CN113486798A - Training plan making processing method and device based on causal relationship - Google Patents
Training plan making processing method and device based on causal relationship Download PDFInfo
- Publication number
- CN113486798A CN113486798A CN202110764535.6A CN202110764535A CN113486798A CN 113486798 A CN113486798 A CN 113486798A CN 202110764535 A CN202110764535 A CN 202110764535A CN 113486798 A CN113486798 A CN 113486798A
- Authority
- CN
- China
- Prior art keywords
- data
- athlete
- sports
- performance
- athletic
- 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.)
- Pending
Links
- 238000012549 training Methods 0.000 title claims abstract description 43
- 230000001364 causal effect Effects 0.000 title claims abstract description 23
- 238000003672 processing method Methods 0.000 title claims abstract description 8
- 230000033001 locomotion Effects 0.000 claims abstract description 52
- 238000000034 method Methods 0.000 claims abstract description 31
- 230000008859 change Effects 0.000 claims abstract description 10
- 230000000386 athletic effect Effects 0.000 claims description 47
- 230000037147 athletic performance Effects 0.000 claims description 19
- 230000001133 acceleration Effects 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 14
- 238000009472 formulation Methods 0.000 claims description 2
- 239000000203 mixture Substances 0.000 claims description 2
- 230000001149 cognitive effect Effects 0.000 description 38
- 230000015654 memory Effects 0.000 description 27
- 230000008569 process Effects 0.000 description 14
- 230000019771 cognition Effects 0.000 description 12
- 241000282414 Homo sapiens Species 0.000 description 10
- 230000008447 perception Effects 0.000 description 10
- 230000006870 function Effects 0.000 description 8
- 210000004556 brain Anatomy 0.000 description 7
- 238000012512 characterization method Methods 0.000 description 6
- 238000003745 diagnosis Methods 0.000 description 6
- 238000013507 mapping Methods 0.000 description 6
- 238000003860 storage Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 5
- 230000006399 behavior Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 230000006872 improvement Effects 0.000 description 4
- 238000003754 machining Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000009192 sprinting Effects 0.000 description 3
- 241000282412 Homo Species 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000010365 information processing Effects 0.000 description 2
- 230000007787 long-term memory Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000003278 mimic effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000003936 working memory Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 210000000577 adipose tissue Anatomy 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000000988 bone and bone Anatomy 0.000 description 1
- 230000036992 cognitive tasks Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000010924 continuous production Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 230000037149 energy metabolism Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000037361 pathway Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000009182 swimming Effects 0.000 description 1
- 230000001225 therapeutic effect Effects 0.000 description 1
- 230000008448 thought Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The application discloses a training plan making processing method and device based on causal relationship, the method comprises the following steps: acquiring first motion data, and comparing the first motion data with the second motion data; sequentially replacing the values of the data items with gaps between the first sports data and the second sports data with the values of the corresponding data items in the second sports data, and determining the score of the first athlete after each data item in the first sports data is replaced; determining the difference between the score obtained after each change of data and the score of the second athlete; determining data items that affect the first athlete's performance based on the gap; to determine a training program for the first athlete. The problem caused by the fact that a new training plan is made by depending on personal experience of a coach in the prior art is solved, the pertinence and the scientificity of the training plan are improved, and assistance is provided for improving the performance of athletes to a certain extent.
Description
Technical Field
The application relates to the field of sports, in particular to a training plan making and processing method and device based on causal relationship.
Background
In recent years, sports has greatly benefited from the application of artificial intelligence due to advances in data collection and processing power, advances in statistics, particularly deep learning, increases in computing resources, and an increasing amount of economic activity. However, in addition to detection, recognition and statistical analysis, there are more fundamental problems that motion artificial intelligence has to solve.
In the case of sports training, this is a continuous process between athletes (athletes, students or hobbyists) and their coaches (coaches, teachers or organizers) to guide their activities and organize their training sessions. While two intelligent sports training (1ST) systems are now available to assist athletes in different training sessions, how well a training is achieved depends on the coach's experience and personal abilities.
In particular, after the athlete trains to a certain degree, the performance of the athlete is slowly improved, and a new training plan needs to be made for the athlete. The creation of new training programs for athletes in the prior art relies on the personal experience of a coach. This, depending on the ability of the coach, may affect the performance improvement of the athlete.
Disclosure of Invention
The embodiment of the application provides a training plan making processing method and device based on causal relationship, so as to at least solve the problem caused by the fact that the prior art makes a new training plan by relying on personal experience of a coach.
According to one aspect of the application, a training plan making processing method based on causal relationship is provided, and the method is characterized by comprising the following steps: obtaining first athletic data obtained by a first athlete while performing an athletic performance, wherein the first athletic data includes at least one data item; comparing the first athletic data with pre-configured second athletic data obtained by a second athlete in performing the athletic performance, wherein the performance of the second athlete in the athletic performance is better than the performance of the first athlete in the athletic performance, and the second athletic data includes the same data items as the first athletic data; sequentially replacing the values of the data items with gaps between the values of the data items in the first sports data and the second sports data with the values of the corresponding data items in the second sports data, and determining the achievement of the first athlete after each replacement of one data item in the first sports data; determining a difference between the performance obtained after each change of one of the first athletic data and the performance of the second athlete; determining data items that affect the first athlete's performance based on the gap; determining a training program for the first athlete based on data items that affect the first athlete's performance.
