CN114832277A - Rope skipping mode identification method and rope skipping - Google Patents
Rope skipping mode identification method and rope skipping Download PDFInfo
- Publication number
- CN114832277A CN114832277A CN202210553143.XA CN202210553143A CN114832277A CN 114832277 A CN114832277 A CN 114832277A CN 202210553143 A CN202210553143 A CN 202210553143A CN 114832277 A CN114832277 A CN 114832277A
- Authority
- CN
- China
- Prior art keywords
- acceleration sensor
- rope
- rope skipping
- characteristic
- time window
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000001133 acceleration Effects 0.000 claims abstract description 198
- 238000012549 training Methods 0.000 claims abstract description 21
- 238000012544 monitoring process Methods 0.000 claims abstract description 9
- 238000001914 filtration Methods 0.000 claims description 36
- 238000012952 Resampling Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000012567 pattern recognition method Methods 0.000 claims 5
- 238000010586 diagram Methods 0.000 description 11
- 238000005070 sampling Methods 0.000 description 7
- 238000009434 installation Methods 0.000 description 5
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 235000008694 Humulus lupulus Nutrition 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 239000011435 rock Substances 0.000 description 1
- 238000004579 scanning voltage microscopy Methods 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B5/00—Apparatus for jumping
- A63B5/20—Skipping-ropes or similar devices rotating in a vertical plane
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B71/00—Games or sports accessories not covered in groups A63B1/00 - A63B69/00
- A63B71/06—Indicating or scoring devices for games or players, or for other sports activities
- A63B71/0619—Displays, user interfaces and indicating devices, specially adapted for sport equipment, e.g. display mounted on treadmills
- A63B71/0669—Score-keepers or score display devices
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B2220/00—Measuring of physical parameters relating to sporting activity
- A63B2220/17—Counting, e.g. counting periodical movements, revolutions or cycles, or including further data processing to determine distances or speed
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B2220/00—Measuring of physical parameters relating to sporting activity
- A63B2220/40—Acceleration
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63B—APPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
- A63B2220/00—Measuring of physical parameters relating to sporting activity
- A63B2220/80—Special sensors, transducers or devices therefor
Landscapes
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Physical Education & Sports Medicine (AREA)
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Maintenance And Inspection Apparatuses For Elevators (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a rope skipping mode identification method and a rope skipping, wherein the identification method comprises the steps of obtaining historical signal waveforms of an acceleration sensor and a Hall sensor on a rope skipping, identifying different rope skipping modes according to the falling edge of the historical signal waveform of the Hall sensor, dividing corresponding time windows, extracting rope shaking characteristics, and performing model training on a classifier corresponding to the rope skipping mode respectively by adopting the rope shaking characteristics to obtain a classifier; when skipping rope, monitoring the falling edge in real time to identify a time window of any rope skipping mode, dividing the time window, extracting real-time rope swinging characteristics and inputting the characteristics into a corresponding classifier to obtain classification probability, and classifying the rope skipping mode in the time window into the rope skipping mode and marking when the classification probability is greater than a preset classification probability threshold of the corresponding rope skipping mode. The invention combines the signals of the acceleration sensor and the Hall sensor to automatically identify double-swing and multi-swing of the rope skipping of the user, and improves the identification accuracy rate through multi-dimensional data.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a rope skipping mode identification method and a rope skipping.
Background
The current intelligent skipping rope generally uses an acceleration sensor to identify a user by single shaking, double shaking or multiple shaking, but the identification accuracy is not very high.
Disclosure of Invention
The present invention is directed to solving at least one of the problems of the prior art. Therefore, the invention provides a method for identifying a rope skipping mode, which adopts the technical problems that:
a rope skipping mode identification method comprises the following steps that a rope skipping comprises an acceleration sensor and at least two Hall sensors, a magnet is installed on one side of a rotating shaft of the rope skipping, the Hall sensors are evenly distributed at intervals along the circumferential direction of the rotating shaft by taking the rotating shaft as a center, and the acceleration sensor is installed on the rope skipping, and the method comprises the following steps:
step 1, obtaining historical signal waveforms of an acceleration sensor and a Hall sensor on a skipping rope, and identifying and dividing time windows of different skipping rope modes according to falling edges of the historical signal waveforms of the Hall sensor;
extracting corresponding rope swinging characteristics from historical signal waveforms of the acceleration sensor in all time windows of each rope skipping mode, and performing model training on a classifier corresponding to the rope skipping mode by adopting the rope swinging characteristics of each rope skipping mode to obtain the classifier of each rope skipping mode;
and 2, during rope skipping, monitoring the signal waveform of the Hall sensor, identifying and dividing a time window of any rope skipping mode according to a falling edge on the signal waveform of the Hall sensor, extracting real-time rope shaking characteristics of the rope skipping mode in the time window, inputting the real-time rope shaking characteristics into a classifier corresponding to the rope skipping mode to obtain classification probability, classifying the rope skipping mode in the time window into the rope skipping mode and marking the rope skipping mode when the classification probability is greater than a preset classification probability threshold value of the corresponding rope skipping mode, and otherwise, not marking the rope skipping mode.
Preferably, the specific method for identifying and dividing the time window of any one rope skipping mode according to the falling edge on the signal waveform of the hall sensor comprises the following steps:
counting after a falling edge appears on a signal waveform of any one Hall sensor, and waiting for the falling edge of the other Hall sensor to appear until M falling edges are accumulated, wherein M is A (N +1), A is the number of the Hall sensors, and N is the rope shaking time of one jump in the rope skipping mode;
and calculating whether the interval between every two adjacent falling edges is less than preset time, and if so, dividing the accumulated signal waveforms between the M falling edges into a time window with one jump of N swing.
Preferably, in step 2, the method for identifying and dividing the time window of any one of the rope skipping modes according to the falling edge on the hall sensor signal waveform further includes: and (3) sequentially identifying the time window of any rope skipping mode by taking each falling edge as a starting point according to the appearance sequence of the falling edge from the first falling edge appearing on the signal waveform of any Hall sensor and classifying the rope skipping modes.
