direction of an elevator displacement (i.e. either up or down) can be sensed by an accelerometer since the elevator acceleration will either add to or subtract from the gravity signal (which is always in the upward direction).
Therefore, according to a first aspect of the invention, there is provided a method for detecting the motion of a user or object in an elevator, the method comprising measuring the acceleration experienced by the user or object to obtain a series of acceleration measurements; processing the series of acceleration measurements to identify a peak and a trough therein that are associated with the start and end of an elevator motion; identifying a section of the acceleration measurements corresponding to the elevator motion from the identified peak and trough; and determining an indication of the change in elevation of the user or object during the elevator motion from the identified section of the acceleration measurements.
In some embodiments, the method further comprising the steps of determining a noise level signal from the series of acceleration measurements, the noise level signal indicating the level of noise in each of the acceleration measurements; and comparing the noise level signal to a noise threshold; and the step of processing the acceleration measurements comprises processing only those acceleration measurements where the corresponding part of the noise level signal is less than the noise threshold.
In other embodiments, the step of processing comprises comparing the amplitude of each acceleration measurement to a peak threshold and a trough threshold, the peak threshold being higher than the trough threshold; identifying a candidate peak in the acceleration measurements as a group of consecutive acceleration measurements whose amplitudes exceed the peak threshold and identifying a candidate trough in the acceleration measurements as a group of consecutive acceleration measurements whose amplitudes are below the trough threshold.
In those embodiments, the step of processing preferably further comprises discarding any identified candidate peak or identified candidate trough where the duration of the respective group of acceleration measurements is less than a minimum time threshold.
In further embodiments, the method further comprises the steps of determining a noise level signal from the series of acceleration measurements, the noise level signal indicating the level of noise in each of the acceleration measurements; and comparing the noise level signal to a noise threshold; and wherein the step of processing the acceleration measurements further comprises discarding any identified candidate peak and identified candidate trough where the corresponding part of the noise level signal exceeds the noise threshold.
In yet further embodiments, the step of processing further comprises evaluating the candidate peaks and candidate troughs to identify an elevator motion, with an elevator motion being identified by a candidate peak followed by a candidate trough or a candidate trough followed by a candidate peak.
Alternatively, the step of processing further comprises evaluating the candidate peaks and candidate troughs to identify an elevator motion, with an elevator motion being identified as a candidate peak followed by a candidate trough within a predetermined time period or as a candidate trough followed by a candidate peak within a predetermined time period.
In some embodiments, the step of identifying a section of the acceleration measurements corresponding to the elevator motion from the identified peak and trough comprises determining the start time of the elevator motion by searching the acceleration measurements before the earliest one of the identified peak and trough until an acceleration measurement is found where the acceleration is constant and/or the acceleration substantially corresponds to acceleration due to gravity; and determining the end time of the elevator motion by searching the acceleration measurements after the latest one of the identified peak and trough until an acceleration measurement is found where the acceleration is constant and/or the acceleration substantially corresponds to acceleration due to gravity.
In some embodiments, the step of determining an indication of the change in elevation of the user or object during the elevator motion from the identified section of the acceleration measurements comprises double integrating the acceleration measurements between the start time and the end time of the elevator motion.
In further embodiments, the step of processing further comprises applying a filter to the obtained acceleration measurements to suppress high-frequency noise in the measurements prior to identifying peaks and troughs.
In some embodiments, the method further comprises the step of using the indication of the change in elevation of the user or object during the elevator motion to update a previously-determined value for the height of the user or object relative to sea level, a previously-determined value for the height of the user or object relative to the ground or a previously-determined value for the floor of a building or other structure that the user or object is on.
According to a second aspect of the invention, there is provided a computer program product comprising computer-readable code embodied therein, the computer readable code being configured such that, upon execution by a suitable computer or processor, the computer or processor performs the method as described above.
According to a third aspect of the invention, there is provided an apparatus for detecting motion of a user or object in an elevator, the apparatus comprising a processor that is configured to receive measurements of the acceleration experienced by the user or object, process the acceleration measurements to identify a peak and a trough therein that are associated with the start and end of an elevator motion, identify a section of the acceleration measurements corresponding to the elevator motion from the identified peak and trough, and determine an indication of the change in elevation of the user or object during the elevator motion from the identified section of the acceleration measurements.