Further, the motion data comprises at least one of: the sports data comprises the first sports data and/or the second sports data, and the sports parameters comprise physical sports values of the body and/or energy generated by the body when the athlete engages in the sports.
Further, the motion parameters include at least one of: velocity, acceleration, rotational velocity, frequency of body part motion, force generated by a body part.
Further, determining a data item that impacts the first athlete's performance based on the gap comprises: determining a most influential data item for the first athlete's performance based on the gap, wherein the most influential data item is: the difference in athletic performance between the first athlete and the second athlete is minimized after the data item in the first athletic data is replaced with the corresponding data item in the second athletic data.
Further, determining a data item that impacts the first athlete's performance based on the gap comprises: arranging data items influencing the first athlete performance according to the gap to obtain a queue; determining a training program for the first athlete based on data items that affect the first athlete's performance comprises: determining a training program for the first athlete based on a front predetermined number of data items in the queue.
According to another aspect of the present application, there is also provided a training plan making process apparatus based on causal relationship, including: the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring first motion data obtained by a first athlete during motion, and the first motion data comprises at least one data item; the comparison module is used for comparing the first sports data with second sports data which is pre-configured and obtained by a second athlete in the sports, wherein the performance of the second athlete in the sports is better than that of the first athlete in the sports, and the second sports data comprises the same data items as the first sports data; the replacing module is used for sequentially replacing the values of the data items with gaps with the second sports data in the first sports data with the values of the corresponding data items in the second sports data and determining the score of the first athlete after each data item in the first sports data is replaced; a first determining module for determining a difference between a performance obtained after each change of one of the first athletic data and a performance of the second athlete; a second determining module for determining data items that affect the first athlete's performance based on the gap; a third determination module to determine a training program for the first athlete based on data items that affect the first athlete's performance.
Further, the motion data comprises at least one of: the sports data comprises the first sports data and/or the second sports data, and the sports parameters comprise physical sports values of the body and/or energy generated by the body when the athlete engages in the sports.
Further, the motion parameters include at least one of: velocity, acceleration, rotational velocity, frequency of body part motion, force generated by a body part.
Further, the second determination module is configured to: determining a most influential data item for the first athlete's performance based on the gap, wherein the most influential data item is: the difference in athletic performance between the first athlete and the second athlete is minimized after the data item in the first athletic data is replaced with the corresponding data item in the second athletic data.
Further, the second determining module is used for arranging data items which have influence on the first athlete achievement according to the gap to obtain a queue; the third determination module is to determine a training program for the first athlete based on a front predetermined number of data items in the queue.
In an embodiment of the present application, obtaining first athletic data obtained by a first athlete while performing an athletic performance is employed, wherein the first athletic data includes at least one data item; comparing the first athletic data with pre-configured second athletic data obtained by a second athlete in performing the athletic performance, wherein the performance of the second athlete in the athletic performance is better than the performance of the first athlete in the athletic performance, and the second athletic data includes the same data items as the first athletic data; sequentially replacing the values of the data items with gaps between the values of the data items in the first sports data and the second sports data with the values of the corresponding data items in the second sports data, and determining the achievement of the first athlete after each replacement of one data item in the first sports data; determining a difference between the performance obtained after each change of one of the first athletic data and the performance of the second athlete; determining data items that affect the first athlete's performance based on the gap; determining a training program for the first athlete based on data items that affect the first athlete's performance. The problem caused by the fact that a new training plan is made by depending on personal experience of a coach in the prior art is solved, the pertinence and the scientificity of the training plan are improved, and assistance is provided for improving the performance of athletes to a certain extent.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application. In the drawings:
FIG. 1 is a flow chart of a causal relationship-based training plan formulation process according to an embodiment of the present application.
FIG. 2 is a schematic diagram of a causal cognitive computation framework for locomotion according to an embodiment of the present application.
Detailed Description
It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict. The present application will be described in detail below with reference to the embodiments with reference to the attached drawings.
It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer-executable instructions and that, although a logical order is illustrated in the flowcharts, in some cases, the steps illustrated or described may be performed in an order different than presented herein.
In the present embodiment, a causal relationship-based training plan making processing method is provided, and fig. 1 is a flowchart according to an embodiment of the present application, as shown in fig. 1, the method includes the following steps:
step S102, acquiring first motion data obtained by a first athlete during motion, wherein the first motion data comprises at least one data item;
in this step, the motion data may include at least one of: the sports data comprises the first sports data and/or the second sports data, and the sports parameters comprise physical sports values of the body and/or energy generated by the body when the athlete engages in the sports. The motion parameters may include at least one of: velocity, acceleration, rotational velocity, frequency of body part motion, force generated by a body part.
