Method for analyzing a player's game in a racket sport

Abstract

The invention relates to a computer-implemented method (100) for analyzing a game of a player (300) in a racket sport, comprising: receiving (101) sensor data (401) from an accelerometer (403), wherein the accelerometer (403) is arranged on an arm (301) of the player (300); Analyzing (103) the sensor data (401) of the accelerometer (403) by an analysis module (200) and detecting player movements (300) based on the sensor data (401); Determine (105) based on the detected movements of the player (300) at least one game parameter of the player (300); and Display (107) of the determined game parameter.

Description

[0001] The present invention relates to a method for analyzing a game of a player of a racket sport. State of the art

[0002] Methods for analyzing a game of a player in a racket sport are known from the prior art.

[0003] It is an object to provide an improved method for analyzing a game of a player of a racquet sport.

[0004] The problem is solved by the method of the independent claim. Advantageous embodiments are the subject of the dependent claims.

[0005] According to one aspect, there is provided a computer-implemented method for analyzing a game of a player of a racket sport, comprising: Receiving sensor data from an acceleration sensor, wherein the acceleration sensor is arranged on an arm of the player; Analyzing the sensor data of the accelerometer by an analysis module and detecting movements of the player based on the sensor data; Determining at least one game parameter of the player based on the detected movements of the player; and Display the determined game parameter.

[0006] This provides the technical advantage of providing an improved method for analyzing a player's game in a racket sport. The claimed method can be executed exclusively on sensor data from an acceleration sensor and does not require additional information, for example, from another gyroscope sensor. Instead, only the sensor data from the acceleration sensor is analyzed by a suitably designed analysis module, and based on the sensor data from the acceleration sensor, the player's movements during the badminton game are determined or recognized. Based on the detected movements, at least one of the player's game parameters is determined and can, if necessary, be displayed to the player or another person.

[0007] According to one embodiment, the game parameter includes: stroke type, stroke strength, number of strokes in a rally, player mobility, playing style.

[0008] This provides the technical advantage that relevant parameters of the badminton game can be determined using the method according to the invention based on the sensor data from the acceleration sensor. This allows a precise analysis of the player to be performed.

[0009] According to one embodiment, the analysis module comprises a first sub-analysis module and a second analysis module, wherein the analysis of the sensor data comprises: Detecting, by the first sub-analysis module based on the sensor data, whether the movement performed by the player is a hitting movement; and Detection by the second sub-analysis module based on the sensor data of the stroke type of the stroke motion performed by the player if a stroke motion was detected by the first sub-analysis module.

[0010] This provides the technical advantage that the first sub-analysis module can perform a preliminary analysis of the sensor data. The first sub-analysis module can interpret the player's movements in such a way that it can distinguish hitting movements from movements that are not performed to execute a stroke in the racquet sport. If the first sub-analysis module detects a hitting movement, the second sub-analysis module is activated, which in turn is designed to determine the type of stroke performed by the player.

[0011] According to one embodiment, the sensor data are only passed to the second partial analysis module as input data if a striking movement has been detected by the first partial analysis module.

[0012] This allows for the technical advantage of saving computing time and capacity by activating the second sub-analysis module to determine the impact type of the impact movement only if the first sub-analysis module has clearly detected an impact movement. If no impact movement is detected, the second sub-analysis module remains inactive, thus avoiding unnecessary execution of the second sub-analysis module.

[0013] According to one embodiment, the analysis module further comprises a time determination module, wherein the analysis of the sensor data comprises: Determining a time period between successive strokes of the player by the time determination module; and Determine whether a detected stroke is part of a rally based on the time span by the timing module.

[0014] This offers the technical advantage of enabling not only the stroke type of individual stroke movements, but also the analysis of multiple consecutive stroke movements. This allows, for example, an entire rally to be analyzed. Furthermore, it can be determined whether a stroke movement or a stroke executed belongs to an existing rally, which can be used to determine, for example, the length of the completed rallies. This allows for a more precise analysis of the racket sports player's game.

[0015] According to one embodiment, the analysis module further comprises an energy determination module, wherein the analysis of the sensor data comprises: Determining a hitting force of a hitting movement performed by the player by the energy determination module based on the sensor data.