In some embodiments, the processor is further configured to determine a noise level signal from the acceleration measurements, the noise level signal indicating the level of noise in each of the acceleration measurements; and compare the noise level signal to a noise threshold; and wherein the processor is configured to identify a peak and a trough in the acceleration measurements by processing only those acceleration measurements where the corresponding part of the noise level signal is less than the noise threshold.
In other embodiments, the processor is configured to identify a peak and a trough in the acceleration measurements by comparing the amplitude of each acceleration measurement to a peak threshold and a trough threshold, the peak threshold being higher than the trough threshold; and identifying a candidate peak in the acceleration measurements as a group of consecutive acceleration measurements whose amplitudes exceed the peak threshold and identifying a candidate trough in the acceleration measurements as a group of consecutive acceleration measurements whose amplitudes are below the trough threshold.
In those embodiments, the processor is further configured to identify a peak and a trough in the acceleration measurements by discarding any identified candidate peak or identified candidate trough where the duration of the respective group of acceleration measurements is less than a minimum time threshold.
In further embodiments, the processor is further configured to determine a noise level signal from the acceleration measurements, the noise level signal indicating the level of noise in each of the acceleration measurements; and compare the noise level signal to a noise threshold; and wherein the processor is configured to identify a peak and a trough in the acceleration measurements by discarding any identified candidate peak and identified candidate trough where the corresponding part of the noise level signal exceeds the noise threshold.
In yet further embodiments, the processor is further configured to evaluate the candidate peaks and candidate troughs to identify an elevator motion, with an elevator motion being identified by a candidate peak followed by a candidate trough or a candidate trough followed by a candidate peak.
In other embodiments, the processor is further configured to evaluate the candidate peaks and candidate troughs to identify an elevator motion, with an elevator motion being identified as a candidate peak followed by a candidate trough within a predetermined time period or as a candidate trough followed by a candidate peak within a predetermined time period.
In some embodiments, the processor is configured to identify a section of the acceleration measurements corresponding to the elevator motion from the identified peak and trough by determining the start time of the elevator motion by searching the acceleration measurements before the earliest one of the identified peak and trough until an acceleration measurement is found where the acceleration is constant and/or the acceleration substantially corresponds to acceleration due to gravity; and determining the end time of the elevator motion by searching the acceleration measurements after the latest one of the identified peak and trough until an acceleration measurement is found where the acceleration is constant and/or the acceleration substantially corresponds to acceleration due to gravity.
In some embodiments, the processor is configured to determine an indication of the change in elevation of the user during the elevator motion from the identified section of the acceleration measurements by double integrating the acceleration measurements between the start time and the end time of the elevator motion.
In some embodiments, the processor is further configured to apply a filter to the obtained acceleration measurements to suppress high-frequency noise in the measurements prior to identifying peaks and troughs.
In some embodiments, the processor is further configured to use the indication of the change in elevation of the user or object during the elevator motion to update a previously-determined value for the height of the user or object relative to sea level, a previously-determined value for the height of the user or object relative to the ground or a previously-determined value for the floor of a building or other structure that the user or object is on.
In yet further embodiments, the apparatus further comprises an accelerometer that is configured to be attached to a user or object.
Embodiments of the invention will now be described, by way of example only, with reference to, and as shown in, the following drawings, in which:
The indication of the elevation determined by the processor 6 can be provided in a number of different forms. In particular, the indication of the elevation of the apparatus 2 can indicate the height of the apparatus 2 relative to the ground, the height of the apparatus 2 relative to sea level (more commonly referred to as altitude) and/or the height of the apparatus 2 relative to a previously determined elevation. Preferably, however, the indication of the elevation of the apparatus 2 indicates the floor or level of the building that the apparatus 2 is on (for example, ground floor, first floor, etc.).
The accelerometer 4 is preferably a tri-axial accelerometer that provides measurements of the acceleration in three-dimensions. In this case, the signal output by the accelerometer 4 to the processor 6 can comprise a respective signal for each of the measurement axes of the accelerometer 4. In some embodiments, the accelerometer 4 is a micro-electromechanical system (MEMS) accelerometer, although it will be appreciated that other types of accelerometer 4 can be used. The accelerometer 4 can typically operate at a sampling frequency of 30 Hz or 50 Hz, although it will be appreciated that other sampling frequencies can be used.
Although not illustrated in
Exemplary steps in the processing performed by the processor 6 to determine the elevation indication are shown in
In the pre-processing step 101, the signal from the accelerometer 4 undergoes pre-processing in order to remove, or at least reduce, high-frequency noise from the acceleration measurements.