Preferably, when the time windows corresponding to the classification marks of any two rope skipping modes overlap, and the rope skipping mode of the previous time window is classified as one rock N hops, N is equal to [2, N ],
if the distance between the starting point of the next time window and the starting point of the previous time window is greater than or equal to n × A-1 falling edges, the classification marks of the rope skipping modes corresponding to the two time windows are reserved;
and if the distance between the starting point of the next time window and the starting point of the previous time window is less than n × A-1 falling edges, deleting the classification mark of the rope skipping mode corresponding to the next time window.
Preferably, if two different time windows are identified according to the same falling edge of the hall sensor and the rope skipping modes of the corresponding time windows are classified into two different rope skipping modes, the classification marks of the time windows with the larger number of falling edges and the rope skipping modes of the corresponding time windows are reserved, and the classification marks of the time windows with the smaller number of falling edges and the rope skipping modes of the corresponding time windows are deleted.
Preferably, when the rope skipping mode is one-jump double-shake, the time window corresponding to the one-jump double-shake is divided into a first part and a second part, wherein the first part is a time window from the first falling edge to the 1+ A falling edge, the second part is a time window from the 1+ A falling edge to the 1+2A falling edge,
the rope swinging characteristics corresponding to one jump and double swings comprise:
a characteristic a, which is the ratio of the duration of the first portion to the duration of the second portion;
characteristic ofb 1 The relative position of the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part and the second part in the time window of the one-jump double-shaking,
characteristic c 1 A difference between a maximum value of a signal waveform on each axis of the acceleration sensor in the first section and a maximum value in the second section;
characteristic d 1 The difference between the minimum value of the signal waveform on each axis of the acceleration sensor in the first part and the minimum value of the signal waveform on each axis of the acceleration sensor in the second part;
characteristic e 1 The ratio of the energy average value of the signal waveform on each axis of the acceleration sensor in the first part to the energy average value of the second part;
characteristic f 1 Is as a feature c 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic d 1 Sum of absolute values of the values on three axes of the acceleration sensor 1 Sum of absolute values of values on three axes of the acceleration sensor, for a three-axis acceleration sensor c 1 、d 1 、e 1 The features are calculated to obtain 3 values, and the sum of the absolute values of the 3 values of the features is calculated to obtain the feature f 1 ，f 1 Having three values, distribution being c 1 Sum of absolute values of three values, d 1 Sum of absolute values of three values, e 1 The sum of the absolute values of the three values;
characteristic b 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part and the second part is at the relative position in the time window of the one-jump double-swing;
characteristic c 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the difference between the maximum value of the signal waveform on each axis of the acceleration sensor in the first part and the maximum value in the second part is obtained;
characteristic d 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the acceleration sensor is arranged on each shaftThe difference between the minimum value of the signal waveform in the first portion and the minimum value in the second portion;
the characteristic e2 is that after the signal waveform of the acceleration sensor is subjected to band-pass filtering at 0.5-6 Hz, the signal waveform on each axis of the acceleration sensor is in the ratio of the energy mean value of the first part to the energy mean value of the second part;
characteristic f 2 Is as a feature c 2 Characteristic d 2 And feature e 2 Sum of absolute values on each axis of the acceleration sensor.
Preferably, when the rope skipping mode is one-jump three-shake, the time window corresponding to the one-jump three-shake is divided into a first part, a second part and a third part, wherein the first part is a time window from a first falling edge to a 1+ A falling edge, the second part is a time window from a 1+ A falling edge to a 1+2A falling edge, the third part is a time window from a 1+2A falling edge to a 1+3A falling edge,
rope swinging characteristics corresponding to one jump and three swings comprise:
characteristic g, being the time ratio between the first portion, the second portion and the third portion, the time ratio between the first portion and the second portion, the time ratio between the second portion and the third portion;
characteristic h 1 The relative position of the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part, the second part and the third part in the time window of one-jump and three-swing is shown;
characteristic i 1 A difference between maximum values of the signal waveform on each axis of the acceleration sensor at the first portion and the second portion;
characteristic j 1 A difference between maximum values of the signal waveform on each axis of the acceleration sensor at the second portion and the third portion;
feature k 1 A difference between minimum values of the waveform of the signal on each axis of the acceleration sensor in the first portion and the second portion;
characteristic l 1 A difference between minimum values of the signal waveform on each axis of the acceleration sensor in the second portion and the third portion;
feature m 1 The ratio of the energy mean values of the signal waveform on each axis of the acceleration sensor in the first part and the second part is obtained;
characteristic n 1 The ratio of the energy mean values of the signal waveform on each axis of the acceleration sensor in the second part and the third part;
characteristic o 1 Is a characteristic i 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic j 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic k 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic l 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic m 1 Sum of absolute values of the values on three axes of the acceleration sensor characteristic n 1 The sum of the absolute values of the values on the three axes of the acceleration sensor;
characteristic h 2 After the acceleration data are subjected to band-pass filtering of 0.5-6 Hz, the maximum peak values of the signal waveforms on each axis of the acceleration sensor in the first part, the second part and the third part are in the relative positions in the time window of one-jump three-swing;
characteristic i 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference of the maximum values of the signal waveforms on each axis of the acceleration sensor in the first part and the second part is calculated;
characteristic j 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference of the maximum values of the signal waveforms on each axis of the acceleration sensor in the second part and the third part is calculated;
feature k 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference between the minimum values of the signal waveforms on each axis of the acceleration sensor in the first part and the second part is calculated;
characteristic l 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference between the minimum values of the signal waveform on each axis of the acceleration sensor in the second part and the third part is calculated;
feature m 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 HzThe ratio of the energy mean values of the signal waveform on each axis of the acceleration sensor in the first part and the second part;
characteristic n 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the energy average value ratio of the signal waveform on each axis of the acceleration sensor in the second part and the third part is calculated;
the characteristic o2 is that after the acceleration data is subjected to band-pass filtering at 0.5-6 Hz, the characteristic i 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic j 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic k 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic l 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic m 2 Sum of absolute values of the values on three axes of the acceleration sensor characteristic n 2 The sum of the absolute values of the values on the three axes of the acceleration sensor.
In the above contents:
the relative position of the corresponding time window of the maximum peak value is calculated by using normalization, the starting point of the corresponding time window is set as 0, the end point of the corresponding time window is set as 1, and the relative position of the maximum peak value is (the time of the maximum peak value appearing-the time of the starting point)/(the time of the end point-the time of the starting point);
the method for calculating the energy value of any sampling point of the signal waveform comprises the following steps: the method for calculating the mean value of the energy of the time window and the square value or the absolute value of the acceleration sampling value of the sampling point comprises the following steps: and averaging the energy values of all the sampling points in the time window in the part to obtain the energy average value.