Step S104, comparing the first sports data with second sports data which is pre-configured and obtained by a second athlete in the sports, wherein the performance of the second athlete in the sports is better than that of the first athlete in the sports, and the data items of the second sports data are the same as those of the first sports data;
step S106, sequentially replacing the values of the data items with gaps with the second sports data in the first sports data with the values of the corresponding data items in the second sports data, and determining the score of the first athlete after each replacement of one data item in the first sports data;
step S108, determining the difference between the achievement obtained after each change of one item of data in the first athletic data and the achievement of the second athlete;
step S110, determining data items influencing the first athlete achievement according to the gap;
step S112, determining a training program for the first athlete based on the data item that affects the first athlete' S performance.
There are various ways to determine the training program in steps S110 and S112, for example, the most influential data item on the first athlete' S performance may be determined according to the gap, wherein the most influential data item is: the difference in athletic performance between the first athlete and the second athlete is minimized after the data item in the first athletic data is replaced with the corresponding data item in the second athletic data. For another example, data items that affect the first athlete's performance may also be ranked into a queue based on the gap; a training program for the first player is then determined based on the front predetermined number of data items in the queue.
Through the steps, the problem caused by the fact that a new training plan is made by depending on personal experience of a coach in the prior art is solved, the pertinence and the scientificity of the training plan are improved, and help is provided for improving the performance of athletes to a certain extent.
In this embodiment, the following steps may also be taken to match the second athlete:
step S202, acquiring a physiological parameter of a first athlete, wherein the physiological parameter is used for indicating a physiological characteristic of the first athlete, and the difference is used as a basis for improving the achievement of the first athlete;
in this step, the physiological parameter may be various, and any physiological parameter related to the sport performed by the first athlete may be used for the determination, for example, the physiological parameter may include at least one of the following: sex, height, weight, age, body fat rate.
Step S204, acquiring a second athlete matched with the physiological parameters of the first athlete, wherein the first athlete and the second athlete are engaged in the same sport;
the physiological parameters can be matched in various ways, for example, the physiological parameters of at least one athlete are acquired and stored in advance; and finding the athletes with the same number or the smallest difference with the physiological parameters of the first athlete as the second athlete from the physiological parameters of the at least one athlete. In this alternative embodiment, a plurality of physiological parameters of an athlete can be converted into relative values, and then the value corresponding to the athlete is calculated according to the weight value and the relative value corresponding to each physiological parameter, and the value of a second athlete matched with the first athlete is within a predetermined range from the value corresponding to the first athlete. If there are a plurality of second athletes within the predetermined range, the athlete with the smallest difference in height and weight may be selected as the second athlete.
Step S206, acquiring first movement data obtained by the first athlete during the movement;
step S208, comparing the first sport data with second sport data which is configured in advance and obtained by the second athlete in the sport, wherein the performance of the second athlete in the sport is better than the performance of the first athlete in the sport;
the first motion data and the second motion data in the above steps S206 and S208 may be collectively referred to as motion data, and the motion data may include at least one of: the motion posture of the athlete and the motion parameters of the athlete during the motion. The exercise parameter may be a physical exercise value of the body and/or energy generated by the body when the athlete is engaged in the exercise. For example, speed and acceleration during running, swing arm speed during swimming, spin speed during throwing of a hammer ball, weight lifting, muscle strength generated, etc.
And step S210, determining the difference between the first athlete and the second athlete in the process of performing the sports according to the comparison result.
In this step, as an alternative embodiment, the difference of the motion parameters can be directly compared. Alternatively, in another alternative embodiment, where the athletic data includes athletic poses of the athlete, the obtaining of the athletic poses obtained by the first athlete and/or the second athlete while exercising may include: shooting the first athlete and/or the second athlete by a camera to obtain a video file during the sports; extracting the athletic maneuver of the first athlete and/or the second athlete from the video file.
In the following examples, which are described in connection with sprinting, a first player is a champion and a second player is a champion. In this embodiment, a causal cognitive computing framework (alternatively referred to as a system or platform) is proposed to mimic the cognitive process of a human trainer in motor training.
The system in this embodiment may be composed of 4 computing modules: perception, cognitive characterization, knowledge recall, and planning diagnosis (the parts that implement these four functions are referred to as "modules" or "sections" in the system of the present embodiment). The first of these four modules is used to obtain athletic data, while the other 3 modules are used to compare and process athletic data based on the athlete's athletic data, and the last 2 modules are inferred by causality.
In this embodiment, the data of elite athletes as knowledge rather than preset rules is referred to as "champion model". To make the determination, the motion characteristic parameters of the body of the user (e.g., athlete) are compared with the knowledge and then the portions of the user that need improvement are obtained. Based on this framework, in the embodiment, by an IST sprint system, the motion characteristic parameters of Elite athletes in the process of 100m competition, the system simulates the thinking cognition information processing mode of a coach without manually setting the parameters or the pre-designed rules by a user. The present embodiment will be explained below.