[0016] This provides the technical advantage that, in addition to the type of stroke or the rallies, the strength of the strokes executed can also be analyzed.

[0017] According to one embodiment, the analysis module comprises at least one artificial intelligence trained to recognize movements of the player.

[0018] This allows the technical advantage of using artificial intelligence to enable precise analysis of the sensor data and reliable detection of the racket athlete's movement.

[0019] According to one embodiment, the racket sport includes: badminton, tennis, table tennis.

[0020] This makes it possible to achieve the technical advantage that the method according to the invention can be applied to various racket sports.

[0021] Computing unit configured to execute the method for analyzing a game of a player of a racket sport according to one of the preceding embodiments.

[0022] Computer program product comprising instructions which, when the program is executed by a data processing unit, cause the data processing unit to carry out the method for analyzing a game of a player of a racket sport according to one of the preceding embodiments.

[0023] Embodiments of the invention are described with reference to the following figures. The figures show: Fig. 1 is a schematic representation of a system for analyzing a game of a player of a racquet sport according to one embodiment; Fig. 2 is a further schematic representation of the system for analyzing a game of a player of a racket sport according to another embodiment; Fig. 3 is a flowchart of an apparatus for analyzing a game of a player of a racket sport according to one embodiment; and Fig. 4 a schematic representation of a computer program product.

[0024] Fig. 1 shows a schematic representation of a system 400 for analyzing a game of a player 300 of a racquet sport according to one embodiment.

[0025] In the embodiment shown, the system 400 for analyzing a game of a player 300 in a racket sport comprises at least one acceleration sensor 403. The acceleration sensor 403 is arranged on an arm 301 of the racket athlete 300. In the embodiment shown, the racket athlete 300 is a tennis / badminton / squash player. Alternatively, the invention can also be applied to any other racket sports.

[0026] The system 400 further includes a computing unit 409 on which the analysis module 200 can be executed. In the embodiment shown, the computing unit 409 is embodied in an external mobile device 407. The external mobile device 407 can be, for example, a cell phone, a tablet, a laptop, or a desktop computer.

[0027] In the embodiment shown, the acceleration sensor 403 records sensor data 401 during the game of the player 300 and sends the sensor data 401 to the mobile device 407 via a wireless data transmission 413. The analysis module 200 is configured to carry out the inventive method 100 for analyzing a game of a player 300 in a racquet sport. For this purpose, the analysis module 200 detects movements of the player 300 based on the sensor data 401 and determines at least one game parameter 405 of the player 300 based on the detected movements. This parameter is then displayed to the player 300 or another person, for example, a coach of the player 300.

[0028] According to a further embodiment, the mobile device 407 can also be designed as a smartwatch that the player 300 wears on his wrist. The acceleration sensor 403 can also be integrated into the smartwatch. The player 300 can thus read the displayed game parameters 405 even during the game, during breaks in the game, or during training.

[0029] Game parameters 405 can, for example, include a stroke type executed, such as a forehand, a backhand, a serve, a topspin, a slice, a drop, or other stroke types. Furthermore, game parameter 405 can describe the strength of the strokes executed. Furthermore, game parameter 405 can, for example, include a number of consecutive strokes executed, for example, during a rally. Game parameter 405 can also indicate the length or number of rallies executed. Furthermore, game parameter 405 can classify a playing style of player 300, for example, as an offensive or defensive playing style. The playing style can be determined, for example, by the number of strokes or the stroke type or even the stroke strength, with an offensive playing style, for example, having strokes with a higher stroke strength than a defensive playing style.In addition, the game parameter 405 can indicate an agility of the player 300, for example the frequency with which the player 300 executes strokes directly near the net or the frequency of consecutive alternating forehand and backhand strokes, which indicates a high mobility of the player 300.

[0030] Fig. 2 shows another schematic representation of the system 400 for analyzing a game of a player 300 of a racket sport according to another embodiment.

[0031] In the embodiment shown, the analysis module 200 comprises a first sub-analysis module 201 and a second sub-analysis module 203. The first sub-analysis module 201 serves to perform a preliminary analysis of the sensor data 401 provided as input data. The sensor data 401 is provided as input data to the first sub-analysis module 201 and the second sub-analysis module 203 in overlapping sample windows 411. The overlapping sample windows 411, in which the sensor data 401 is sampled, ensure that no stroke movements executed by the player 300 are ignored in the analysis. The sample rate can be in the range of seconds or microseconds.

[0032] Upon receipt of the sensor data 401, the first partial analysis module 201 first performs a preliminary analysis of the sensor data 401. The first partial analysis module 201 determines the movements performed by the player 300 based on the sensor data 401. Furthermore, the first partial analysis module 201 detects whether the movements performed by the player 300 are striking movements of the racquet sport or are arbitrary movements, such as swinging the arms 301 while running from one playing position to the next.

[0033] If the first sub-analysis module 201 detects that the movement performed by the player 300 is a hitting movement, the sensor data 401 are also provided as input data to the second sub-analysis module 203. The second sub-analysis module 203 is configured to detect a hit type, such as a forehand, a backhand, a serve, a topspin, a slice, a drop, or other hit types of the respective racket sport being played, from the hitting movement.

[0034] The movement results or stroke type results of the first and second partial analysis modules 201, 203 thus determined are then summarized in a prediction result 209. The prediction results 209 are further evaluated in a result evaluation 211, in which, for example, the reliability of the detected movements is determined.

[0035] The result evaluation 211 takes into account the overlap of the sample windows 411 of the sensor data 401 and considers that due to the overlap of the sample windows 411, impact movements may be detected multiple times. The result evaluation 211 takes into account this potentially multiple detection of impact movements, which is based exclusively on the overlap of the sample windows 411.

[0036] In the embodiment shown, the analysis module 200 further comprises a time analysis module 205. The time analysis module 205 is configured to determine a time span between the executed hitting movements. Based on the determined time span between successively executed hitting movements, the time analysis module 205 can further determine whether an executed hitting movement is part of a rally or whether a new rally is started by the executed hitting movement. If the time span between two successively executed hitting movements exceeds a predetermined limit, the subsequent hitting movement of the two hitting movements is assigned to a new rally. The time analysis module 205 can also determine the length of executed rallies, for example the duration and / or the number of strokes within a rally.

[0037] According to the embodiment shown, the analysis module 200 further comprises an energy determination module 207. The energy determination module 207 is configured to determine a hitting force of the executed hits of the player 300 based on the sensor data 401.

[0038] Game statistics can be provided to Player 300 or the other person. These statistics include the number of shots executed, a specific shot type, a specific strength, length, and number of rallies, or other game-relevant parameters. The displayed statistics can also include, if access to additional information is available, statistics on Player 300's heart rate or calorie consumption.

[0039] According to one embodiment, the first and second partial analysis modules 201, 203 are embodied as appropriately trained artificial intelligences. The artificial intelligences can be embodied, for example, as artificial neural networks, or as convolutional neural networks (CNNs), recurrent neural networks (RNNs) with long short-term memory (LSTMs), and / or bidirectional long short-term memory (Bi-LSTMs).

[0040] The analysis module 200 can be trained on training data of different athletes, different genders, playing styles, age groups and playing abilities through supervised or unsupervised learning.

[0041] Alternatively, the first and second partial analysis modules 201, 203 can also be implemented in a common neural network.

[0042] Fig. 3 shows a flowchart of a method 100 for analyzing a game of a player 300 of a racquet sport according to one embodiment.

[0043] To analyze a game of a player 300 of a racket sport, in a first method step 101, sensor data 401 of an acceleration sensor 403, which is arranged on an arm 301 of the player 300, is first received.

[0044] In a further method step 103, the sensor data 401 of the acceleration sensor 403 are analyzed by an analysis module 200 and movements of the player 300 are recognized based on the sensor data 401.

[0045] For this purpose, in a further method step 109, a first partial analysis module 201 detects, based on the sensor data 401, whether the movements performed by the player 300 are hitting movements of the respective racket sport being performed.

[0046] In a further method step 111, a second partial analysis module 203 recognizes the stroke type of the stroke movements performed by the player 300 based on the sensor data 401 if stroke movements were determined by the first partial analysis module 201.

[0047] Furthermore, in a further method step 113, time intervals between successive hitting movements of the player 300 are determined by a time analysis module 205. The time analysis module 205 acts on the prediction results 209 of the first and second sub-analysis modules 201, 203, i.e., on the detected hitting movements, and is configured to determine a time interval between the executed hitting movements.

[0048] In a further method step 115, the time analysis module 205 further detects, based on the time period, whether executed hitting movements are part of a rally or whether a new rally is started by the hitting movements.

[0049] In a further method step 117, an energy determination module 207 further determines a hitting force of the hitting movements performed by the player 300 based on the sensor data 401.

[0050] In a further method step 105, at least one game parameter 405 is determined based on the detected movements of the player 300. The game parameter 405 can include, for example, the stroke types, the stroke strength, the number of strokes in a rally, an agility of the player 300, a playing style, or an intention to score a point.

[0051] In a further method step 107, the determined game parameter 405 is displayed in a corresponding display.

[0052] Fig. 4 shows a schematic representation of a computer program product 500 comprising instructions which, when executed by a data processing unit, cause the data processing unit to carry out the method 100 for analyzing a game of a player 300 of a racket sport.

[0053] In the embodiment shown, the computer program product 500 is stored on a storage medium 501. The storage medium 501 can be any storage medium known from the prior art.

Claims

[1] A computer-implemented method (100) for analyzing a game of a player (300) of a racket sport, comprising: Receiving (101) sensor data (401) of an acceleration sensor (403), wherein the acceleration sensor (403) is arranged on an arm (301) of the player (300); Analyzing (103) the sensor data (401) of the acceleration sensor (403) by an analysis module (200) and detecting movements of the player (300) based on the sensor data (401); Determining (105) at least one game parameter (405) of the player (300) based on the detected movements of the player (300); and displaying (107) the determined game parameter (405). [2] The method (100) of claim 1, wherein the game parameter (405) comprises: Stroke type, stroke strength, number of strokes in a rally, player mobility, playing style. [3] Method (100) according to claim 1 or 2, wherein the analysis module (200) comprises a first partial analysis module (201) and a second partial analysis module (203), and wherein the analysis (103) of the sensor data (401) comprises: Detecting (109) by the first partial analysis module (201) based on the sensor data (401) whether the movement performed by the player (300) is a hitting movement; and Detecting (111) by the second partial analysis module (203) based on the sensor data (401) the stroke type of the stroke movement performed by the player (300) if a stroke movement was detected by the first partial analysis module (201). [4] Method (100) according to claim 3, wherein the sensor data (401) are only passed to the second partial analysis module (203) as input data if a striking movement has been detected by the first partial analysis module (201). [5] Method (100) according to one of the preceding claims, wherein the analysis module (200) further comprises a time analysis module (205), and wherein the analysis (103) of the sensor data (401) comprises: Determining (113) a time period between successive stroke movements of the player (300) by the time analysis module (205); and Determining (115) based on the time period by the time analysis module (205) whether a detected stroke movement is part of a rally. [6] Method (100) according to one of the preceding claims, wherein the analysis module (200) further comprises an energy determination module (207), and wherein the analysis (103) of the sensor data (401) comprises: Determining (117) a striking force of a striking movement performed by the player (300) by the energy determination module (207) based on the sensor data (401). [7] Method (100) according to one of the preceding claims, wherein the analysis module (200) comprises at least one artificial intelligence trained to recognize movements of the player (300). [8] Method (100) according to one of the preceding claims, wherein the racket sport comprises: badminton, tennis, table tennis. [9] Computing unit (409) configured to carry out the method (100) for analyzing a game of a player (300) of a racket sport according to any one of the preceding claims 1 to 8. [10] Computer program product (500) comprising instructions which, when the program is executed by a data processing unit, cause the data processing unit to carry out the method (100) for analyzing a game of a player (300) of a racket sport according to one of the preceding claims 1 to 8.

Images