Preferably, the training method of the classifier of each rope skipping mode comprises the following steps:
randomly extracting a plurality of time windows, resampling the extracted time windows, enabling the proportion of the time window of the rope skipping mode to the time window of the non-rope skipping mode in the extracted time windows to be 1:1, and performing model training on the classifier by taking rope shaking characteristics of the resampled time windows as training data.
Preferably, the preset classification probability threshold value of each rope skipping mode is automatically adjusted according to whether the rope skipping mode is marked or not and the time interval between the rope skipping mode and the last rope skipping mode marked.
This application still provides a rope skipping, the rope skipping includes acceleration sensor and two at least hall sensor, and magnet is installed to the pivot one side of rope skipping, hall sensor with the pivot is the even distribution of its circumference interval of center edge, acceleration sensor installs on the rope skipping, include:
the rope skipping mode identification module is used for monitoring the signal waveform of the Hall sensor in real time to identify a time window of a rope skipping mode and dividing the time window when rope skipping is carried out;
the real-time rope swinging feature extraction module is used for extracting real-time rope swinging features corresponding to a rope skipping mode on a signal waveform of the acceleration sensor in a current time window;
the classifier of each rope skipping mode is obtained by performing model training on the classifier through the rope swinging characteristics of each rope skipping mode, and is used for receiving the real-time rope swinging characteristics of the rope skipping mode and outputting the real-time classification probability of the current rope skipping mode;
the rope skipping mode marking module is used for comparing the real-time classification probability with a preset classification probability threshold value of the rope skipping mode, when the real-time classification probability is larger than the preset classification probability threshold value of the corresponding rope skipping mode, the rope skipping mode of the time window is classified into the rope skipping mode, otherwise, the rope skipping mode is not marked;
and the rope skipping mode counting module is used for counting the number of marks of each rope skipping mode respectively.
The invention has the beneficial effects that: the invention combines the signals of the acceleration sensor and the Hall sensor to automatically identify double-swing and multi-swing of the rope skipping of the user, and improves the identification accuracy rate through multi-dimensional data.
Drawings
The above and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a schematic view showing the installation of a Hall sensor in embodiment 1;
FIG. 2 is one of the schematic views of the installation of the Hall sensor;
FIG. 3 is a second schematic view of the installation of the Hall sensor;
FIG. 4 is a diagram showing a time window of one-hop double-shaking divided according to a historical signal waveform of a Hall sensor in embodiment 1;
FIG. 5 is a time window identified as one-hop double-shaking in example 1;
FIG. 6 is a schematic diagram of the time window of one-jump double-shaking divided into a first part and a second part by using the 3 rd and 5 th falling edges in example 1;
FIG. 7 is a diagram of the time window of one-jump three-shake divided into the first part, the second part and the third part by the 3 rd, the 5 th and the 7 th falling edges in example 2;
fig. 8 is a block diagram of the skipping rope in example 3.
Detailed Description
Reference will now be made in detail to the present preferred embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to like elements throughout.
In the description of the present invention, a plurality of means is two or more, and greater than, less than, more than, etc. are understood as excluding the present number, and greater than, less than, etc. are understood as including the present number. If the first and second are described for the purpose of distinguishing technical features, they are not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated or implicitly indicating the precedence of the technical features indicated.
In the description of the present invention, it should be understood that the orientation or positional relationship referred to in the description of the orientation, such as the upper, lower, front, rear, left, right, etc., is based on the orientation or positional relationship shown in the drawings, and is only for convenience of description and simplification of description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus, should not be construed as limiting the present invention.
In the present invention, unless explicitly defined otherwise, the terms "disposed," "mounted," "connected," and the like are to be understood in a broad sense, and for example, may be directly connected or indirectly connected through an intermediate; can be fixedly connected, can also be detachably connected and can also be integrally formed; may be a mechanical connection; either as communication within the two elements or as an interactive relationship of the two elements. The specific meaning of the above-mentioned words in the present invention can be reasonably determined by those skilled in the art in combination with the detailed contents of the technical solutions.
The application provides a rope skipping mode's identification method, and rope skipping mode is that a jump N shakes, and N is integer and more than or equal to 2, and the rope skipping includes acceleration sensor and two at least hall sensor, and magnet is installed to the pivot one side of rope skipping, hall sensor with the pivot is the even distribution of its circumference interval along the center, acceleration sensor installs on rope skipping, and the method includes following step:
step 1, obtaining historical signal waveforms of an acceleration sensor and a Hall sensor on a skipping rope, and identifying and dividing time windows of different skipping rope modes according to falling edges of the historical signal waveforms of the Hall sensor;
extracting corresponding rope shaking characteristics from historical signal waveforms of the acceleration sensor in all time windows of each rope skipping mode, and performing model training on a classifier of the corresponding rope skipping mode by adopting the rope shaking characteristics of each rope skipping mode to obtain the classifier of each rope skipping mode;
and 2, during rope skipping, monitoring the signal waveform of the Hall sensor, identifying and dividing a time window of any rope skipping mode according to a falling edge on the signal waveform of the Hall sensor, extracting real-time rope shaking characteristics of the rope skipping mode in the time window, inputting the real-time rope shaking characteristics into a classifier corresponding to the rope skipping mode to obtain classification probability, classifying the rope skipping mode in the time window into the rope skipping mode and marking the rope skipping mode when the classification probability is greater than a preset classification probability threshold value of the corresponding rope skipping mode, and otherwise, not marking the rope skipping mode.
