The present invention relates to a method for determining a heartbeat rate of a user from an acoustic heartbeat signal detected in an ear of the user and a heartbeat rate measuring device for detecting a heartbeat rate of a user from an acoustic heartbeat signal detected in the ear of the user.
According to an embodiment, a method for detecting a heartbeat rate of a user is provided. The heartbeat rate is determined from an acoustic heartbeat signal. The acoustic heartbeat signal is detected in an ear of the user. According to the method, the heartbeat signal is sampled with a predefined sampling rate. Furthermore, a plurality of heartbeat rate values are provided. According to the method, the sampled heartbeat rate signal is correlated with each of the plurality of heartbeat rate values. Then, the heartbeat rate value with the best matching correlation to the sampled heartbeat signal is determined and this heartbeat rate value is assigned as the heartbeat rate of the user.
When a user of a mobile device, for example a mobile telephone, a personal digital assistant, a mobile navigation system, a mobile media player or a mobile computer, is doing exercises or workout, for example jogging, walking, climbing, or biking, the user may listen to music being played back by the mobile device during exercise. Furthermore, the user may be interested in his or her current heartbeat rate to control the training. A mobile device supporting the above-described method of the present invention is adapted to determine the heartbeat rate by detecting the acoustic heartbeat signal in the ear of the user. This can be simply accomplished by a microphone integrated into an in-ear headphone providing the music for the user. According to the above-described embodiment of the present invention, a correlation technique is used to determine the current heartbeat rate of the user and thus the technique is very robust against for example noise or disturbances. Furthermore, the technique can be carried out at a very low sampling rate which results in a low power consumption and a low amount of processing power of the mobile device.
According to an embodiment, the plurality of heartbeat rate values is based on the predefined sampling rate. By choosing the heartbeat rate values based on the predefined sampling rate, a high correlation between the sampled heartbeat rate signal and the thus provided heartbeat rate values can be achieved which makes the technique even more robust.
Furthermore, the detected heartbeat signal may be filtered with a low pass filter. The low pass filter may have a cut off frequency in a range from 4 Hz to 10 Hz, preferably 5 Hz. Furthermore, an absolute value of the detected heartbeat signal may be formed. Filtering and forming an absolute value may remove noise and disturbing signals to make the heartbeat rate determination more robust.
According to another embodiment, correlating the sampled heartbeat signal with each of the plurality of heartbeat rate values comprises for each heartbeat rate value the following: For each heartbeat rate value a plurality of cumulation values are provided. The number of the plurality of cumulation values is defined by the heartbeat rate value and the predefined sampling rate. When a new sample value of the sampled heartbeat signal is sampled this sample value is added to a cumulation value of the plurality of cumulation values. Thereby, consecutive sample values of the sampled heartbeat signal are added to consecutively and cyclicly addressed cumulation values.
Thus, for each heartbeat rate value a list of cumulation values is provided which is addressed consecutively and cyclicly. A new sample value is added to the currently addressed cumulation value of each of the heartbeat rate values. Then, the next cumulation value is addressed for each of the heartbeat rate values. The length of each list, i.e. the number of cumulation values for each list, is different for each heartbeat rate value and defined by the heartbeat rate value and the predefined sampling rate. When a detected heartbeat rate signal correlates with a heartbeat rate value, in the corresponding cumulation value list some of the cumulation values will grow very strong whereas other cumulation values of the same list will grow very slow. When the detected heartbeat signal does not correlate to the heartbeat rate value, in the corresponding list of cumulation values the cumulation values will grow more uniformly. Thus, the cumulation values clearly indicate a correlation between the heartbeat signal and the corresponding heartbeat rate value. Performing the above-described correlation requires only very little computing power and can therefore be performed on a mobile device having restrictions due to processing power and energy.
According to an embodiment, determining the heartbeat rate value with the best matching correlation comprises a determining of a difference value between a minimum cumulation value and a maximum cumulation value of the plurality of cumulation values relating to the heartbeat rate value for each of the heartbeat rate values. Thus, for each of the heartbeat rate values a separate difference value is determined. Then, the maximum difference value of these difference values is determined and the heartbeat rate value with the maximum difference value is defined as the heartbeat rate value which correlates best to the detected heartbeat signal of the user.
According to an embodiment, the number of the plurality of cumulation values for each heartbeat rate value is defined by a ratio of the predefined sampling rate to the heartbeat rate value.
According to another embodiment, a cumulation value is reduced by a predetermined factor before adding a sample value to the cumulation value. The value of the predetermined factor may be in the range of 0.8 to 0.95, preferably 0.9. By reducing the cumulation value before adding a new sample value a “forget” function is realized which allows that the correlation automatically adapts continuously to a varying heartbeat rate of the user.
The sampling rate for sampling the acoustic heartbeat signal may be in the range of 10 to 200 Hz, preferably 22.05 Hz.
According to another embodiment, a heartbeat rate measuring device is provided. The device comprises a microphone adapted to detect an acoustic heartbeat signal of a user in an ear of the user, a sampling unit adapted to sample the heartbeat signal with a predefined sampling rate, and a correlation unit. The correlation unit provides a plurality of heartbeat rate values and is adapted to correlate the sampled heartbeat signal with each of the plurality of heartbeat rate values and to determine the heartbeat rate value with the best matching correlation.
According to an embodiment, the plurality of heartbeat rate values is based on the predefined sampling rate. The predetermined sampling rate may be in the range of 10 to 200 Hz, preferably 22.05 Hz.
The heartbeat measuring device may comprise furthermore a low pass filter adapted to filter the detected heartbeat signal. The low pass filter may have a cut off frequency in a range from 4 Hz to 10 Hz, preferably 5 Hz.
Furthermore, the heartbeat rate measuring device may comprise an absolute value unit adapted to form an absolute value of the detected heartbeat signal.
According to an embodiment, the correlation unit is adapted to provide for each heartbeat rate value a plurality of cumulation values. The number of the plurality of cumulation values is defined by the heartbeat rate value and the predefined sampling rate. The correlation unit may be further adapted to add a sample value of the sampled heartbeat signal to a cumulation value of the plurality of cumulation values for each heartbeat rate value. Consecutive sampling values are added to consecutively and cyclicly addressed cumulation values for each heartbeat rate value.
Furthermore, the correlation unit may be adapted to determine for each heartbeat rate value a difference value defining a difference between a minimum cumulation value and a maximum cumulation value of the plurality of cumulation values relating to the heartbeat rate value, and to define the heartbeat rate value with a maximum difference value as the heartbeat rate value with the best matching correlation.
The number of the plurality of cumulation values for each heartbeat rate value may be defined by a ratio of the predefined sampling rate to the heartbeat rate value.
According to an embodiment, the correlation unit is adapted to reduce the cumulation value by a predetermined factor before adding the sample value to the cumulation value. The predetermined factor may be in the range from 0.8 to 0.95, preferably 0.9.
According to an embodiment, a mobile device is provided. The mobile device comprises a heartbeat rate measuring device. The heartbeat rate measuring device comprises a microphone adapted to detect an acoustic heartbeat signal of a user in an ear of the user, a sampling unit adapted to sample the heartbeat signal with a predefined sampling rate, and a correlation unit providing a plurality of heartbeat rate values. The correlation unit is adapted to correlate the sampled heartbeat signal with each of the plurality of heartbeat rate values, and to determine the heartbeat rate value with the best matching correlation.
The mobile device may comprise a mobile phone, personal digital assistant, a mobile navigation system, a mobile media player, or a mobile computer.
Although specific features described in the above summary and the following detailed description are described in connection with specific embodiments, it is to be understood that the features of the embodiments described can be combined with each other unless it is noted otherwise.
Hereinafter, exemplary embodiments of the invention will be described with reference to the drawings.
In the following, exemplary embodiments of the present invention will be described in detail. It is to be understood that the following description is given only for the purpose of illustrating the principles of the invention and is not to be taken in a limiting sense. Rather, the scope of the invention is defined only by the appended claims and not intended to be limited by the exemplary embodiments hereinafter.
It is to be understood that the features of the various exemplary embodiments described herein may be combined with each other unless specifically noted otherwise.
As described in connection with
Therefore, the heartbeat rate detection unit 207 performs a detection based on an energy detection of the audio waveform. This allows the signal to be greatly downsampled as the pulse beat resides only at very low frequencies, for example 5 Hz and below. This reduces the amount of signal processing and thus offers low power implementations.
For a robust detection the downsampled audio signal is correlated against a series of impulse streams which differ in period (interval) and in time offset.
Whenever a new downsampled value of the heartbeat signal is provided by the downsampling unit 303 to the correlation unit 304 (
W(mod(k,Ti),i)=αW(mod(k,Ti),i)+(1−α)xk
wherein
W(r, c) is the matrix element of row r with r in the range from 0 to Ti−1 and column c with a range of 1 to 5 in the example of
mod(a, b) is the modulo operation a mod b,
k is a counter counting the input samples starting at 0,
Ti is the period of the considered column with i in the range of 1 to 5 in the example of
α is a predetermined factor, and
xk is the value of sample k of the downsampled heartbeat signal.
Thus, each column of the matrix of
The modification or update process defined in the above-described equation uses an exponential forget function with the parameter α to average between new and past samples. For a large α more averaging will be carried out and correlation over more energy bursts will be applied. However, the system will have a slower response to a change in the heartbeat rate. α may be in the range of 0.8 to 0.95, preferably 0.9.
When modifying the matrix values, one column of
Operation of the heartbeat rate measuring device as shown in connection with
Finally,
The above-described algorithm is very robust and operates even when noise and disturbances are added to the acoustic heartbeat signal. The heartbeat rate is based on a number of consecutive energy bursts and gaps between the energy bursts. For further noise suppression, filtering is applied by an exponential “forget” function characterized by an “forget” parameter α. For larger α the method will be more robust at the expense of a longer latency and a slower response speed.
As in the above-described method very robust correlation techniques are used, the described method is very robust against disturbances and noise. Furthermore, the described techniques are carried out at a very low sampling rate which results in a low power consumption. Furthermore, the amount of processing power is very low. Therefore, the above-described method and the above-described heartbeat rate measuring device can be advantageously integrated into a mobile device, for example a mobile phone or a mobile reproduction device.
While exemplary embodiments have been described above, various modifications may be implemented in other embodiments. For example, the sampling rate after the downsampling unit 303 may be adapted to meet the requirements of a desired range and resolution for the heartbeat rate detection. Furthermore, the number of columns and rows in the matrix of
Finally, it is to be understood that all the embodiments described above are considered to be comprised by the present invention as it is defined by the appended claims.