Limit scooter movement assisting method and limit scooter capable of recording real-time movement state

Abstract

The invention discloses a limit scooter movement assisting method and a limit scooter capable of recording real-time movement states, and the method comprises the steps: building a standard movement database through limit scooter standard movement track data collected through motion capture, and building a limit scooter movement track model; the limit scooter collects and records part of feature data of user motion to obtain a complete motion trail of the user by means of the model, then compares the motion trail of the user with a standard motion trail closest to the motion trail to obtain a difference point between the user motion and the standard trail motion, and enables the difference point to correspond to the feature data of the user motion. Therefore, the user can be informed of how to adjust the posture to be closer to the standard movement track during movement, so that the user is assisted in completing the limit scooter movement.

Description

Technical Field

[0001] The present invention relates to the field of extreme scooters, and in particular to an extreme scooter motion assisting method and an extreme scooter for recording real-time motion status. Background Art

[0002] Extreme scooters are scooters used for stunts and extreme sports, suitable for various fancy moves and jumps in city streets, skate parks and other venues. Its rich stunts are popular among young people, but at the same time, it also tests the user's physical coordination, sense of balance and athletic ability, which is not friendly to novices who want to get started with this sport. Existing extreme scooters usually have simple mechanical structures, and new users can only imitate and learn by referring to learning videos. Summary of the invention

[0003] In view of the problems existing in the prior art, this specification provides an extreme scooter motion assistance method and an extreme scooter that records real-time motion status. The standard motion of the extreme scooter is recorded by motion capture vision, and a standard motion trajectory database for the extreme scooter motion is established. The standard motion trajectory database is used as a training set to train a motion trajectory prediction model for the user's characteristic motion data. After obtaining the model, it is only necessary to collect part of the characteristic data of the user's motion to obtain the user's complete motion trajectory with the help of the model, and then compare the user's motion trajectory with its closest standard motion trajectory to obtain the difference between the user's motion and the standard trajectory motion. The difference is corresponding to the characteristic data of the user's motion, and the user can be informed of how to adjust the posture during the motion to be closer to the standard motion trajectory, so that a new user of the extreme scooter can also quickly get started and learn some standard movements of the extreme scooter with the assistance of the system, which greatly reduces the threshold of this sport.

[0004] To achieve the above objectives, the present invention provides, on one hand, an extreme scooter motion assisting method, comprising:

[0005] Step S1: establishing a standard motion database and building an extreme scooter motion trajectory model based on a standard data set consisting of n one-dimensional initial standard data and extreme scooter standard motion trajectory data collected by motion capture, the process includes: collecting data, establishing a database, training data and building a model;

[0006] Step S2: collecting a feature data set consisting of m one-dimensional motion feature data of a single motion of the user, and obtaining the motion trajectory of the user based on the feature data with the aid of the model;

[0007] Step S3: Determine the difference between the user's motion trajectory and the standard trajectory, and reversely map at most m motion feature data in the feature data set of the user's current motion process based on the difference;

[0008] Step S4: Providing user movement posture adjustment suggestions based on the feature data.

[0009] In the extreme scooter motion assisting method, preferably, in step S1, the standard data set consisting of n one-dimensional standard data collected in the data collection process is obtained by the following algorithm:

[0010] Determine an initial k value corresponding to a point o on the complex plane,

[0011] For n initial one-dimensional standard data x1, x2, ..., x n , calculate the reachable distance:

[0012] reach-dist(x,o)=max(dist(x,o),kd(o))

[0013] Among them, dist(x,o) is the true distance from point x to point o, and kd(o) is the k distance of point o (that is, the distance from point o to its kth nearest neighbor)

[0014] Calculate x i The k-nearest neighbor reachable density is:

[0015]

[0016] Then the decision threshold corresponding to each of the initial one-dimensional standard data is

[0017]

[0018] Judgment: If M(x i )>1, the point is a bad point and is removed from the standard data set.

[0019] If M(x i )≤1, then the point is a normal point and is included in the standard data set. In the extreme scooter motion assisting method, preferably, k=15.

[0020] In the extreme scooter motion assisting method, preferably, the training process adopts a deep neural network for training, and the different initial standard data x i is the independent variable, and the physical trajectory data y of the extreme scooter collected by vision i The true value of the required predictive variable is obtained by deep neural network training with the help of PyTorch framework to obtain the standard motion trajectory model of the extreme scooter.