Based on the above, embodiments of the present application are proposed:
example 1
The embodiment specifically provides a method for identifying one-hop double-shake, and referring to fig. 1 to 3, installation schematic diagrams of several hall sensors are shown, in the embodiment, a jump rope is provided with a rotating shaft connected with a rope, the jump rope further comprises a controller, a magnet, an acceleration sensor and two hall sensors, fig. 1 is an installation schematic diagram of the hall sensors in the embodiment, and the method for identifying one-hop double-shake in the embodiment comprises the following steps:
step 1, historical signal waveforms of an acceleration sensor and a Hall sensor on a skipping rope are obtained, one-skipping double-shaking is identified according to the historical signal waveforms of the Hall sensor, corresponding time windows are divided, corresponding rope shaking characteristics are extracted from the historical signal waveforms of the acceleration sensor in all the time windows of the one-skipping double-shaking, model training is conducted on a classifier through the corresponding rope shaking characteristics of the one-skipping double-shaking, and the classifier of the one-skipping double-shaking is obtained;
and 2, during rope skipping, monitoring signal waveforms of the Hall sensors in real time until a one-jump double-swing time window is identified, dividing the time window, extracting real-time rope swinging characteristics of the one-jump double-swing in the current time window, inputting the real-time rope swinging characteristics into a one-jump double-swing classifier to obtain classification probability, classifying a rope skipping mode in the time window into the one-jump double-swing mode and marking the rope skipping mode in the time window as the one-jump double-swing mode when the classification probability is larger than a preset classification probability threshold corresponding to the one-jump double-swing mode, and otherwise, not marking the rope skipping mode.
Referring to fig. 4, a schematic diagram of a time window for dividing a one-hop double-shaking according to a historical signal waveform of a hall sensor in this embodiment is shown, and a waveform diagram of the hall sensor in fig. 4 is a waveform diagram of two hall sensors.
The specific method for identifying and dividing the time window of any rope skipping mode according to the falling edge signal waveform on the Hall sensor signal waveform comprises the following steps:
counting is started after falling edges appear on the signal waveform of any one Hall sensor, the falling edges of the other Hall sensor are waited for appearing until M falling edges are accumulated, M is A (N +1), wherein A is the number of the Hall sensors, N is the rope shaking frequency of one jump in the rope skipping mode, whether the interval between every two adjacent falling edges is smaller than preset time or not is calculated, and if yes, the accumulated signal waveform between the M falling edges is divided into a time window of one jump N shaking.
In this embodiment, when 6 falling edges are accumulated, whether an interval between every two adjacent falling edges is less than 1s is calculated, and if yes, a time window of one-hop double shaking is divided. And 6 descending edges are selected to represent that the rotating shaft rotates for 2.5 circles, and if a complete double-swing action exists in the 2.5 circles, any swing rope characteristic is not lost.
Corresponding to the determination of the time window of one-jump three-shake, 6 falling edges are changed into 8 falling edges on the basis of the double-shake window;
corresponding to the determination of the time window of one-jump four-swing, 8 falling edges are changed into 10 falling edges on the basis of the three-swing window;
and corresponding to the determination of the time window of one-jump five-swing, on the basis of a four-swing window, changing 10 falling edges into 12 falling edges.
The sampling rate of the Hall sensor can be the same as that of the acceleration sensor or an integral multiple of the acceleration sensor, and when the Hall sensor is the integral multiple of the sampling rate of the acceleration sensor, the starting point and the end point of the corresponding acceleration data can be obtained only by dividing the number of the falling edges of the Hall sensor at the starting point and the end point of the time window by the multiple.
The sampling rate of the acceleration sensor and the Hall sensor is more than or equal to 50 Hz.
Referring to fig. 5, the arrows show the time window divided into one-hop double-shaking.
Specifically, in step 2, the method for identifying and dividing the time window of any one rope skipping mode according to the falling edge on the hall sensor signal waveform further includes: and (3) sequentially identifying the time window of any rope skipping mode by taking each falling edge as a starting point according to the appearance sequence of the falling edge from the first falling edge appearing on the signal waveform of any Hall sensor, and classifying and marking the rope skipping modes.
The method traverses all the falling edges, and avoids omission in the division of the time window.
Preferably, when two time windows with skipping rope mode marks overlap, and the skipping rope mode of the previous time window is one rocking N jumps, N is equal to [2, N ],
if the distance between the starting point of the next time window and the starting point of the previous time window is greater than or equal to n × A-1 falling edges, keeping the marks of the rope skipping modes of the two time windows;
and if the distance between the starting point of the next time window and the starting point of the previous time window is less than n × A-1 falling edges, deleting the mark of the rope skipping mode of the next time window.
For example, the current time window is classified as one-jump double-swing, if the distance between the starting point of the previous time window and the starting point of the next time window is greater than or equal to 3 falling edges, the rope skipping mode marks of the two time windows are reserved, and if the distance between the starting point of the previous time window and the starting point of the next time window is less than 3 falling edges, the rope skipping mode mark of the next time window is deleted;
and if the distance between the starting point of the previous time window and the starting point of the next time window is less than 5 falling edges, deleting the skipping rope mode mark of the next time window.
By adopting the method, the overlapped time windows are screened, so that the time windows corresponding to the classification marks of any two rope skipping modes are prevented from being overlapped and colliding with each other.
Preferably, the skipping pattern is marked for a time window with a large number of falling edges if the same falling edge is used as a starting point, time windows with different lengths are identified and divided, and the time windows are classified into different skipping patterns.
For example, when historical rope skipping data is counted, if a time window of one-jump double-swing and one-jump triple-swing is identified and divided by taking the same falling edge as a starting point, and rope skipping modes of the two time windows are respectively classified into one-jump double-swing and one-jump triple-swing, marking the rope skipping mode of the time window of one-jump triple-swing is performed, and not marking the time window of one-jump double-swing.
For example, when a user jumps, a time window of one-jump double-swing is identified according to the falling edge of the Hall sensor, the rope skipping of the time window is classified into the one-jump double-swing, the marking of the one-jump double-swing on the time window is delayed, the signal waveform of the Hall sensor is continuously monitored, when the time window of the one-jump double-swing is identified and divided by taking the starting point of the time window of the one-jump double-swing as the starting point and is classified into the one-jump triple-swing, the rope skipping mode of the time window corresponding to the one-jump triple-swing is marked as the one-jump triple-swing, and the time window of the one-jump double-swing is not marked.
