• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The present invention relates to a detection method for an electronic apparatus comprising a housing containing an electronic module supplied with electrical energy by an electrical energy storage means, said electronic module comprising a computing unit connected to an acceleration sensor and to a memory unit, said electronic module also being connected to a display means so as to display information related to data of said acceleration sensor. The method is particularly intended to evaluate the movements made during a volleyball match.
    公开号:CH711373A2
    申请号:CH00903/16
    申请日:2016-07-14
    公开日:2017-01-31
    发明作者:Stolpe Alexander；Streicher Markus；Taraben Adib；Eckelmann Christoph
    申请人:Swatch Ag；
    IPC主号:
    专利说明:

    FIELD OF THE INVENTION
    The present invention relates to a detection method for an electronic device comprising a housing containing an electronic module supplied with electrical energy by an electrical energy storage means, said electronic module comprising a calculation unit connected to a sensor of acceleration and to a memory unit, said electronic module also being connected to a display means so as to display information related to data of said acceleration sensor.
    PRIOR ART
    A portable electronic object, such as an electronic device or, for example, a portable timepiece, comprises a housing generally consisting of a middle part closed by a bottom and a glass. An electronic module is disposed within the housing to provide information. This information could be information related to an activity. These activities are mainly related to fitness such as running or walking.
    To this end, the electronic module comprises at least one sensor, an acceleration sensor for example. The electronic circuit also includes a computing unit for using data from the acceleration sensor and a memory for recording them.
    An algorithm is implemented in the computing unit. Said algorithm is programmed to process said data so as to obtain data representative of an activity.
    [0005] Currently, however, it is difficult to have an algorithm capable of accurately determining the activity. For example, in the case of a pedometer, it is very difficult for the computing unit and the algorithm to determine whether the user is walking or running. However, the distinction between a walking activity and a running activity is important to calculate the distance traveled.
    Therefore, if it is difficult to determine the action for simple activities such as running or walking, it is more difficult to provide a portable object capable of detecting complex actions such as volleyball.
    Indeed, in volleyball, we can distinguish different categories of volleyball techniques. In addition to this, the force of impact on the ball, hereinafter referred to as "smash power" is determined.
    BRIEF SUMMARY OF THE INVENTION
    An object of the present invention is to provide an algorithm, used for example in a volleyball activity, capable of detecting and characterizing the types of contact with the ball in a beach volleyball match such as services, passes, dives, etc.
    The invention therefore relates to a method of managing an electronic device comprising a housing containing an electronic module supplied with electrical energy by a means for storing electrical energy, said electronic module comprising a calculation unit connected to a sensor. a memory unit, said electronic module also being connected to a display means so as to display information related to data of said acceleration sensor, the method comprising the steps of:acquire data from the acceleration sensor;detecting potential activities from data acquired using a high-pass filter to detect rapid changes in the measured acceleration, and using a low-pass filter to filter, in the data transmitted by the high-pass filter, the standard of transmitted data to detect only significant activities;filtering above a second threshold, in the data transmitted by the low-pass filter, the maxima of the data standard having passed the high-pass filter, to reject a larger part of the irrelevant activities;Prepare the data for the ranking by filtering the acceleration data, by low-pass filtering, before and after the peak / impact, progressing towards the impact from both directions, and sub-sampling them starting with the value la closer to the impact, the subsampled data passing to the classification stage in which the algorithm is trying to determine which activity has occurred;characterize the technique on the basis of a nearest neighbor algorithm using the L1 standard.
    According to a first advantageous embodiment, the method further comprises a step of determining the power of blow of using a bandpass filter to filter the acceleration data in a certain window before and after the suddenly, the sum on the norm of the past band filtered data being correlated with the velocity of the ball.
    According to a second advantageous embodiment, the computing unit is able to switch from a low power mode in which said computing unit is dormant, to a normal mode in which said computing unit s' active if the acceleration sensor, which is constantly acquiring data with an approximately appropriate frequency, measures an acceleration that exceeds a determined threshold.
    According to a third advantageous embodiment, the low-pass filter and the high-pass filter use an infinite impulse response filter (IIR filter) of second order.
    BRIEF DESCRIPTION OF THE FIGURES
    Other features and advantages of the present invention will appear on reading the description which follows, given solely by way of non-limiting example, and with reference to the accompanying drawings in which:<tb> figs. 1 and 2 <SEP> are functional diagrams that illustrate the management method according to the invention;<tb> fig. 3 <SEP> is a functional diagram that illustrates a characterization step according to the invention.