[0021] In the extreme scooter motion assisting method, preferably, the training process also includes a process of performing difference calibration using trajectory model results corresponding to different initial standard data:

[0022] Establish a piecewise loss function:

[0023] where y i represents the actual value, i.e., the real trajectory data of the extreme scooter collected by vision, f(x i ) is the trajectory value given by the trajectory model, δ is the threshold parameter, that is, the change point of the segmented loss function, and the value of δ is 4.

[0024] In the extreme scooter motion assisting method, preferably, in step S3, the process of determining the difference between the user motion trajectory and the standard trajectory model includes:

[0025] The cosine similarity between the user's motion trajectory and the standard trajectory model in the standard model library is calculated in the following steps:

[0026] Define the user motion trajectory data g(x j ), where x j is the n one-dimensional motion feature data collected during the user's motion, and f(x i ) is the trajectory value given by the trajectory model, and g(x j )、f(x i ) are discretized into numerical vectors:

[0027] F=[f(x1),f(x2),...,f(x m )], G=[g(x1),g(x2),...,g(x m )]

[0028] Then the cosine similarity between the two can be expressed as:

[0029]

[0030] The value of the cosine similarity is determined. If there is a standard trajectory model that makes the cosine similarity greater than 0.95, all the standard trajectory models that meet the requirements are given. If there is no standard trajectory model that makes the cosine similarity greater than 0.95, a standard trajectory model that maximizes the cosine similarity is given.

[0031] In the extreme scooter motion assisting method, preferably, the process of back-mapping at most m motion feature data in the feature data set of the user's motion process based on the difference comprises:

[0032] Computes the absolute value of the difference between two numeric vectors

[0033] Z=|FY|,

[0034] Vector Z is a vector composed of the absolute values ​​of the differences between the elements at corresponding positions in the two. Each element in Z is sorted by size to obtain a new vector:

[0035] Z'=(z1,z2,...,z m ,...,z n ),

[0036] The first m elements correspond to the m user motion feature data (x1, x2, ..., x m ), which is the at most m motion feature data.

[0037] In the extreme scooter motion assisting method, preferably, based on the inverse mapped at most m motion feature data, a specific user posture corresponding to the at most m motion feature data is obtained, and then the user is informed of motion posture adjustment suggestions based on the difference between the user motion trajectory corresponding to the specific user posture and the standard trajectory model.

[0038] In the extreme scooter motion assisting method, preferably, the motion characteristic data is collected by a speed sensor.

[0039] On the other hand, the present specification also provides an extreme scooter that records real-time motion status and executes the extreme scooter motion assistance method as described above.

[0040] The extreme scooter motion assisting method and the extreme scooter for recording real-time motion status of the present invention provide an extreme scooter that is easy for new users to learn. By building a database of standard movements of the extreme scooter and a trajectory prediction model, the trajectory prediction of the user's movement and the difference analysis compared with the standard movement can be realized, so that the user can independently learn and calibrate his own movements with the help of the extreme scooter, which greatly reduces the entry threshold of the sport and is also conducive to the promotion of the extreme scooter sport. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of this specification, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 A schematic diagram of the process of an extreme scooter motion assisting method provided according to an embodiment of the present application is shown;

[0043] Figure 2 A schematic diagram showing the steps of collecting a standard data set consisting of n one-dimensional standard data in a data collection process provided in an embodiment of the present application;

[0044] Figure 3 A schematic diagram showing the entire process of collecting motion trajectories and training models for an extreme scooter provided in an embodiment of the present application before leaving the factory is shown;

[0045] Figure 4 A schematic diagram of an extreme scooter provided in some embodiments of the present application is shown; DETAILED DESCRIPTION

[0046] The following description provides specific application scenarios and requirements of this specification, with the purpose of enabling those skilled in the art to make and use the contents of this specification. Various local modifications to the disclosed embodiments will be apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of this specification. Therefore, this specification is not limited to the embodiments shown, but to the widest scope consistent with the claims.

[0047] The terms used herein are only used for the purpose of describing specific example embodiments and are not restrictive. For example, unless the context clearly indicates otherwise, as used herein, the singular forms "a", "an" and "the" may also include plural forms. When used in this specification, the terms "include", "comprise" and / or "contain" mean that the associated integers, steps, operations, elements and / or components exist, but do not exclude the existence of one or more other features, integers, steps, operations, elements, components and / or groups or that other features, integers, steps, operations, elements, components and / or groups may be added in the system / method.