In this embodiment the exercise training can be divided into 4 phases including planning, implementation, control and evaluation.
A program is an objective to be achieved by training, wherein the objective may be athletic performance achieved after completion of training within a certain time. In an exercise system, the goal is typically set manually by the user, and then the exercise schedule is presented after judgment by the remote trainer, or by the system.
The training execution phase is a phase in which the prepared training is executed on the athlete. The primary role of the trainer at this stage is to monitor the exercises and collect data. For example, sensors or cameras may be used, and more monitoring devices may be incorporated in the system of the present embodiment as needed to acquire data.
Control is the comparison of the practice actually performed by the athlete with the planned practice. This stage includes extensive analysis, for example, analysis and identification of athletic movements may be performed by a computer, and then the planned athletic information may be compared to the actual athlete's performance. This stage is improved in this embodiment by cognitive characterization.
The evaluation is used to measure the performance of the athlete. The output of this phase will be the basis for the next training cycle from the plan. The evaluation can use a machine learning evaluation model trained by big data to analyze and evaluate the performance of the athlete and output a training plan of the next stage.
Cognitive computations are involved in the control phase, and the cognitive characterization can be processed using a cognitive model. The cognition model is applied to the real world and can be divided into two parts of object cognition and thing cognition. An object is an objectively stored entity, can be seen, touched and felt, is abstract, and is difficult to recognize relative to the object.
In the cognition model, the WHAT is a term for thinking cognition of everything and things by human beings, is the most recognized real world, is the ultimate true theory, law and principle, and has uniqueness. In order to achieve a certain expected result, human beings need to use the rules summarized by the predecessors to complete different transactions through innovation, so as to achieve the expected result, which is called "HOW" in the cognitive model, and is a means, a method and the like for realizing the method, and the like, and the method has diversity. To improve or solve the problem based on the real situation that has occurred, it is necessary to analyze with known experience and rules to find out the possible cause of the situation, which is called "WHY" in the cognitive model, and it is a verification of the rules summarized by the hypothesis.
The cognitive model is focused on intelligent processing of mass data: sensing, expressing, memorizing, judging, and acting. There are many current cognitive models, such as the ACT-R system and the Soar system.
ACT-R (Adaptive Character of four-dimensional) is a cognitive behavior system structure, is a theoretical model about human cognitive mechanism, aims to finally disclose human organization knowledge and generate thinking and movement rules of intelligent behaviors, and is verified based on neurobiology research results. The ACT-R is seemingly similar to a programming language platform, and the platform is constructed based on the results of many psychological studies, but the model constructed based on the ACT-R reflects the cognitive behaviors of human beings. The ACT-R realizes the construction of a cognitive model of a specific task through programming, researchers use a cognitive theory built in the ACT-R and necessary hypothesis and knowledge description of the specific task to construct the cognitive model of the specific task, the effectiveness of the model is verified through comparing a model result and an experimental result, and the ACT-R releases a version 6.0 at present and realizes the support of different system operation platforms. ACT-R is a cognitive architecture that is used to simulate and understand the theory of human cognition. ACT-R attempts to understand how humans organize knowledge and produce intelligent behavior. The goal of ACT-R is to enable the system to perform various cognitive tasks for humans, such as capturing human perception, thought, and behavior. The basic framework of ACT-R consists of a series of modules, each of which handles a specific type of information. For example, the vision module identifies objects in the vision zone, the hand motion control module and the statement module retrieve information from memory, and the target module tracks current targets and intentions. The modules act through a central generative system in concert. The centrally-generated system is not perceived by other active modules, but only reacts to the information stored in the caches of these modules.
SOAR is a framework developed by Newwell et al in 1986 called "Universal Intelligence", the letters of SOAR mean states, operators and results, simply applying operators to change states and produce results. SOAR mainly discusses the problems of knowledge, thinking, intelligence, memory, etc., and is a cognitive structure with a very wide application range. The SOAR model is a general problem solving program, based on the knowledge block theory, by utilizing the memory based on the rules, the search control knowledge and the operational characters are obtained, so that the SOAR model can learn from the experience, can remember how to solve the problem, and uses the experience and the knowledge in the subsequent problem solving process to realize the general problem solving. The SOAR consists of only a single long-term memory, encoded as a production rule, and a working memory, encoded as a graph structure. The evaluation of the intelligent agent on the current environment and condition is stored based on the working memory of the symbol, the next operation is selected by utilizing the recall related knowledge in the long-term memory through the decision cycle of inputting, state description, proposing an operator, comparing an operator, selecting an operator, applying an operator and outputting until the target state is reached.
The two systems Soar and ACT-R are dedicated to computer vision and natural language processing.
In the system of the present embodiment, there are two aspects to cognitive computation: coaches and athletes. The cognitive results of the coach are acted upon by the athlete, which is hard to cover by Soar and ACT-R.