    DETAILED DESCRIPTION
    The portable object according to the present invention is a wristwatch contained in a housing. The housing forms a housing in which an electronic circuit is placed. Said electronic circuit comprises a calculation unit or processing unit for the management of the electronic circuit, a sensor for acquiring data and a memory for recording these data. The portable object also includes display means such as an LCD screen for displaying the data.
    This portable object can be used during a sports activity such as volleyball. In volleyball, there are many different types of volleyball techniques that can be distinguished, such as diving ("dig" in English), diving with a jump, the taut pass, the lobed pass, the attack, the attack in suspension, and so on. In addition, the force of impact on the ball, hereinafter referred to as "smash power", is determined. A common feature of volleyball techniques is that the ball normally touches the forearm in the immediate vicinity of the wrist. This provides the opportunity to acquire acceleration data using a sensor associated with a wristwatch to determine and characterize the techniques used.
    The sensor is a 3D acceleration sensor with a measurement range between + 16g and -16g, and it is attached to the wrist so that its axes are aligned with the forearm. Parallel to the bones of the forearm, perpendicular to the palm of the hand and horizontal when the arm is turned down, which is useful because most of the movements of the arms during a volleyball match ball are causing distinct accelerations along these axes. The acceleration sensor is configured to acquire data with a frequency between 30 Hz and 50 Hz, preferably between 35 Hz and 45 Hz, most preferably about 40 Hz in the low power mode.
    The computing unit of the portable object further comprises an algorithm capable of processing said acceleration data so as to obtain data representative of an activity. In the present description, the activity is volleyball. The volleyball action detection algorithm is based on a multi-level approach as seen in figs. 1 and 2 . An activity is defined by a change in the acceleration data above a certain threshold. In the watch, an activity interruption is used so that when the acceleration sensor detects an activity, it is notified to the operating system and the acceleration data is retrieved. If the sensor does not detect activity and therefore does not send an operating interrupt to the operating system, the operating system never retrieves the acceleration data and therefore never starts the algorithm .
    In the first step of the detection algorithm, the acceleration sensor is set to an alarm threshold mode, it continuously acquires data in low power mode at about 40 Hz, stores them in a circular buffer, and checks whether an acceleration greater than a determined threshold is measured. This threshold should be between 1 g and 3 g, preferably between 1.5 g and 2.5 g and most preferably be about 2 g. In the case where an acceleration greater than the threshold is measured, the microcontroller that executes the algorithm wakes up and the acceleration data contained in the circular buffer are transferred to the microcontroller. To explain more precisely, the acceleration sensor acquires data at a certain frequency. In the present case 40 Hz. The acceleration data is first stored in the acceleration sensor itself until the operating system retrieves the acceleration data into pieces. By generating a "watermark" interrupt, the operating system tells the acceleration sensor how many groups of data the acceleration sensor must store until it notifies the operating system. The operating system then retrieves the acceleration data from the sensor, and accordingly moves the data from the sensor memory to the operating system. The acceleration sensor then fills its own memory again with the new acceleration data and again notifies the operating system when 20 sets of acceleration data are stored. The limit 20 of our case is called "watermark" level. Whenever the watermark level is reached, the acceleration sensor sends a system interrupt to the operating system, the operating system then retrieving the sensor data.
    The second step of the detection algorithm is to detect potential volleyball activities. To do this, the acceleration data is high-pass filtered with an efficient second-order infinite impulse response (IIR) filter. The cutoff frequency must be between 5 Hz and 15 Hz, preferably between 7 Hz and 13 Hz and most preferably be about 9 Hz. This is done to detect rapid changes in measured acceleration.
    To detect only important activities, namely a contact of the ball with the arm, the norm (sum squared xyz axes) filtered data high-pass is calculated and transmitted to a low-pass filter (again a times with a second-order IIR filter). The cutoff frequency should be between 1 Hz and 10 Hz, preferably between 2 Hz and 6 Hz and most preferably be about 3 Hz.
    The occurrence of a volleyball technique is probably reflected by peaks of this standard above a second threshold. This second threshold should be between 0.1 g and 1 g, preferably between 0.2 g and 0.6 g, most preferably about 0.3 g. At this level, about 55% to 70% of the maxima detected were irrelevant activities, and the remaining 30% to 45% were interesting volleyball activities. The response time of the IIR filters is taken into account for the raw acceleration data for all subsequent operations.
    In a third step, maxima, above the second threshold, of the norm of the data having passed the high-pass filter, filtered low-pass, are analyzed in another filter to reject a larger part of the activities. irrelevant.
    The analysis of the data showed that most of the irrelevant maxima of the filter described in the second step had a relatively low value of the high pass maxima and a low value for the standard of the raw acceleration data just before the peak (on average between 5 and 10 values of accelerations before the peak), while most of the relevant volleyball activities are characterized by a high peak in the high-pass filtered data and / or a high value of 'raw acceleration.
    This filter is implemented by dividing the 2D space (peak value of the high-pass filtered standard and raw acceleration standard before the peak) by a line, the peaks above the line undergoing a further processing and the peaks below the line (relatively low filtered low pass peaks and low raw acceleration standard) are rejected. This reduces the ratio of non-relevant activities to the relevant volleyball (ball-hit, ie hitting) activities to about 1: 1. There are about 25 to 35 acceleration measurements raw (for each acceleration sensor axis) before and there are approximately 15 to 25 after the high-pass filtered peak (ie likely contact with the ball). The data is then transferred to the next function.
    In the fourth step of the algorithm, the data are prepared for classification. The raw 3D acceleration data before and after the peak / hit (excluding peak acceleration data) are low-pass filtered from both ends to the peak, which means that in a centric data interval on the peak, the filtering is done from the data at the end of the interval to the data at the peak. The 3D raw acceleration data is also downsampled by a factor between 2 and 4, starting with the value closest to the keystroke (typing at the ends). The low-pass filter may be the same IIR filter as the filter described and used in the second step.
    The downsampled data is then passed to the classification stage of an algorithm that attempts to determine what type of activity and preference of volleyball activity has occurred (also eliminating non-competitive activities). relevant).
    In a fifth step, the characterization of the volleyball technique is performed as seen in FIG. 3. Said characterization of the volleyball technique is based on a nearest-neighbor algorithm using the L1 standard in the subsampled acceleration data space. The L1 standard is calculated in the direction of cluster centers, each of which is uniquely associated with a volleyball technique (ie jumping attacking, serving low, etc.) or several other common activities that occur during a normal volleyball game (applause, checks with another or fan activities like applauding or doing the ola). Each volleyball technique is tied to a memory location. So, whenever the volleyball technique occurs, the number in the memory slot is incerated. So, in the end, it is possible to display the number of occurrences of each volleyball technique with a timestamp from the moment it occurred.
    In an alternative embodiment, the method according to the present invention also comprises a sixth step. In this sixth step, the smash power, which is correlated with the speed of the ball, is determined. The power of smash is determined as follows: when a volleyball technique has been detected the raw acceleration data, in a certain window about 0.6 s to 0.8 s before and after the strike, are bandpass filtered with a second-order IIR filter. The sum on the standard bandpass filtered data is strongly correlated with the velocity of the ball. Since the watch uses only two digits to display the integral corresponding to the power of smash, the sum on the filtered bandpass data is expressed in a scale from 0 to 99.
    Said power is associated with a volleyball technique. Said power is also stored in the memory. The memory could be managed such that for each detected volleyball activity, the power of said detected activity is recorded. It will be understood that for each smash, each pass or each service, the power is recorded. The portable object is therefore able to display, for each occurrence, the power of the volleyball activity.
    Whenever a volleyball activity is detected and saved by the volleyball algorithm, the operating system of the watch is notified. In this way, the watch can update the display.
    It will further be understood that various modifications and / or improvements obvious to one skilled in the art can be made to the various embodiments which are the subject of the present description without departing from the scope of the invention defined by the appended claims .
    It will be understood that this method could be used for other activities such as basketball, baseball, tennis or badminton. In addition, it will be understood that this method could be used to differentiate a player's applause from a strike or a service. Therefore, the effectiveness of the portable object comprising said method is higher. It is also possible to have a specific function in the portable object regarding applause or ola.
    权利要求:
    Claims (6)
    [1]
    A detection method for an electronic apparatus comprising a housing containing an electronic module supplied with electrical energy by an electrical energy storage means, said electronic module comprising a calculation unit connected to an acceleration sensor and to a control unit. memory, said electronic module also being connected to a display means so as to display information related to data of said acceleration sensor, the method comprising the steps of:- acquire data from the acceleration sensor;Detect potential activities from data acquired using a high-pass filter to detect rapid changes in the measured acceleration, and using a low-pass filter to filter, in the data transmitted by the high-pass filter , a calculated standard of transmitted data to detect only significant activities;- filtering above a second threshold, in the data transmitted by the low-pass filter, the maxima of the norm of the data having passed the high-pass filter, to reject a larger part of the irrelevant activities and to determine a peak / hit of this standard;- prepare the data for the classification by filtering the acceleration data, by low-pass filtering, before and after the peak / strike, progressing towards the striking from both directions, and subsampling them starting with the value closer to the keystroke, the subsampled data passing to the classification step in which an algorithm is trying to determine which activity has occurred;- characterize activities on the basis of a nearest neighbor algorithm using the L1 standard in the subsampled accelerations space.
    [2]
    2. Method according to claim 1, characterized in that it further comprises a step of determining the striking power by using a bandpass filter to filter the acceleration data in a certain window before and after the shot, the sum on the norm of the past band filtered data being correlated with the velocity of the ball.
    [3]
    The method according to claim 1 or 2, wherein the computing unit is able to switch from a low power mode, wherein said computing unit is dormant, to a normal mode in which said computing unit is activated if the acceleration sensor which continuously acquires data at a frequency of approximately between 30 and 50 Hz, measures acceleration data above a determined threshold.
    [4]
    4. The method of claim 3, wherein the determined threshold is between 1.5 and 2.5 g.
    [5]
    The method of any of the preceding claims, wherein the low-pass filter and the high-pass filter use a second-order filter.
    [6]
    6. The method of claim 1, wherein the second threshold is between 0.2 and 0.6 g.
    类似技术:
    公开号 | 公开日 | 专利标题
    JP5803962B2|2015-11-04|Information processing apparatus, sensor apparatus, information processing system, and recording medium
    FR2981561A1|2013-04-26|METHOD FOR DETECTING MOTION SENSOR ACTIVITY, CORRESPONDING DEVICE AND COMPUTER PROGRAM
    EP2483758B1|2019-06-26|System and method for recognizing gestures
    EP1586353B1|2007-01-10|Method and device for measuring efficacy of a sportive activity
    EP3058440A1|2016-08-24|Calculating pace and energy expenditure from athletic movement attributes
    FR2990357A1|2013-11-15|METHOD FOR ANALYZING THE PLAY OF A USER OF A RACKET
    EP3120902A1|2017-01-25|Information processing apparatus, information processing method, and recording medium
    CH711373A2|2017-01-31|A method for detecting and processing accelerations, in particular during a volleyball match.
    JP6136920B2|2017-05-31|Analysis device, recording medium, and analysis method
    CN107393260B|2020-04-21|Sedentariness reminding method and device and wrist type sedentariness reminder
    EP2467061A1|2012-06-27|System and method for detecting an epileptic seizure in a prone epileptic person
    US10272324B2|2019-04-30|Classifying collision events using inertial and audio data
    US10004947B2|2018-06-26|Detection method for volley-ball function
    JP6187558B2|2017-08-30|Information processing apparatus, information processing system, and recording medium
    FR2967273A1|2012-05-11|SENSOR DETECTION DEVICE, DETECTION METHOD AND CORRESPONDING COMPUTER PROGRAM
    EP3641896A1|2020-04-29|System and method for collecting and processing data associated with players playing a ball game
    EP3761063A1|2021-01-06|Portable instrument for managing a sports or fitness activity
    FR3082722A1|2019-12-27|METHOD OF ANALYZING A CROWD OF A WALKING PEDESTRIAN
    CH716382A2|2021-01-15|Portable instrument for managing a sports or wellness activity.
    FR3028343A1|2016-05-13|METHOD FOR DETECTING THE FALL OF A HUMAN SUBJECT AND CORRESPONDING ACTIMETRIC DEVICE
    FR3053490A1|2018-01-05|SENSOR, STATISTICAL ANALYSIS SYSTEM, AND METHOD FOR DETERMINING AND DISPLAYING THE HEIGHT OF A SPORTS JUMP
    CH713220A2|2018-06-15|Method for detecting and calculating the duration of a jump.
    US20160310788A1|2016-10-27|Analysis device, recording medium, and analysis method
    WO2016066949A1|2016-05-06|Instrumented cycling shorts
    同族专利:
    公开号 | 公开日
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    法律状态:
    2020-01-15| AZW| Rejection (application)|
    优先权:
    申请号 | 申请日 | 专利标题
    CH10962015|2015-07-28|
    [返回顶部]