[0048] These and other features of the present specification, as well as the operation and function of the related elements of the structure, and the economy of combination and manufacture of the parts can be significantly improved in view of the following description. Reference is made to the accompanying drawings, all of which form a part of this specification. However, it should be clearly understood that the drawings are for illustration and description purposes only and are not intended to limit the scope of this specification.

[0049] The flowcharts used in this specification illustrate the operations implemented by the system according to some embodiments in this specification. It should be clearly understood that the operations of the flowcharts may not be implemented in sequence. On the contrary, the operations may be implemented in reverse order or simultaneously. In addition, one or more other operations may be added to the flowchart. One or more operations may be removed from the flowchart.

[0050] Figure 1 A schematic diagram of a process P100 of an extreme scooter motion assisting method provided according to an embodiment of the present application is shown. The process includes:

[0051] S110: Establish a standard motion database and build an extreme scooter motion trajectory model.

[0052] There are a large number of actions that have been developed and basically standardized by users in extreme scooter sports, such as Bunny Hop / Dolphin Jump / Bunny Jump (a basic jumping action, which is achieved by bending the knees and pushing the scooter with the feet), TailWhip / Dragon Tail Swing (this is an action in which the handlebars are quickly rotated by the rider's foot movements during the jump), G-turn / front wheel touches the ground and twists 180 degrees (in gliding, the front wheel touches the ground and twists 180 degrees, and the rider and scooter complete the turn at the same time), etc. In some embodiments of the present application, the establishment of a standard motion database comes from the motion capture of professional extreme scooter athletes, that is, the process of collecting data: first, let the professional athletes wear clothes with specific markers (Markers), and when the athletes perform some of the above-mentioned standard actions, multiple high-speed cameras are used to capture the markers on the target from different angles, and then the three-dimensional space coordinate data is obtained through computer processing, so that the complete motion trajectory of the athlete can be recorded. By collecting a large number of different sets of standard motion actions, a database of extreme scooter standard motion trajectories can be established. However, at the user end, it is unrealistic for the user to obtain his own motion trajectory and assist in learning by means of motion capture. Although the method of relying on multi-channel sensors to monitor the motion posture of the extreme scooter in the air in real time and output the final motion trajectory is feasible, it will inevitably increase the cost of the extreme scooter, and it is difficult to ensure that the multi-channel sensors can operate stably under the working conditions of the extreme scooter. In the embodiment of the present application, only one channel of sensor is used in conjunction with machine learning to solve this problem. That is, the extreme scooter standard motion trajectory database established in the previous text is used as a training set to train the extreme scooter motion trajectory model. The specific principle can be briefly described as: collecting n one-dimensional initial standard data during the athlete's movement to form a standard data set. The one-dimensional here means that the collected data is composed of the same type of data, so the one-dimensional initial standard data can be collected by one channel of sensor, which can be the speed, acceleration, angular momentum, etc. of a certain point on the extreme scooter during the athlete's movement. In theory, there is a certain mapping relationship between the standard data set and the standard motion trajectory database, so in theory, the deep learning method can also be used to obtain a mathematical model that can obtain the final trajectory by inputting the initial data. In some embodiments of the present application, in order to ensure the accuracy of the model, the collected initial data needs to be denoised.

[0053] Since the initial data such as velocity and acceleration are all vectors, the standard data set consisting of n one-dimensional standard data collected during the data collection process can be obtained by the following algorithm, referring to Figure 2 , including: determining an initial k value, corresponding to a point o on the complex plane, for n initial one-dimensional standard data x1, x2, ..., x n; Calculate the reachable distance: reach-dist(x,o)=max(dist(x,o),kd(o)), where dist(x,o) is the actual distance from point x to point o, and kd(o) is the k distance of point o (i.e. the distance from point o to its kth nearest neighbor); Calculate x i The k-nearest neighbor reachable density is: Then the decision threshold corresponding to each of the initial one-dimensional standard data is:

[0054] Judgment: If M(x i )>1, the point is a bad point and is removed from the standard data set;

[0055] If M(x i )≤1, the point is a normal point and is included in the standard data set.

[0056] It can be seen from the above algorithm that in the process of data collection, it determines the distance between each collected data point and its nearest neighbor data, that is, the density of each point is compared with the density of its neighbors (other surrounding points), so it is very suitable for excluding bad points in the vector data set. Taking the one-dimensional standard data as an example of speed, generally speaking, the movement speed of an extreme scooter will not exceed 30km / h, corresponding to a k value of less than 20, preferably, the k value is 15.