When the rope skipping mode is one-jump double-swing, a time window corresponding to the one-swing double-jump is divided into a first part and a second part, wherein the first part is a time window from a first falling edge to a 3 rd falling edge, the second part is a time window from the 3 rd falling edge to a 5 th falling edge,
the rope swinging characteristics corresponding to one jump and double swings comprise:
a characteristic a, which is the ratio of the duration of the first portion to the duration of the second portion;
characteristic b 1 The relative position of the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part and the second part in the time window of the one-hop double-rocking is shown;
characteristic c 1 A difference between a maximum value of a signal waveform on each axis of the acceleration sensor in the first portion and a maximum value in the second portion;
characteristic d 1 The difference between the minimum value of the signal waveform on each axis of the acceleration sensor in the first part and the minimum value of the signal waveform on each axis of the acceleration sensor in the second part;
characteristic e 1 For each of the acceleration sensorsThe ratio of the mean value of the energy of the signal waveform on the axis in the first part to the mean value of the energy of the second part;
characteristic f 1 Is as a feature c 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic d 1 Sum of absolute values of the values on three axes of the acceleration sensor 1 The sum of the absolute values of the values on the three axes of the acceleration sensor;
characteristic b 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part and the second part is at the relative position in the time window of the one-jump double-swing;
characteristic c 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the difference between the maximum value of the signal waveform on each axis of the acceleration sensor in the first part and the maximum value in the second part is obtained;
characteristic d 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the difference between the minimum value of the signal waveform on each axis of the acceleration sensor in the first part and the minimum value of the signal waveform on each axis of the acceleration sensor in the second part is obtained;
characteristic e 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the ratio of the energy mean value of the signal waveform on each axis of the acceleration sensor in the first part to the energy mean value of the second part is calculated;
characteristic f 2 Is as a feature c 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic d 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic e 2 The sum of the absolute values of the values on the three axes of the acceleration sensor.
Referring to fig. 6, a schematic diagram of a one-hop double-roll time window divided into a first portion and a second portion using the 3 rd and 5 th falling edges is shown.
The method for carrying out model training on the classifier by utilizing the corresponding rope swinging characteristics of one-jump double swinging comprises the following steps:
70% of the time windows of all one-hop double-shaking are randomly drawn for model training, and the other 30% are used for verifying the performance of the model.
Resampling 70% of randomly extracted time windows, ensuring that the proportion of time windows of 'one-jump double shaking' and 'non-jump double shaking' is 1:1, and performing model training on a classifier by using rope shaking characteristics of the resampled time windows as training data, wherein 'double shaking' is a positive example, and 'non-double shaking' is a negative example, and the model training can use classifiers such as SVM, random forest, logistic regression, GBDT, XGBOOST, LIGHT TBGM and the like;
and inputting the rope swinging characteristics of 30% of the time window into a trained classifier to obtain a classification result and corresponding probability, uniformly using 0.5 as a threshold value during verification, wherein the case that the probability is greater than the threshold value is a positive case, and when the precision and recall ratio of the verified classification result and the actual classification result meet requirements, the performance of the classifier passes the verification and can be used on rope skipping, otherwise, data needs to be continuously supplemented to train the classifier until the precision and recall ratio meet the requirements.
The preset classification probability threshold value of each rope skipping mode is automatically adjusted according to whether the rope skipping mode is marked or not and the time interval from the last marking of the rope skipping mode, the classification probability threshold value is improved to reduce misjudgment when one-jump double-shaking is not marked, and the threshold value is reduced to improve the recognition rate when one-jump double-shaking is found.
For example, the initial classification probability threshold is set to 0.85;
when the rope skipping mode of a time window is marked as one-jump double-swing, the classification probability threshold value of the one-jump double-swing is automatically adjusted to 0.5;
when the time distance is more than 5s from the end point of a time window marked as one-jump double-shake, the classification probability threshold value of the one-jump double-shake is automatically adjusted to 0.7
When the time distance is more than 10s from the end point of a time window marked as one-jump double-shake, the classification probability threshold value of the one-jump double-shake is automatically adjusted to 0.85.
Example 2
The embodiment 2 provides a one-jump three-swing identification method, wherein a skipping rope is provided with a rotating shaft connected with a rope, the skipping rope further comprises a controller, a magnet, an acceleration sensor and at least two Hall sensors, the magnet is installed on one side of the rotating shaft, the Hall sensors are uniformly distributed around the rotating shaft at intervals along the circumferential direction of the rotating shaft, the acceleration sensor is installed on the skipping rope, and the acceleration sensor and the Hall sensors are respectively and electrically connected with the controller, and the method comprises the following steps:
step 1, historical signal waveforms of an acceleration sensor and a Hall sensor on a skipping rope are obtained, one-skip three-shake is recognized according to the historical signal waveforms of the Hall sensor, corresponding time windows are divided, corresponding rope shaking characteristics are extracted from the historical signal waveforms of the acceleration sensor in all the time windows of the one-skip three-shake, model training is conducted on a classifier through the corresponding rope shaking characteristics of the one-skip three-shake, and a classifier corresponding to a rope skipping mode is obtained;
and 2, during rope skipping, monitoring signal waveforms of the Hall sensors in real time until a one-jump three-shake time window is identified, dividing the time window, extracting real-time rope-skipping characteristics of one-jump three-shake in the current time window, inputting the real-time rope-skipping characteristics into a one-jump three-shake classifier to obtain classification probability, classifying rope-skipping modes in the time window into one-jump three-shake modes and marking the rope-skipping modes corresponding to the time window as the one-jump three-shake modes when the classification probability is larger than a preset classification probability threshold value of the one-jump three-shake modes, and otherwise, not marking the rope-skipping modes.