The signal output from the accelerometer 4 may have a different scale than the one needed in subsequent processing steps (for example the integration step 1051). For example, the scale may be in units of some analog-to-digital converter. Therefore, in step 1011, the scale is normalized. To normalize the scale (preferably to units of m/s2), the original accelerometer signal values are divided by a constant, the value of which depends on the ‘before’ and ‘after’ scales involved. Furthermore, this step can require the subtraction or addition of a constant in order to normalize the zero-level of the signal. Techniques for implementing this scale normalization will be familiar to those skilled in the art.
It should be noted that the amplitude scale used in
After scale normalization, the scale normalized accelerometer signal is normalized to calculate the vector length (step 1013). The vector length S can be calculated using:
S=√(X2+Y2+Z2)
where X, Y and Z are the components of acceleration along the x-, y- and z-axes of the accelerometer 4 at a particular sampling instant.
After the length (magnitude) of the acceleration vector at each sampling instant has been determined in step 1013, a filter is applied to the time-series of acceleration vector lengths to remove or reduce noise, in particular high-frequency noise (step 1015). The filter applied is preferably a 1-second smoothing filter, although those skilled in the art will be aware of other suitable types of filter that can be used to remove the high-frequency noise components. The high-frequency noise, can comprise frequencies of above around 5 Hz, although other threshold frequencies can be used.
In addition to outputting the filtered acceleration signal, step 1015 also preferably outputs a signal indicating the level of the high-frequency noise, as this signal can be used in the peak detection and classification step 103 to reject noise-related peaks.
The noise level signal can be derived by first subtracting the filtered acceleration signal from the acceleration length signal output by step 1013 to obtain a noise signal. Then, the noise level signal can be derived by either (i) rectifying the noise signal (i.e. taking absolute values) and then performing low-pass filtering, (ii) rectifying the noise signal and for each sample in the noise signal taking the sum of the surrounding samples or (iii) for each sample in the noise signal, calculating the variance in the group of surrounding samples.
The low-pass filtering in option (i) can be performed using the same or similar filter as that used to remove the high-frequency noise from the acceleration length signal output by step 1013 (so for example a 1-second smoothing filter), although those skilled in the art will appreciate that other types of filter can be used. For the selection of surrounding samples in options (ii) and (iii), a window size of 1 second is suitable. Furthermore, instead of variance, other measures of spread can be used, like standard deviation or inter-quartile range. In addition, in principle any operation that results in a measure of the ‘energy’ of a signal can be used to determine the noise level signal.
After pre-processing, the peaks and troughs in the filtered accelerometer signal that are associated with the motion of an elevator are detected and classified in step 103. As can be seen in
For example, when the elevator is to ascend, the elevator accelerates upwards from rest to a constant velocity. This upwards acceleration is represented in the signal in
Likewise, when the elevator is to descend, the elevator accelerates downwards from rest to its preset velocity. This downwards acceleration is represented in the signal in
Therefore, peak detection and classification is performed in step 103 to detect these successive accelerations and decelerations. In particular, the peak detection and classification aims to detect consecutive peaks and troughs that represent motion of an elevator, with a peak followed by a trough representing an ascent, and a trough followed by a peak representing a descent. Step 103 can achieve this using a number of features typically associated with elevator motion, including duration and amplitude.
In a first step of the peak detection and classification, step 1031, the filtered accelerometer signal is analysed to detect candidate peaks and troughs.
In a preferred embodiment, detection of candidate positive peaks is performed as follows. Firstly, the filtered accelerometer signal is scanned for samples whose amplitudes exceed an amplitude threshold value. The amplitude threshold value for detecting candidate positive peaks will be set to a value that is higher than the acceleration due to gravity (i.e. referring to
The period of time covered by each group of consecutive samples whose magnitudes are above the amplitude threshold value is compared to a minimum time period. Exemplary values for the minimum time period can include 0.5 s, is or 1.5 s, although it will be appreciated that other values can be used. A candidate positive peak is identified as any group of consecutive samples whose magnitudes are above the amplitude threshold value and that cover a period of time greater than the minimum time period. The sample with the largest amplitude within that candidate peak is marked as the central peak time. The use of a minimum time period allows, for example, impacts (which appear as relatively sharp peaks in the accelerometer signal) to be excluded from the subsequent processing steps.