[0057] After obtaining the initial data that meets the requirements, different initial standard data x i is the independent variable, the visually collected extreme scooter standard trajectory data is the true value of the required prediction variable, and the extreme scooter trajectory model can be obtained by deep neural network (DNN) training with the help of PyTorch framework.

[0058] In some embodiments of the present application, the difference calibration can also be performed using similar actions and actual trajectories performed by athletes during the motion capture process, that is, different batches of initial motion data correspond to the same motion trajectory, and different initial motion data belonging to the same action are brought into the trajectory model, and the difference calibration is performed with the visual trajectory of motion capture, and the following algorithm is executed:

[0059] Establish a piecewise loss function:

[0060] where y i represents the actual value, i.e., the real trajectory data of the extreme scooter collected by vision, f(x i) is the trajectory value given by the trajectory model, δ is the threshold parameter, i.e., the change point of the segmented loss function. From the above algorithm, it can be seen that the value of δ determines the value of the above loss function. Since both relatively standard actions (small errors) and actions with large deviations (large errors) may occur in the movement of extreme scooters, the existence of the δ value can artificially adjust the sensitivity of this loss function to errors. In some embodiments of the present application, δ is preferably 4. The calibration process is a step to minimize the above loss function during the training process.

[0061] Figure 3 The entire process of collecting the motion trajectory and training the model of the extreme scooter before leaving the factory is shown.

[0062] It should be noted that, since the above steps can be completed in the manufacturer's testing phase and do not require the participation of specific users, the above steps can be completed before the extreme scooter is delivered to the user and stored in the cloud or stored on each extreme scooter shipped by means of a storage medium, and the user can call and use the motion trajectory model at any time. S120: Collect a feature data set of the user's motion, and obtain the user's motion trajectory based on the motion trajectory model.

[0063] In some embodiments of the present application, when a user uses an extreme scooter to perform an action, the extreme scooter collects characteristic data of the user's movement and forms a characteristic data set. Figure 4 An extreme scooter 100 described in some embodiments of the present application is shown, which at least includes a handlebar 110, a base plate 120, and a wheel 130. Taking the speed sensor as an example, in the action of the extreme scooter, at least the speed sensors need to be arranged at the handlebar 110, the base plate 120, and the wheel 130, corresponding to the three basic actions of "turning the handlebar", "adjusting the posture in the air", and "accelerating on the ground" and the motion state speed monitoring of the combination of the three basic actions. The inventors of the present application have found in long-term research that only relying on the speed sensor combined with the motion trajectory model described above, a complete motion trajectory can be obtained in real time during a motion action, and the accuracy of the trajectory depends on the number of sampling points.

[0064] In some embodiments of the present application, a feature data set consisting of m one-dimensional motion feature data of a single motion of a user is collected, and based on this feature data with the help of the model, the user's motion trajectory can be obtained. Since the user's trajectory data is usually used to compare with the standard trajectory data described above, m≤n.

[0065] S130: Determine the difference between the user's motion trajectory and the standard trajectory, and reflect it in the feature data of the user's motion.

[0066] After obtaining the user's motion trajectory, it is necessary to compare the motion trajectory with the standard trajectory to inform the user of the differences between his or her motion and the standard motion. However, considering the large number of standard motion trajectories in the standard motion trajectory database, it is necessary to find the standard trajectory that is closest to the user's single motion for comparison. The process is obtained by the following steps:

[0067] The cosine similarity between the user's motion trajectory and the standard trajectory model in the standard model library is calculated in the following steps:

[0068] Define the user motion trajectory data g(x j ), where x j is the m one-dimensional motion feature data collected during the user's motion, and f(x i ) is any trajectory value in the trajectory model library, and in the entire motion trajectory interval, both are discretized into numerical vectors: g(x j )、f(x i ) are all discretized into numerical vectors: F = [f(x1),f(x2),...,f(x m )], G=[g(x1),g(x2),...,g(x m )], which is equivalent to re-discretizing both trajectory functions into a set of m sets of mappings of independent variables and dependent variables, and using cosine similarity to determine the closeness of the user's single motion trajectory to any trajectory value in the trajectory model library, thereby determining the action the user wants to complete.