The specific method for identifying one-hop double-swing from the historical signal waveform of the hall sensor is similar to embodiment 1, except that M is 8:
when the rope skipping mode is one-skipping three-rocking, the time window corresponding to the one-skipping three-rocking is divided into a first part, a second part and a third part, wherein the first part is a time window from a first falling edge to a 3 rd falling edge, the second part is a time window from the 3 rd falling edge to a 5 th falling edge, the third part is a time window from the 5 th falling edge to a 7 th falling edge,
rope swinging characteristics corresponding to one jump and three swings comprise:
characteristic g, being the time ratio between the first portion, the second portion and the third portion, the time ratio between the first portion and the second portion, the time ratio between the second portion and the third portion;
characteristic h 1 The relative position of the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part, the second part and the third part in the time window of one-jump and three-swing is shown;
characteristic i 1 A difference between maximum values of the signal waveform on each axis of the acceleration sensor at the first portion and the second portion;
characteristic j 1 A difference between maximum values of the signal waveform on each axis of the acceleration sensor at the second portion and the third portion;
feature k 1 A difference between minimum values of the waveform of the signal on each axis of the acceleration sensor in the first portion and the second portion;
characteristic l 1 A difference between minimum values of the signal waveform on each axis of the acceleration sensor in the second portion and the third portion;
feature m 1 The ratio of the energy mean values of the signal waveform on each axis of the acceleration sensor in the first part and the second part is obtained;
characteristic n 1 The ratio of the energy mean values of the signal waveform on each axis of the acceleration sensor in the second part and the third part;
characteristic o 1 Is a characteristic i 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic j 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic k 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic l 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic m 1 Sum of absolute values of the values on three axes of the acceleration sensor characteristic n 1 The sum of the absolute values of the values on the three axes of the acceleration sensor;
characteristic h 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the waveform of a signal on each axis of the acceleration sensor isThe relative position of the maximum peak value of the first part, the second part and the third part in the time window of one-jump three-swing;
characteristic i 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference of the maximum values of the signal waveforms on each axis of the acceleration sensor in the first part and the second part is calculated;
characteristic j 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference of the maximum values of the signal waveforms on each axis of the acceleration sensor in the second part and the third part is calculated;
feature k 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference between the minimum values of the signal waveforms on each axis of the acceleration sensor on the first part and the second part is calculated;
characteristic l 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference between the minimum values of the signal waveform on each axis of the acceleration sensor in the second part and the third part is calculated;
feature m 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the energy average value ratio of a signal waveform on each axis of the acceleration sensor in the first part and the second part is calculated;
characteristic n 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the energy average value ratio of the signal waveform on each axis of the acceleration sensor in the second part and the third part is calculated;
characteristic o 2 After the acceleration data is subjected to band-pass filtering at 0.5-6 Hz, the characteristic i 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic j 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic k 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic l 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic m 2 Sum of absolute values of the values on three axes of the acceleration sensor 2 The sum of the absolute values of the values on the three axes of the acceleration sensor.
Referring to fig. 7, a schematic diagram of a time window of one-jump three-swing using 3 rd, 5 th and 7 th falling edges is divided into a first part, a second part and a third part.
The method of model training the classifier using the corresponding rope sway feature of one jump and three sways is the same as in example 1.
The corresponding rope swinging characteristics of the first-jump fourth swing and the first-jump fifth swing can extend the calculation mode of the corresponding rope swinging characteristics of the first-jump third swing, the time window is divided into a first part, a second part, a third part and a fourth part in advance by the first-jump fourth swing, the time window is divided into a first part, a second part, a third part, a fourth part and a fifth part in advance by the first-jump fifth swing, and rope swinging characteristics corresponding to the first-jump fourth swing and the first-jump fifth swing are calculated by adopting the method.
Example 3
Based on the above, embodiment 3 provides a skipping rope, and the skipping rope includes acceleration sensor and at least two hall sensors, and magnet is installed to the pivot one side of skipping rope, hall sensor uses the pivot as the even distribution of its circumference interval along its circumference as the center, acceleration sensor installs on skipping rope, includes:
the rope skipping mode identification module is used for monitoring the signal waveform of the Hall sensor in real time to identify a time window of a rope skipping mode and dividing the time window when rope skipping is carried out;
the real-time rope swinging feature extraction module is used for extracting real-time rope swinging features of a corresponding rope skipping mode on a signal waveform of the acceleration sensor in a current time window;
the classifier of each rope skipping mode is obtained by performing model training on the classifier through the rope swinging characteristics of each rope skipping mode, and is used for receiving the real-time rope swinging characteristics of the rope skipping mode and outputting the real-time classification probability of the current rope skipping mode;
the rope skipping mode marking module is used for comparing the real-time classification probability with a preset classification probability threshold value of the rope skipping mode, when the real-time classification probability is larger than the preset classification probability threshold value of the corresponding rope skipping mode, the rope skipping mode of the time window is classified into the rope skipping mode and marked, otherwise, the rope skipping mode is not marked;
and the rope skipping mode counting module is used for counting the number of marks of each rope skipping mode respectively so as to count the number of skipping ropes of each rope skipping mode.
It is to be understood that the present invention is not limited to the above-described embodiments, and that equivalent modifications and substitutions may be made by those skilled in the art without departing from the spirit of the present invention, and that such equivalent modifications and substitutions are to be included within the scope of the appended claims.
Claims (10)
1. The rope skipping mode identification method is characterized in that a rope skipping comprises an acceleration sensor and at least two Hall sensors, a magnet is installed on one side of a rotating shaft of the rope skipping, the Hall sensors are evenly distributed at intervals along the circumferential direction of the rotating shaft by taking the rotating shaft as a center, the acceleration sensor is installed on the rope skipping, and the method comprises the following steps:
step 1, obtaining historical signal waveforms of an acceleration sensor and a Hall sensor on a skipping rope, and identifying and dividing time windows of different skipping rope modes according to falling edges of the historical signal waveforms of the Hall sensor;
extracting corresponding rope swinging characteristics from historical signal waveforms of the acceleration sensor in all time windows of each rope skipping mode, and performing model training on a classifier corresponding to the rope skipping mode by adopting the rope swinging characteristics of each rope skipping mode to obtain the classifier of each rope skipping mode;
and 2, during rope skipping, monitoring the signal waveform of the Hall sensor, identifying and dividing a time window of any rope skipping mode according to a falling edge on the signal waveform of the Hall sensor, extracting real-time rope shaking characteristics of the rope skipping mode in the time window, inputting the real-time rope shaking characteristics into a classifier corresponding to the rope skipping mode to obtain classification probability, classifying the rope skipping mode in the time window into the rope skipping mode and marking the rope skipping mode when the classification probability is greater than a preset classification probability threshold value of the corresponding rope skipping mode, and otherwise, not marking the rope skipping mode.
2. The method for identifying the rope skipping mode according to claim 1, wherein the specific method for identifying and dividing the time window of any one rope skipping mode according to the falling edge on the signal waveform of the Hall sensor comprises the following steps:
counting after a falling edge appears on a signal waveform of any one Hall sensor, and waiting for the falling edge of the other Hall sensor to appear until M falling edges are accumulated, wherein M is A (N +1), A is the number of the Hall sensors, and N is the rope shaking time of one jump in the rope skipping mode;
and calculating whether the interval between every two adjacent falling edges is less than preset time, and if so, dividing the accumulated signal waveforms between the M falling edges into a time window with one jump of N swing.