Although exemplary values have been provided for the amplitude threshold value and minimum time period, it will be appreciated that the values for these two parameters should generally be set in combination. So, where a lower amplitude threshold value is used, a longer time period might be selected, and vice versa. It will also be appreciated that optimum values for the amplitude threshold value and minimum time period will depend on the specific type of elevator that the user is in. Optimum values for various types of elevator can be predetermined and stored in a database for use by the apparatus 2 when the apparatus 2 is in a building that contains a known type of elevator (for example with the building being determined on the basis of the last available satellite positioning system measurement).
Although not essential, it is possible to apply a maximum time period to identify the candidate positive peaks, which means that any peak that has a duration longer than the maximum time period will be excluded from the subsequent processing steps. However, it will be appreciated that ‘long-duration peaks’ are unlikely to occur in the accelerometer signal. Nevertheless, an exemplary value for the maximum time period is 10 s, although it will be appreciated that other values can be used.
The effect of this processing is illustrated in
It will be appreciated that the above peak detection can be performed in substantially real time on the filtered accelerometer signal, in which case the signal can be scanned and as soon as a sample is identified whose amplitude exceeds the amplitude threshold value, a timer is started. The scanning of the signal continues and the timer runs until a sample is found that has an amplitude value below the amplitude threshold value. The time that has elapsed between the ‘threshold crossings’ is then compared to the minimum time period for peak duration. If the time difference is sufficiently large (i.e. the elapsed time exceeds the minimum time period), a candidate peak is detected and the sample with the largest value within that time range is marked as the central peak time.
It will be noted that candidate troughs (negative peaks) in the accelerometer signal can be detected in a similar fashion by using a negative peak/trough amplitude threshold value that is set below the value of acceleration due to gravity (for example 1% or 2% below the value for acceleration due to gravity), and identifying candidate negative peaks/troughs as those groups of consecutive samples whose amplitudes are below that threshold value and that cover a time period greater than the minimum time period. The detection of candidate peaks and troughs is preferably performed in parallel.
In the next step of the peak detection and classification, step 1033, the candidate peaks and troughs identified in step 1031 are evaluated to determine if any result from noise in the accelerometer signal. Any candidate peak or trough determined to result from noise in the accelerometer signal is discarded and not considered in subsequent processing steps. In particular, step 1033 takes the output of step 1031 and uses the noise level signal from the filtering step (step 1015) to determine which, if any, of the detected peaks and troughs have been induced into the signal by high frequency noise.
As filtering step 1015 is basically a moving average operation, high-frequency noise can be eliminated provided that it does not affect the average signal level. Thus, filtering step 1015 is reasonably effective where the noise components are of short duration and low amplitude (so they have limited effect on the average) and are evenly spread into positive and negative deviations (so they cancel out). However, it has been found that prolonged, high-energy, high-frequency noise can effectively induce a low-frequency distortion into the accelerometer signal. Such noise might be caused by, for example, the user of the apparatus 2 jumping up and down.
So, if the noise level signal (i.e. representing the energy of the noise) determined in step 1015 is high, then this can indicate that the noise has affected the accelerometer signal even into lower frequencies, and an associated candidate peak or trough in the signal may relate to noise rather than accelerations that have actually been caused by an elevator.
Therefore, in step 1033, for each of the candidate peaks and troughs (preferably for each sample or samples in a detected peak or trough), the noise level signal is evaluated to determine whether the noise level is high for any of the samples forming that peak or trough. This evaluation can comprise comparing the noise level signal to a threshold value, and if the noise level signal exceeds the threshold value at the same time that a peak or trough is detected in step 1031, that peak or trough can be discarded as being ‘noise-induced’. The noise level threshold value depends on the details of the filtering and the noise level signal calculation (step 1015). Where the noise level signal is determined using option (i)—rectifying the noise signal (i.e. taking absolute values) and then performing low-pass filtering—the noise level threshold can be set to 0.8 ms−2, although it will be appreciated that other values can be used.
After rejection of any noise-induced peaks or troughs in step 1033, the remaining candidate peaks and troughs are evaluated to determine which correspond to a single elevator displacement (step 1035). As indicated above, an elevator motion comprises a sequential pair of oppositely signed peaks (either a negative peak/trough followed by a positive peak for a descent, or the other way round for an ascent).
To qualify as a valid pair, an upper limit is placed on the permitted time duration between consecutive peaks and troughs. The upper limit can be set according to the maximum expected elevator displacement duration. If no subsequent positive/negative peak is detected within the maximum permitted time duration from an initial negative/positive peak, then it can be assumed that the initial peak was not as a result of elevator motion. The analysis can then be repeated for the next detected peak or trough.
Once a valid peak and trough pair has been identified, the duration of the elevator motion is determined. This can be achieved by determining the start time of the motion and the end time of the motion (it will be appreciated that the ‘time’ can be indicated by the appropriate samples in the signal from the accelerometer 4). Using the central peak times of the peak and trough, the starting time can be determined by searching backwards from the earliest central peak time and the end time can be determined by searching forwards from the latest central peak time until either, or both, (i) the signal no longer increases or decreases in amplitude (whichever is applicable) or (ii) the signal reaches a value that is sufficiently close (i.e. within a threshold amount) to the magnitude of acceleration due to gravity.
The output of step 1035 comprises sections of the filtered accelerometer signal corresponding to detected pairs of positive and negative peaks, as defined by the determined start and end times of the elevator motion.
An exemplary result of the processing in step 103 is shown in
The sections identified in step 103 are passed to step 105 in which the displacement due to detected elevator motion is determined.
Firstly, in step 1051, a double integration with respect to time is performed on each accelerometer signal section to determine the vertical distance travelled during that particular motion. The start and end times of the section (i.e. the start time of the first peak/trough and the end time of the second trough/peak) provide the lower and upper bounds for the integration respectively.
Optionally, prior to step 1051, the ‘flat’ parts of each accelerometer signal section can be smoothed to improve the accuracy of the double integration. It is also possible to carry out a (further) signal conditioning step between the two integration operations.
Then, optionally, the determined vertical distance can be converted into a measure of the number of floors traversed up or down during the motion by dividing the determined vertical distance by the or an average floor height (step 1053), and rounding the result to the nearest integer. The average floor height can be predetermined, and stored in the memory module 8. An exemplary value for the average floor height is 4 metres, but it will be appreciated that this value can vary significantly depending on the building, building type (e.g. residential or commercial), the prevailing building regulations, etc. In some embodiments, average floor height values for different building or structure types can be stored in the memory module 8, and an appropriate value selected by the apparatus 2 depending on the building or structure that the user is in (which could be determined based on the most recent measurement obtained by the satellite positioning system.
In step 1055, the number of floors traversed during a particular motion is added (or subtracted, as appropriate) from a previous indication of the floor level of the user to give the current floor level. So, for example if it had previously been determined that the user was on the second floor and the apparatus 2 determines that the user has gone up three floors in an elevator, step 1055 will output the user's position as being on the fifth floor.
In some implementations of the invention, the apparatus 2 and associated method can be implemented in a device, such as a mobile telephone, smart phone or PDA, that typically already include an accelerometer and the required processor.
It will also be appreciated that the apparatus 2 can be implemented in any system where it is useful to determine the elevation of a user in a building or other indoor environment or structure, such as a system for tracking the location of particular patients prone to wandering in a hospital (in which case the elevation determined according to the invention can be communicated to a monitoring station to enable a nurse or other healthcare professional to locate the patient and intervene to return the patient to the correct location), or a system for providing general indoor location tracking that complements an outdoor location tracking system, such as GPS. Furthermore, it will be appreciated that the apparatus 2 can be used to provide an indication of the elevation of an object (including an elevator itself) associated with the apparatus 2, rather than a user.
In the case where the apparatus 2 is used to track patients in a hospital (and in similar implementations), the apparatus 2 can be adapted to be attached to a part of the body of the user, such as the waist, trunk, thorax, pelvis or sternum, and can comprise a suitable arrangement (for example a belt or strap) for attaching the apparatus 2 to that part of the body.
Finally, it will be appreciated that in some embodiments of the invention, the accelerometer 4 can be provided in a separate device to the device that comprises the processor 6. In this case, the accelerometer 4 can be attached to the user or object being tracked, and the output of the accelerometer 4 can be transmitted to the processor 6 (using either a wired or wireless connection) for subsequent processing.
There is therefore provided an improved method and apparatus for detecting the motion of the user in an elevator and for determining the elevation of the user from that motion, for example in terms of the floor or level of a building that the user is on.
While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments.
Variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure and the appended claims. In the claims, the word “comprising” does not exclude other elements or steps, and the indefinite article “a” or “an” does not exclude a plurality. A single processor or other unit may fulfil the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a 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 telecommunication systems. Any reference signs in the claims should not be construed as limiting the scope.
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/IB2012/056967 | 12/5/2012 | WO | 00 | 5/20/2014 |
| Number | Date | Country | |
|---|---|---|---|
| 61567714 | Dec 2011 | US |