[0069] The cosine similarity between the two can be expressed as:

[0070]

[0071] The value of the cosine similarity is determined. If the two are exactly the same, the cosine similarity is 1 and the cosine distance is 0; if the two are completely opposite, the cosine similarity is -1 and the cosine distance is 2. In an embodiment of the present application, if it is determined that there is a standard trajectory model that makes the cosine similarity greater than 0.95, then all standard trajectory models that meet the requirements are given. If there is no standard trajectory model that makes the cosine similarity greater than 0.95, then a standard trajectory model that maximizes the cosine similarity is given. The reason for using cosine similarity is that this indicator pays more attention to the impact of the direction of the vector on the result, and is more applicable to the trajectory model of the present application that contains vector independent variables.

[0072] S140: Providing user movement posture adjustment suggestions based on the inverse mapping data.

[0073] After finding the standard motion trajectory model that is closest to the user's motion, it is necessary to find the source of the difference between the user's motion and the standard motion as a basis for assisting in calibrating the user's motion, that is, to reversely map the difference to at most m motion feature data in the feature data set during a single motion of the user, including the following algorithm:

[0074] Compute the absolute value of the difference between two numeric vectors:

[0075] Z=|FY|,

[0076] Vector Z is a vector composed of the absolute values ​​of the differences between the elements at corresponding positions in the two. Each element in Z is sorted by size to obtain a new vector:

[0077] Z'=(z1,z2,...,z m ,...,z n ),

[0078] The first m elements correspond to the m user motion feature data (x1, x2, ..., x m ), which is the at most m motion feature data. After the at most m motion feature data are obtained through inverse mapping, a specific user posture corresponding to the at most m motion feature data is obtained, and then the user is informed of motion posture adjustment suggestions based on the difference between the user motion trajectory corresponding to the specific user posture and the standard trajectory model.

[0079] Taking the Bunny Hop (jumping by bending the knees and pushing the scooter with the feet) as an example, a complete extreme scooter action auxiliary process P200 is:

[0080] S210: The user starts to move, and the sensor collects motion parameters of the user's motion in real time.

[0081] The process of a user performing a complete Bunny Hop action includes: accelerating by pushing off the ground, bending the knees to drive the jump, maintaining the posture in the air, decelerating and stopping after landing. Taking the speed sensor arranged on the extreme scooter as an example, the speed sensor will collect a set of speed data in at least each of the above four action intervals, and bring the collected speed data into the trajectory model to obtain the trajectory model of the user's current movement.

[0082] S220: Upload the current motion data to the standard motion trajectory database and complete motion comparison.

[0083] When the user issues an action assistance command, the extreme scooter will upload the trajectory model of the user's current movement obtained by the model to the standard motion trajectory database, and conclude that the user is performing a Bunny Hop action this time, and then analyze the difference between the user's action and the standard Bunny Hop action.

[0084] S330: Analyze the differences and provide exercise suggestions to the user.

[0085] After analysis, it is found that the biggest difference between the user's Bunny Hop motion trajectory and the standard Bunny Hop motion trajectory falls in the initial motion interval, that is, the ground-pushing acceleration stage. The corresponding motion characteristic data is the initial speed of the ground-pushing acceleration stage. The suggestion fed back to the user is: "It is recommended to push harder when pushing off the ground to obtain a larger initial speed."

[0086] The present application also provides a scooter that records real-time motion status, which monitors the user's motion status in real time with the help of a sensor set thereon and executes the extreme scooter motion assistance method described above. In order to achieve this purpose, the scooter that records real-time motion status is also provided with a user interaction interface, and the user can use the user interaction interface to execute the extreme scooter motion assistance method described above. In some embodiments, the scooter that records real-time motion status can be connected to the user's mobile terminal such as a mobile phone, a personal computer, etc. through Bluetooth protocol wireless communication technology, wireless local area network communication technology (Wi-Fi), etc., and the user can execute the extreme scooter motion assistance method described above through the mobile terminal.

[0087] In some possible implementations, various aspects of this specification may also be implemented in the form of a program product, which includes a program code. When the program product is run on a computer device, the program code is used to enable the computer program to execute a workshop visualization operation and maintenance method based on BIM technology described in this specification. The program product for implementing the above method may include program code using a portable compact disk read-only memory (CD-ROM) and may be run on a computer device or a server device. However, the program product of this specification is not limited to this. In this specification, the non-temporary storage medium may be any tangible medium containing or storing a program, which may be used by an instruction execution system or used in combination with it. The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage chip. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor system, device or device, or any combination of the above.

[0088] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only for example and does not constitute a limitation of this specification. Although not explicitly stated here, those skilled in the art may make various modifications, improvements and corrections to this specification. Such modifications, improvements and corrections are suggested in this specification, so such modifications, improvements and corrections still belong to the spirit and scope of the exemplary embodiments of this specification.

Claims

1. An extreme scooter motion assisting method, characterized in that: It includes the following steps: Step S1: establishing a standard motion database and building an extreme scooter motion trajectory model based on a standard data set consisting of n one-dimensional initial standard data and extreme scooter standard motion trajectory data collected by motion capture, the process includes: collecting data, establishing a database, training data and building a model; Step S2: collecting a feature data set consisting of m one-dimensional motion feature data of a single motion of the user, and obtaining the motion trajectory of the user based on the feature data with the aid of the model; Step S3: Determine the difference between the user's motion trajectory and the standard trajectory, and reversely map at most m motion feature data in the feature data set of the user's motion process based on the difference; Step S4: Providing user movement posture adjustment suggestions based on the feature data.

2. The extreme scooter motion assisting method according to claim 1, characterized in that: In step S1, the standard data set consisting of n one-dimensional standard data collected in the data collection process is obtained by the following algorithm: Determine an initial k value corresponding to a point o on the complex plane, For n initial one-dimensional standard data x1, x2, ..., x n , calculate the reachable distance: reach-dist(x,o)=max(dist(x,o),kd(o)); Among them, dist(x,o) is the true distance from point x to point o, and kd(o) is the distance from point o to its kth nearest neighbor; Calculate x i The k-nearest neighbor reachable density is: Then the decision threshold corresponding to each of the initial one-dimensional standard data is Judgment: If M(x i )>1, the point is a bad point and is removed from the standard data set; If M(x i )≤1, the point is a normal point and is included in the standard data set.

3. The extreme scooter motion assisting method according to claim 2, characterized in that: k＝15。 4. The extreme scooter motion assisting method according to claim 1, characterized in that: The training process uses a deep neural network for training, and different initial standard data x i is the independent variable, and the physical trajectory data y of the extreme scooter collected by vision i The extreme scooter motion trajectory model is obtained by performing deep neural network training with the help of PyTorch framework.

5. The extreme scooter motion assisting method according to claim 4, characterized in that: The training process also includes a process of performing difference calibration using trajectory model results corresponding to different initial standard data: Establish a piecewise loss function: where y i represents the actual value, i.e., the real trajectory data of the extreme scooter collected by vision, f(x i ) is the trajectory value given by the trajectory model, δ is the threshold parameter, that is, the change point of the segmented loss function, and the value of δ is 4.

6. The extreme scooter motion assisting method according to claim 1, characterized in that: In step S3, the process of determining the difference between the user's motion trajectory and the standard trajectory model includes: The cosine similarity between the user's motion trajectory and the standard trajectory model in the standard model library is calculated in the following steps: Define the user motion trajectory data g(x j ), where x j is the m one-dimensional motion feature data collected during the user's motion, and f(x i ) is any trajectory value in the trajectory model library, and g(x j )、f(x i ) are all discretized into numerical vectors: F = [f(x1),f(x2),...,f(x m )], G=[g(x1),g(x2),...,g(x m )]; Then the cosine similarity between the two can be expressed as: The value of the cosine similarity is determined. If there is a standard trajectory model that makes the cosine similarity greater than 0.95, all the standard trajectory models that meet the requirements are given. If there is no standard trajectory model that makes the cosine similarity greater than 0.95, a standard trajectory model that maximizes the cosine similarity is given.

7. The extreme scooter motion assisting method according to claim 6, characterized in that: The process of back-mapping at most m motion feature data in the feature data set of the user's current motion process based on the difference comprises: Computes the absolute value of the difference between two numeric vectors Z=|FY|, Vector Z is a vector composed of the absolute values ​​of the differences between the elements at corresponding positions in the two. Each element in Z is sorted by size to obtain a new vector: Z'=(z1,z2,...,z m ,...,With n ), The first m elements correspond to the m user motion feature data (x1, x2, ..., x m ), which is the at most m motion feature data.

8. The extreme scooter motion assisting method according to claim 7, characterized in that: According to the inversely mapped at most m motion feature data, a specific user posture corresponding to the at most m motion feature data is obtained, and then the user is informed of motion posture adjustment suggestions based on the difference between the user motion trajectory corresponding to the specific user posture and the standard trajectory model.

9. The extreme scooter motion assisting method according to any one of claims 1 to 8, characterized in that: The motion characteristic data is collected by a speed sensor.

10. An extreme scooter that records real-time motion status, characterized in that: Execute the extreme scooter motion assisting method as described in any one of claims 1-9.

Images