3. The rope skipping pattern recognition method of claim 2, wherein in the step 2, the method for recognizing and dividing the time window of any one rope skipping pattern according to the falling edge on the signal waveform of the hall sensor further comprises: and (3) sequentially identifying the time window of any rope skipping mode by taking each falling edge as a starting point according to the appearance sequence of the falling edge from the first falling edge appearing on the signal waveform of any Hall sensor, and classifying and marking the rope skipping modes.
4. The rope skipping pattern recognition method according to claim 3,
when two time windows with skipping rope mode marks are overlapped, and the skipping rope mode of the previous time window is one rocking N skips, N belongs to [2, N ],
if the distance between the starting point of the next time window and the starting point of the previous time window is greater than or equal to n × A-1 falling edges, keeping the marks of the rope skipping modes of the two time windows;
and if the distance between the starting point of the next time window and the starting point of the previous time window is less than n × A-1 falling edges, deleting the mark of the rope skipping mode of the next time window.
5. The rope skipping pattern recognition method of claim 3, wherein if time windows with different lengths are recognized and divided with the same falling edge as a starting point and the rope skipping patterns are classified into different rope skipping patterns, the rope skipping patterns are marked for the time window with a larger number of falling edges.
6. The method for identifying the rope skipping mode according to claim 2, wherein when the rope skipping mode is one-jump double-shake, the time window corresponding to the one-jump double-shake is divided into a first part and a second part, wherein the first part is a time window from a first falling edge to a (1 + A) th falling edge, the second part is a time window from the (1 + A) th falling edge to a (1 + 2A) th falling edge,
the rope swinging characteristics corresponding to one jump and double swings comprise:
a characteristic a, which is the ratio of the duration of the first portion to the duration of the second portion;
characteristic b 1 The relative position of the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part and the second part in the time window of the one-hop double-rocking is shown;
characteristic c 1 A difference between a maximum value of a signal waveform on each axis of the acceleration sensor in the first portion and a maximum value in the second portion;
characteristic d 1 The difference between the minimum value of the signal waveform on each axis of the acceleration sensor in the first part and the minimum value of the signal waveform on each axis of the acceleration sensor in the second part;
characteristic e 1 The ratio of the energy average value of the signal waveform on each axis of the acceleration sensor in the first part to the energy average value of the second part;
characteristic f 1 Is as a feature c 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic d 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic e 1 The sum of the absolute values of the values on the three axes of the acceleration sensor;
characteristic b 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part and the second part is within the time window of one-jump double-swingThe relative position of (a);
characteristic c 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the difference between the maximum value of the signal waveform on each axis of the acceleration sensor in the first part and the maximum value in the second part is obtained;
characteristic d 2 After the signal waveform of the acceleration sensor is subjected to band-pass filtering of 0.5-6 Hz, the difference between the minimum value of the signal waveform on each axis of the acceleration sensor in the first part and the minimum value of the signal waveform on each axis of the acceleration sensor in the second part is obtained;
the characteristic e2 is that after the signal waveform of the acceleration sensor is subjected to band-pass filtering at 0.5-6 Hz, the signal waveform on each axis of the acceleration sensor is in the ratio of the energy mean value of the first part to the energy mean value of the second part;
characteristic f 2 Is as a feature c 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic d 2 Sum of absolute values of the values on three axes of the acceleration sensor 2 The sum of the absolute values of the values on the three axes of the acceleration sensor.
7. The method for identifying the rope skipping mode according to claim 1, wherein when the rope skipping mode is one-jump three-shake, the time window corresponding to the one-jump three-shake is divided into a first part, a second part and a third part, wherein the first part is a time window from a first falling edge to a 1+ A falling edge, the second part is a time window from a 1+ A falling edge to a 1+2A falling edge, the third part is a time window from a 1+2A falling edge to a 1+3A falling edge,
rope swinging characteristics corresponding to one jump and three swings comprise:
characteristic g, which is the time ratio of the first part and the second part, and the time ratio of the second part and the third part;
characteristic h 1 The relative position of the maximum peak value of the signal waveform on each axis of the acceleration sensor in the first part, the second part and the third part in the time window of one-jump and three-swing is shown;
characteristic i 1 To accelerateThe difference between the maximum values of the signal waveform on each axis of the degree sensor at the first part and the second part;
characteristic j 1 A difference between maximum values of the signal waveform on each axis of the acceleration sensor at the second portion and the third portion;
feature k 1 A difference between minimum values of the waveform of the signal on each axis of the acceleration sensor in the first portion and the second portion;
characteristic l 1 A difference between minimum values of the signal waveform on each axis of the acceleration sensor in the second portion and the third portion;
feature m 1 The ratio of the energy mean values of the signal waveform on each axis of the acceleration sensor in the first part and the second part is obtained;
characteristic n 1 The ratio of the energy mean values of the signal waveform on each axis of the acceleration sensor in the second part and the third part;
characteristic o 1 Is a characteristic i 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic j 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic k 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic l 1 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic m 1 Sum of absolute values of the values on three axes of the acceleration sensor characteristic n 1 The sum of the absolute values of the values on the three axes of the acceleration sensor;
characteristic h 2 After the acceleration data are subjected to band-pass filtering of 0.5-6 Hz, the maximum peak values of the signal waveforms on each axis of the acceleration sensor in the first part, the second part and the third part are in the relative positions in the time window of one-jump three-swing;
characteristic i 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference of the maximum values of the signal waveforms on each axis of the acceleration sensor in the first part and the second part is calculated;
characteristic j 2 After the acceleration data is subjected to band-pass filtering at 0.5-6 Hz, the waveform of a signal on each axis of the acceleration sensor is in the thirdThe difference between the maximum values of the second and third portions;
feature k 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference between the minimum values of the signal waveforms on each axis of the acceleration sensor in the first part and the second part is calculated;
characteristic l 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the difference between the minimum values of the signal waveform on each axis of the acceleration sensor in the second part and the third part is calculated;
feature m 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the energy average value ratio of signal waveforms on each axis of the acceleration sensor on the first part and the second part is calculated;
characteristic n 2 After the acceleration data is subjected to band-pass filtering of 0.5-6 Hz, the energy average value ratio of the signal waveform on each axis of the acceleration sensor in the second part and the third part is calculated;
characteristic o 2 After the acceleration data is subjected to band-pass filtering at 0.5-6 Hz, the characteristic i is obtained 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic j 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic k 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic l 2 Sum of absolute values of the values on three axes of the acceleration sensor, characteristic m 2 Sum of absolute values of the values on three axes of the acceleration sensor characteristic n 2 The sum of the absolute values of the values on the three axes of the acceleration sensor.