Another computational cognition model (PMJ model for short) of Perception, Memory, and comment was proposed later in 2014, which integrates the cognitive mechanism and computability aspects in a unified framework, and consists of three stages and three pathways. The cognitive computation model generalizes the cognitive process into perception (perception), memory (memory) and judgment (Judgment), and defines a three-stage and multi-path processing framework combining cognition and computation corresponding to analysis, modeling and decision of a computation flow.
The traditional emotion calculation method is generally classic parameter calculation, and the main research idea can be summarized into a two-layer structure of physical characteristics- > target semantics. At the physical feature level, researchers are mainly concerned about accuracy and effectiveness of feature screening and feature extraction. On the basis, the machine learning algorithm helps to realize the associated mapping from the extracted features to the target semantics, so as to realize modeling. When the prediction result of the established model on the data set for the target semantics is not ideal, the method usually adopted is to continuously replace the algorithm of feature extraction or associated modeling until a more ideal prediction result is obtained. To improve this process, one of the important elements is to characterize (representation) the received information. The characterization is the basis of the communication between the brain and the objective world, and when the state of the external world and the characterization symbol in the brain have the same structural mapping, the brain can further output the result of information processing calculation. The human brain extracts the representation from the instant signal in a computable way, and then carries out semantic judgment according to the representation, and the process provides an important hint for improving a classic parameter calculation framework with a two-layer structure.
The whole process of brain cognition can be summarized into three stages of perception, memory and judgment. In a perception stage, the brain extracts a represented symbol based on the acquired information such as vision, hearing and the like; in the memory stage, the representation symbols are subjected to structure mapping or comparison with the representation symbols stored in the brain; and in the judging stage, outputting subjective semantics according to the mapping or comparison result. The core idea of the PMJ cognitive computation model is to correspond the three stages of perception, memory and judgment to the analysis, modeling and decision of a computation flow. By means of the PMJ model, a logic calculation framework which is improved by fusing subjective cognitive elements can be defined, namely a three-layer framework structure of physical characteristics I, cognitive representation I and target semantics I is formed. In the perception stage of the machine, for different modal information such as input text, video, audio and the like, researching an extraction method of cognitive representation based on a cognitive psychology principle; in the memory stage, establishing association mapping between cognitive representation and semantic description; and in the judging stage, outputting a final target result according to the association mapping. Compared with the traditional two-layer structure, the improved framework structure has the research difficulty of defining the cognitive representation directly related to the target semantics and realizing the automatic extraction of the cognitive representation.
In the embodiment, basic analysis is carried out through a sensing stage of a PMJ model, and characteristics such as telemechanical speed, motion attitude and the like are extracted; the memory stage evaluates the performance based on cognitive representation and abstract knowledge; and a judging stage for making a training plan time table. The fast processing path of the PMJ model may mimic the intuition in training; the fine machining path (i.e., the machining path in FIG. 1) is the primary mode of thinking for the trainer; and the feedback processing path may update the knowledge.
The embodiment can be applied to the sprint project, and the same can be applied to other projects. The following description will be made by taking a sprint item as an example.
Due to the fast speed of movement and the high precision requirements in the sprint program, the movement of the athlete cannot be tracked by the wearable sensor as in the popular sports or the long distance running program.
In this embodiment, the sensor system can be used to record the split time of the run with an accuracy on the order of 0.01 seconds or with a high number of frames per second camera for slice pose estimation in a relatively small range.
Although different sensors or motion capture algorithms output different forms of data, the athlete and coach experience the same in the continuous sprint process: first accelerating, then maintaining the maximum speed for a period of time, and finally inevitably decelerating. Such "tempo" is also one of the main concerns of sprint training. Thus, modeling the modal-independent cognitive perception of speed will help the system to more closely approximate the cognitive processing of a human coach. The speed-time curve of a sprint is modeled throughout the acceleration phase of sprinting by using a single exponential function. Thus, the performance of a sprint is determined by 2 parameters: maximum speed and acceleration time constant.
In this embodiment, a champion model (alternatively referred to as champion mode) is also used. The champion model is explained below. The concept of the champion model is to find an elite athlete similar to the athlete's height, weight, age, or other physical condition, and to simulate his training program or technique.
In this embodiment, causal reasoning is used in the planning process to make intelligent decisions, and the causal reasoning combines mathematical tools of "causal relationship" and "reasoning". In this embodiment "causal reasoning" is introduced into intelligent decisions in sports training. In three causal levels, intervention may specify true causal relationships between training and performance, rather than statistical correlations. Legal facts may quantify the potential therapeutic effect of an unfinished training program, which is more attractive for athletic training.
This embodiment also relates to memory networks, and in sports, most decision making studies follow the assumption that decision making and perceptual judgment are based on an internalized knowledge structure operating as an inference engine to consider the "best" decision or the decision that is the "best-fit" task. The memory network is named by its external memory, which includes the neural turing machine. The algorithm of the Neuropter combines content-based and location-based addressing.
The system of the present embodiment will be described with reference to fig. 2.
FIG. 2 is a schematic diagram of a causal awareness computation framework for motion according to an embodiment of the present application, as shown in FIG. 2, the framework including the following modules:
a sensing portion for sensing signals read from sensors, cameras, kinects, etc. to calculate physical speed, bones, etc. in different ways.
And the cognitive representation part is used for converting the physical signal into a cognitive representation.
Knowledge recall is used to select an optimal "knowledge" that is "bounded" from the "memory" based on causal relationships between key features and performance.
And the plan diagnosis part is used for evaluating different training plans through counterfactual according to the knowledge.
There are also two paths for passing information between the different modules, the first of which constitutes a fine machining path (i.e., a machining path) for all the above processes, which is the primary mode of thinking for the trainer. The fast processing path mimics the training intuition, and the feedback processing path can update knowledge.
Extraction of cognitive characterization
The speed-time curve of a sprint is modeled throughout the sprint acceleration phase using the following single exponential function:
this function requires 2 input parameters: maximum horizontal speed V of sprint runnermaxAnd acceleration constant T1. The monotonically increasing single exponential function can only model the acceleration phase of sprinting. To model all 3 phases of a 100 meter sprint, a new function is provided in this embodiment, in which the trend of the relative loss of speed can be expressed by equation (2), where t2 is the deceleration time constant, the larger the value, the slower the speed decrease, and the longer the deceleration process:
from the viewpoint of the kinetic energy metabolism, the acceleration tendency and the deceleration tendency do not act solely on different phases of the competition, but throughout the whole competition. Then the prototype of the entire time-velocity function should be as shown in (3):
the time variable and the acceleration and deceleration time constant are in units of seconds, and the speed variable and the speed are in units of meters per second. VidealIs an ideal limit value which can last for an indefinite period of time without a deceleration condition, and is not appropriate in the conventional document called "maximum speed", and is set to the ideal maximum speed V in the present embodimentidealAnd (4) showing. Modality-independent cognitive features of velocity are thus extracted.
Causal memory network
The champion model is stored in an external memory as knowledge, and the storage structure is as follows:
position: the champion model is located based on the performance recorded into the memory network, since it can only learn from fast-running athletes.
Key (Key): the unchangeable characteristics of the category, height, age, etc. are used as the key to select the champion model in the present embodiment.
Value (Value): the features that can be changed by training are used as Value for counterfactual diagnosis.
Counterfactual diagnosis
In a simplified manner, the DAG under diagnosis contains the champion model and the memory value in performance, i.e. the ideal speed VidealAn acceleration constant T1, and a deceleration constant T2. Then, counter-fact diagnosis is performed, for example, to find the difference between the ideal maximum speed between the athlete and the champion, the acceleration time parameter and the deceleration time constant, and to calculate whether the performance is consistent with the champion, assuming that one of the values is not different. In this way, it is possible to determine which parameter results in a performance that is lower than that of the champion, and then to develop a training program based on the improvement of that parameter.
In this embodiment, an electronic device is provided, comprising a memory in which a computer program is stored and a processor configured to run the computer program to perform the method in the above embodiments.
The programs described above may be run on a processor or may also be stored in memory (or referred to as computer-readable media), which includes both non-transitory and non-transitory, removable and non-removable media, that implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
These computer programs may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks, and corresponding steps may be implemented by different modules. The present embodiment provides an apparatus comprising at least one module, wherein the at least one module is configured to perform the steps of the method described above. The device is called a training plan making processing device based on causal relationship, and comprises: the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring first motion data obtained by a first athlete during motion, and the first motion data comprises at least one data item; the comparison module is used for comparing the first sports data with second sports data which is pre-configured and obtained by a second athlete in the sports, wherein the performance of the second athlete in the sports is better than that of the first athlete in the sports, and the second sports data comprises the same data items as the first sports data; the replacing module is used for sequentially replacing the values of the data items with gaps with the second sports data in the first sports data with the values of the corresponding data items in the second sports data and determining the score of the first athlete after each data item in the first sports data is replaced; a first determining module for determining a difference between a performance obtained after each change of one of the first athletic data and a performance of the second athlete; a second determining module for determining data items that affect the first athlete's performance based on the gap; a third determination module to determine a training program for the first athlete based on data items that affect the first athlete's performance.
Since the modules in the apparatus correspond to the steps in the method described above, the description has been already made in the method embodiment, and no further description is given here.
For example, the second determining module is configured to: determining a most influential data item for the first athlete's performance based on the gap, wherein the most influential data item is: the difference in athletic performance between the first athlete and the second athlete is minimized after the data item in the first athletic data is replaced with the corresponding data item in the second athletic data. For another example, the second determining module is configured to queue data items that affect the first athlete's performance according to the gap; the third determination module is to determine a training program for the first athlete based on a front predetermined number of data items in the queue.
The above are merely examples of the present application and are not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (10)
1. A training plan making processing method based on causal relationship is characterized by comprising the following steps:
obtaining first athletic data obtained by a first athlete while performing an athletic performance, wherein the first athletic data includes at least one data item;
comparing the first athletic data with pre-configured second athletic data obtained by a second athlete in performing the athletic performance, wherein the performance of the second athlete in the athletic performance is better than the performance of the first athlete in the athletic performance, and the second athletic data includes the same data items as the first athletic data;
sequentially replacing the values of the data items with gaps between the values of the data items in the first sports data and the second sports data with the values of the corresponding data items in the second sports data, and determining the achievement of the first athlete after each replacement of one data item in the first sports data;
determining a difference between the performance obtained after each change of one of the first athletic data and the performance of the second athlete;
determining data items that affect the first athlete's performance based on the gap;
determining a training program for the first athlete based on data items that affect the first athlete's performance.
2. The method of claim 1, wherein the motion data comprises at least one of: the sports data comprises the first sports data and/or the second sports data, and the sports parameters comprise physical sports values of the body and/or energy generated by the body when the athlete engages in the sports.
3. The method of claim 2, wherein the motion parameters comprise at least one of: velocity, acceleration, rotational velocity, frequency of body part motion, force generated by a body part.
4. The method of any one of claims 1 to 3, wherein determining a data item that impacts the first athlete's performance based on the gap comprises:
determining a most influential data item for the first athlete's performance based on the gap, wherein the most influential data item is: the difference in athletic performance between the first athlete and the second athlete is minimized after the data item in the first athletic data is replaced with the corresponding data item in the second athletic data.
5. The method according to any one of claims 1 to 3,
determining data items that affect the first athlete's performance based on the gap comprises: arranging data items influencing the first athlete performance according to the gap to obtain a queue;
determining a training program for the first athlete based on data items that affect the first athlete's performance comprises: determining a training program for the first athlete based on a front predetermined number of data items in the queue.
6. A causal relationship-based training plan formulation processing apparatus, comprising:
the system comprises an acquisition module, a processing module and a display module, wherein the acquisition module is used for acquiring first motion data obtained by a first athlete during motion, and the first motion data comprises at least one data item;
the comparison module is used for comparing the first sports data with second sports data which is pre-configured and obtained by a second athlete in the sports, wherein the performance of the second athlete in the sports is better than that of the first athlete in the sports, and the second sports data comprises the same data items as the first sports data;
the replacing module is used for sequentially replacing the values of the data items with gaps with the second sports data in the first sports data with the values of the corresponding data items in the second sports data and determining the score of the first athlete after each data item in the first sports data is replaced;
a first determining module for determining a difference between a performance obtained after each change of one of the first athletic data and a performance of the second athlete;
a second determining module for determining data items that affect the first athlete's performance based on the gap;
a third determination module to determine a training program for the first athlete based on data items that affect the first athlete's performance.
7. The apparatus of claim 6, wherein the motion data comprises at least one of: the sports data comprises the first sports data and/or the second sports data, and the sports parameters comprise physical sports values of the body and/or energy generated by the body when the athlete engages in the sports.
8. The apparatus of claim 7, wherein the motion parameters comprise at least one of: velocity, acceleration, rotational velocity, frequency of body part motion, force generated by a body part.
9. The apparatus of any of claims 6-8, wherein the second determining module is configured to:
determining a most influential data item for the first athlete's performance based on the gap, wherein the most influential data item is: the difference in athletic performance between the first athlete and the second athlete is minimized after the data item in the first athletic data is replaced with the corresponding data item in the second athletic data.
10. The apparatus according to any one of claims 6 to 8,
the second determining module is used for arranging the data items which have influence on the first athlete achievement according to the gap to obtain a queue;
the third determination module is to determine a training program for the first athlete based on a front predetermined number of data items in the queue.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110764535.6A CN113486798A (en) | 2021-07-07 | 2021-07-07 | Training plan making processing method and device based on causal relationship |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110764535.6A CN113486798A (en) | 2021-07-07 | 2021-07-07 | Training plan making processing method and device based on causal relationship |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113486798A true CN113486798A (en) | 2021-10-08 |
Family
ID=77941363
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110764535.6A Pending CN113486798A (en) | 2021-07-07 | 2021-07-07 | Training plan making processing method and device based on causal relationship |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113486798A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1560758A (en) * | 2004-02-25 | 2005-01-05 | 炜 蒋 | Training method for raising resulf of medium-and-long distance running by computer |
| US20130040272A1 (en) * | 2011-05-19 | 2013-02-14 | Trilife Eternal, Llc | System and method for producing customized training plans for multi-discipline endurance athletic competitions |
| EP2824614A1 (en) * | 2013-07-12 | 2015-01-14 | IQO2 bvba | Platform for planning and analyzing sports training for one or more athletes |
| US20160260340A1 (en) * | 2015-02-07 | 2016-09-08 | Alan Discount | Automated sports training workflow, delivery and tracking system |
| JP2017006192A (en) * | 2015-06-17 | 2017-01-12 | セイコーエプソン株式会社 | Advice generation method, advice generation program, and advice generation system |
| CN106377892A (en) * | 2016-09-27 | 2017-02-08 | 深圳市宏电技术股份有限公司 | Method and system for dart motion identification and training based on wearable equipment |
| CN111415720A (en) * | 2020-03-17 | 2020-07-14 | 中山优感科技有限公司 | Training auxiliary method and device based on multi-data acquisition |
| WO2021016426A1 (en) * | 2019-07-23 | 2021-01-28 | Monster Smart Solutions, Llc | Physical training system with machine learning-based training programs |
| CN112364785A (en) * | 2020-11-13 | 2021-02-12 | 中移雄安信息通信科技有限公司 | Exercise training guidance method, device, equipment and computer storage medium |
-
2021
- 2021-07-07 CN CN202110764535.6A patent/CN113486798A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1560758A (en) * | 2004-02-25 | 2005-01-05 | 炜 蒋 | Training method for raising resulf of medium-and-long distance running by computer |
| US20130040272A1 (en) * | 2011-05-19 | 2013-02-14 | Trilife Eternal, Llc | System and method for producing customized training plans for multi-discipline endurance athletic competitions |
| EP2824614A1 (en) * | 2013-07-12 | 2015-01-14 | IQO2 bvba | Platform for planning and analyzing sports training for one or more athletes |
| US20160260340A1 (en) * | 2015-02-07 | 2016-09-08 | Alan Discount | Automated sports training workflow, delivery and tracking system |
| JP2017006192A (en) * | 2015-06-17 | 2017-01-12 | セイコーエプソン株式会社 | Advice generation method, advice generation program, and advice generation system |
| CN106377892A (en) * | 2016-09-27 | 2017-02-08 | 深圳市宏电技术股份有限公司 | Method and system for dart motion identification and training based on wearable equipment |
| WO2021016426A1 (en) * | 2019-07-23 | 2021-01-28 | Monster Smart Solutions, Llc | Physical training system with machine learning-based training programs |
| CN111415720A (en) * | 2020-03-17 | 2020-07-14 | 中山优感科技有限公司 | Training auxiliary method and device based on multi-data acquisition |
| CN112364785A (en) * | 2020-11-13 | 2021-02-12 | 中移雄安信息通信科技有限公司 | Exercise training guidance method, device, equipment and computer storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 张峰;: "科学的训练计划对提高运动成绩的研究", 科技风, no. 21 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11367364B2 (en) | Systems and methods for movement skill analysis and skill augmentation | |
| CN106956271B (en) | Predict the method and robot of affective state | |
| Novatchkov et al. | Fuzzy logic in sports: a review and an illustrative case study in the field of strength training | |
| US20230082953A1 (en) | System and Method for Real-Time Interaction and Coaching | |
| Bačić | Predicting golf ball trajectories from swing plane: An artificial neural networks approach | |
| Shaker et al. | Towards player-driven procedural content generation | |
| CN113457108B (en) | Cognitive characterization-based exercise performance improving method and device | |
| Ma et al. | [Retracted] Recognition of Badminton Shot Action Based on the Improved Hidden Markov Model | |
| US20210245005A1 (en) | Implementation of machine learning for skill-improvement through cloud computing and method therefor | |
| Yan et al. | A review of basketball shooting analysis based on artificial intelligence | |
| Zhang et al. | Transforming sports training through the integration of internet technology and artificial intelligence | |
| CN113486798A (en) | Training plan making processing method and device based on causal relationship | |
| CN113343774B (en) | Fine-grained engineering mechanics diving action simulation and evaluation method | |
| Wen et al. | Modern technology, artificial intelligence, machine learning and internet of things based revolution in sports by employing graph theory matrix approach | |
| CN113476833A (en) | Game action recognition method and device, electronic equipment and storage medium | |
| Yamagiwa et al. | Skill grouping method: Mining and clustering skill differences from body movement BigData | |
| Olsen | Detecting Human Emotions Using Smartphone Accelerometer Data | |
| Zielinski | Live Perception and Real Time Motion Prediction with Deep Neural Networks and Machine Learning | |
| Han et al. | Artificial Intelligence Technologies and Their Application for Reform and Development of Table Tennis Training in Complex Environments | |
| Acikmese et al. | Artificially intelligent assistant for basketball coaching | |
| Hao | Research on the Applications of Artificial Intelligence in Golf | |
| KR102637528B1 (en) | Method and service platform for performance improvement | |
| Ferreira | Classification of table tennis strokes using a wearable device and deep learning | |
| Jiang et al. | Discrete Dynamic Modeling Analysis of Badminton Games Based on Viterbi Algorithm in College Badminton Physical Education | |
| Zahra | Effects of Emotions in Cognitive Based Game |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20211008 |