8. The rope skipping pattern recognition method of claim 1, wherein the training method of the classifier of each rope skipping pattern is as follows:
randomly extracting a plurality of time windows, resampling the extracted time windows, enabling the proportion of the time window of the rope skipping mode to the time window of the non-rope skipping mode in the extracted time windows to be 1:1, and performing model training on the classifier by taking rope shaking characteristics of the resampled time windows as training data.
9. The rope skipping pattern recognition method of claim 1, wherein the predetermined classification probability threshold of each rope skipping pattern is automatically adjusted according to whether the rope skipping pattern is marked or not and the time interval from the last marking of the rope skipping pattern.
10. The utility model provides a skipping rope, its characterized in that, skipping rope includes acceleration sensor and two at least hall sensor, and magnet is installed to the pivot one side of skipping rope, hall sensor uses the pivot is the even distribution of its circumference interval along the center, acceleration sensor installs on skipping rope, include:
the rope skipping mode identification module is used for monitoring the signal waveform of the Hall sensor in real time to identify a time window of a rope skipping mode and dividing the time window;
the real-time rope swinging feature extraction module is used for extracting real-time rope swinging features corresponding to a rope skipping mode on a signal waveform of the acceleration sensor in a current time window;
the classifier of each rope skipping mode is obtained by performing model training on the classifier through the rope swinging characteristics of each rope skipping mode, and is used for receiving the real-time rope swinging characteristics of the rope skipping mode and outputting the real-time classification probability of the current rope skipping mode;
the rope skipping mode marking module is used for comparing the real-time classification probability with a preset classification probability threshold value of the rope skipping mode, when the real-time classification probability is larger than the preset classification probability threshold value of the corresponding rope skipping mode, the rope skipping mode of the time window is classified into the rope skipping mode and marked, otherwise, the rope skipping mode is not marked;
and the rope skipping mode counting module is used for counting the number of marks of each rope skipping mode respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210553143.XA CN114832277B (en) | 2022-05-20 | 2022-05-20 | Rope skipping mode identification method and rope skipping |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210553143.XA CN114832277B (en) | 2022-05-20 | 2022-05-20 | Rope skipping mode identification method and rope skipping |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114832277A true CN114832277A (en) | 2022-08-02 |
| CN114832277B CN114832277B (en) | 2024-02-06 |
Family
ID=82573028
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210553143.XA Active CN114832277B (en) | 2022-05-20 | 2022-05-20 | Rope skipping mode identification method and rope skipping |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114832277B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086698A (en) * | 2018-07-20 | 2018-12-25 | 大连理工大学 | A kind of human motion recognition method based on Fusion |
| CN113332657A (en) * | 2021-06-25 | 2021-09-03 | 成都怡康科技有限公司 | Rope skipping handle, rope skipping counting method and rope skipping |
| CN113627340A (en) * | 2021-08-11 | 2021-11-09 | 广东沃莱科技有限公司 | Method and equipment capable of identifying rope skipping mode |
-
2022
- 2022-05-20 CN CN202210553143.XA patent/CN114832277B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086698A (en) * | 2018-07-20 | 2018-12-25 | 大连理工大学 | A kind of human motion recognition method based on Fusion |
| CN113332657A (en) * | 2021-06-25 | 2021-09-03 | 成都怡康科技有限公司 | Rope skipping handle, rope skipping counting method and rope skipping |
| CN113627340A (en) * | 2021-08-11 | 2021-11-09 | 广东沃莱科技有限公司 | Method and equipment capable of identifying rope skipping mode |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114832277B (en) | 2024-02-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107376247B (en) | A kind of swimming exercise analysis method based on smartwatch and the smartwatch | |
| CN106515724B (en) | Traffic control method and terminal in a kind of tunnel | |
| CN107273816B (en) | Traffic speed limit label detection recognition methods based on vehicle-mounted forward sight monocular camera | |
| CN109686109A (en) | A kind of parking lot security monitoring management system based on artificial intelligence | |
| CN108124485A (en) | For the alarm method of limbs conflict behavior, device, storage medium and server | |
| CN111846046B (en) | System, method and device for detecting safety of bicycle | |
| CN105489019B (en) | A kind of traffic throughput monitor system for dividing vehicle based on double-audio signal collection | |
| CN110366107A (en) | Vehicle communication method and the device for using this method | |
| CN107945311A (en) | A kind of method for early warning of dangerous driving behavior, device, storage medium and server | |
| CN109816987A (en) | A kind of automobile whistle electronic police enforces the law capturing system and its grasp shoot method | |
| CN108446678A (en) | A kind of dangerous driving behavior recognition methods based on skeleton character | |
| CN105856983B (en) | Tyre pressure sensor unit, tire pressure annunciator and vehicle | |
| CN106373461A (en) | Driver training management service system and method | |
| CN106427840A (en) | Method of self-adaptive vehicle driving mode and terminal | |
| CN108694407A (en) | A kind of driving behavior recognition methods based on mobile terminal | |
| CN109846459A (en) | A kind of fatigue driving state monitoring method | |
| CN110444013A (en) | A kind of traffic monitoring system and method for closed section | |
| CN112770293A (en) | Vehicle driving environment safety analysis early warning management cloud platform based on artificial intelligence | |
| CN107160950A (en) | A kind of vehicle running state recognition methods based on CAN | |
| CN113627340B (en) | Method and equipment capable of identifying rope skipping mode | |
| CN114832277A (en) | Rope skipping mode identification method and rope skipping | |
| CN102436578A (en) | Formation method for dog face characteristic detector as well as dog face detection method and device | |
| CN102568158A (en) | Alarming system for automobile vehicle fatigue driving | |
| CN108154283A (en) | Driver driving behavior analysis method and device | |
| CN111016914B (en) | Dangerous driving scene identification system based on portable terminal information and identification method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |