Tennis racket ball hitting data monitoring system based on intelligent sensor

Abstract

The invention discloses a tennis racket batting data monitoring system based on an intelligent sensor, and particularly relates to a training data monitoring technology, which comprises a four-layer collaborative architecture, an intelligent tennis racket, an intelligent bracelet and a communication module, the four-layer collaborative architecture is composed of a sensor layer, an unmanned aerial vehicle collaborative layer, an edge computing layer and a cloud / user layer; a first communication module is arranged in the unmanned aerial vehicle cooperation layer and used for receiving racket data and environment perception, and the unmanned aerial vehicle cooperation layer communicates with the sensor layer in real time. According to the tennis racket ball hitting data monitoring system based on the intelligent sensor, comprehensive monitoring and analysis of tennis ball hitting data are achieved, the system can capture detailed information at the ball hitting moment, and an accurate ball hitting quality score is generated in combination with physiological data and environmental factors of players; scientific training feedback and competition strategy support are provided for players; the self-adaptive feedback control function of the system can dynamically adjust the training intensity according to the physical state of the player, and the training effect is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of training data monitoring, and particularly to a tennis racket hitting data monitoring system based on intelligent sensors. Background Art

[0002] In tennis training and competitions, accurate data monitoring is crucial for improving players' technical levels and competition performances. Traditional tennis training methods mainly rely on coaches' visual observations and experience guidance, lacking precise capture and analysis of detailed data at the moment of hitting the ball, and it is difficult to formulate and adjust personalized training plans in real time. At the same time, the influence of environmental factors on hitting the ball and the changes in players' own physiological states have not been effectively monitored and comprehensively analyzed.

[0003] Although existing tennis racket-related technologies have developed to a certain extent, most are limited to single functions. For example, some tennis rackets with built-in simple sensors can only record the number of hits or basic swing trajectories, with relatively limited functions, and cannot comprehensively and systematically monitor hitting data and conduct in-depth analysis by combining multi-source information. In addition, there is a lack of effective coordination and integration among these technologies, and it is impossible to form a complete description and accurate evaluation of the hitting process.

[0004] Based on such a background, the present invention aims to provide a tennis racket hitting data monitoring system based on intelligent sensors, which comprehensively monitors and analyzes tennis hitting data through multi-source data fusion and collaborative processing, and provides scientific training feedback and competition strategy support for players. Summary of the Invention

[0005] The main purpose of the present invention is to provide a tennis racket hitting data monitoring system based on intelligent sensors, which can effectively solve the problems in the above-mentioned technology.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0007] A tennis racket hitting data monitoring system based on intelligent sensors includes: a four-layer collaborative architecture, an intelligent tennis racket, an intelligent bracelet, and a communication module;

[0008] The four-layer collaborative architecture is composed of a sensor layer, a drone collaboration layer, an edge computing layer, and a cloud / user layer;

[0009] The intelligent tennis racket is composed of a distributed sensing architecture, a spatial collaborative data flow module, and a cross-device cognitive architecture;

[0010] The spatial collaborative data flow module includes a spatio-temporal calibration engine, a data synchronization protocol, and a heterogeneous data fusion interface, and is used to achieve spatio-temporal alignment and dynamic weight allocation among multiple devices;

[0011] The cross-device cognitive architecture consists of a hierarchical decision-making engine and a cognitive closed-loop mechanism, supporting intention prediction and adaptive feedback;

[0012] The intelligent tennis racket is built-in with a rechargeable lithium battery and supports wireless charging;

[0013] The distributed sensing framework is jointly composed of the racket head area and the racket handle area, and the sensor layer is built inside the racket head area and the racket handle area;

[0014] The racket handle area is also built-in with a data processing module and a communication module 2 for interacting with the drone collaboration layer;

[0015] The drone collaboration layer is built-in with a communication module 1 for receiving racket data and environmental perception. The drone collaboration layer communicates with the sensor layer in real time. The drone collaboration layer tracks the 3D trajectories of the ball and the player through a vision sensor and fuses the data with the sensor layer to establish a hitting dynamics model.

[0016] Preferably, the sensor layer is a physical layer composed of multiple sensor modules, and the multiple sensor modules include:

[0017] A micro MEMS microphone array module, a bio-impedance sensor, and a tactile feedback module integrated inside the racket handle area;

[0018] A motion sensor and a deformation sensor integrated in the racket head area.

[0019] Preferably, the micro MEMS microphone array module can capture the sound fingerprint characteristics at the moment of hitting the ball during hitting, and distinguish the hitting material and the wear state;

[0020] The bio-impedance sensor can monitor the grip force distribution and the conductivity of hand sweat in real time to evaluate the grip stability;

[0021] The tactile feedback module communicates with the bio-impedance sensor, simulates the vibration modes of different hitting types according to the user's grip force state monitored by the bio-impedance sensor, and prompts the user of the applicable scenarios for the current grip force through different vibration types;

[0022] The motion sensor is used to capture the swing trajectory, angular velocity, and hitting acceleration;

[0023] The deformation sensor is attached to the racket frame to monitor the bending deformation of the racket body and evaluate the hitting force transmission efficiency.

[0024] Preferably, the drone collaboration layer includes a drone, and the drone is built-in with a vision sensor, an environmental sensor, an information processing module, and a communication module 1;

[0025] Generate a three-dimensional model of the player's movement using the SLAM algorithm with the sensor layer data and the UAV vision data;

[0026] The UAV communication module and the racket communication module share the GPS / PPS clock signal in real time to ensure the alignment of data timestamps. QR code marking points are arranged on the court to establish the mapping relationship between the racket coordinate system and the UAV global coordinate system, and the data collected by the sensor layer is uploaded to the UAV storage in real time;

[0027] The vision sensor synchronously tracks the 3D trajectories of the ball and the player through the real-time communication of communication module 1 and communication module 2, combines the racket data to reconstruct the complete hitting dynamics model, and generates the 3D space coordinates and the hitting data at each position in the 3D space coordinates;

[0028] The environmental sensor is used to monitor the light and wind speed at the current position in real time and analyze the impact of the environment on hitting.

[0029] Preferably, the hitting dynamics model is fused with the sensor layer and the UAV vision data, and 3D modeling and correction are performed in the following ways:

[0030] The motion sensor and the deformation sensor calculate the initial ball speed and rotation speed after the ball is first hit, and the UAV vision sensor tracks the actual flight trajectory of the ball and corrects and reconstructs the 3D modeling motion trajectory of the racket;

[0031] The correction process is divided into two steps and is carried out according to the following formula:

[0032] The force parameter correction formula when the intelligent racket hits the ball:

[0033]

[0034] The above formula is divided into two parts, where The rotational lift term, is the gravity correction term, where F spin is the rotational force, ρ is the air density, v is the ball speed, C L is the lift coefficient (C L The lift coefficient is obtained by combining wind tunnel experiments with high-speed camera trajectory inversion. The experimental conditions include a wind speed range of 0 - 15 m / s and a rotation rate of 50 - 200 rad / s), r is the radius, πr 2 is the cross-sectional area, m is the ball mass, g is the acceleration due to gravity, β is the rotational axis tilt factor (0 ≤ β ≤ 1, β = 1 means the rotational axis is perpendicular to the ground), and θ is the hitting elevation angle.

[0035] Preferably, the smart bracelet is used to provide the golfer's heart rate, breathing rate, and body temperature, evaluate the golfer's fatigue level, and assist in evaluating the quality of the golf stroke. When the smart bracelet detects that the fatigue index exceeds the threshold, the model automatically reduces the weight of the swing speed;

[0036] When the fatigue index exceeds the threshold, in the weight matrix, α (the weight of the racket data) is adjusted as α new = 0.8·α old .

[0037] Preferably, the edge computing layer includes an edge server. The three-way data is uploaded to the edge server for integration. The data includes the smart racket golf stroke data, the drone scanning data, and the smart bracelet data. The edge server generates a real-time golf stroke quality score, predicted landing point coordinates, and physical fitness status warning through multi-source data fusion, and feeds them back to the user layer;

[0038] The integration process adopts a hierarchical-time-sharing-domain processing framework, dynamically allocates computing resources, and is based on the following core fusion formula:

[0039] (1) Space-time alignment model formula:

[0040]

[0041] where t sync is the unified timestamp, t racket is the local time of the racket, P drone is the drone GPS coordinate, c is the radio frequency signal propagation speed, Δt clock is the atomic clock deviation compensation, P global is the global coordinate system mapping, R and T are the rotation matrix and translation vector (obtained through QR code), λ is the inertial navigation compensation coefficient, and δ IMU is the racket IMU cumulative error.

[0042] (2) Multimodal feature fusion formula:

[0043]

[0044] where is the state estimation vector at time k, F is the state transition matrix, W is the sensor weight matrix, Z k is the observation vector (drone vision + bracelet Bluetooth broadcast), and in addition:

[0045]

[0046] where x, y, z are the spherical triangular coordinate system, v x , v y , v zwhere \(v\) is the velocity component, \(\omega\) is the angular velocity of the sphere rotation, \(HR\) is the heart rate, and FatigueIndex is the fatigue index based on heart rate variability;

[0047] The sensor weight matrix \(W = diag(\alpha,\beta,\gamma)\) is dynamically allocated. During the hitting stage, \(\alpha = 0.6\); during the flight stage, \(\beta = 0.8\); in the fatigue state \(y = 0.3\), and \(diag\) is the diagonal matrix.

[0048] Preferably, the following formula is used to optimize and schedule the intelligent racket hitting data, UAV scanning data, and intelligent bracelet data resources:

[0049]

[0050] where \(f_i\) is the frequency of each processor, \(i = 1\) represents the racket side, \(i = 2\) represents the UAV side, \(i = 3\) represents the bracelet side, \(C\) i is the task computing volume in MIPS, \(w\) i is the task priority weight, and \(\eta\) is the energy consumption coefficient.

[0051] Preferably, the spatial collaborative data flow module is responsible for the synchronization, alignment, and dynamic mapping of multi-source sensor data in the spatial dimension, ensuring data consistency of devices such as rackets, UAVs, and bracelets in the global coordinate system;

[0052] The cross-device cognitive architecture is a distributed intelligent framework based on edge computing. By fusing multi-modal data of intelligent rackets, UAVs, and bracelets, it realizes the understanding of hitting intentions and adaptive feedback control.

[0053] Preferably, the spatio-temporal calibration engine realizes dynamic positioning through VCSEL laser beacons and T0F ranging, and the data synchronization protocol adopts quantum entanglement clock synchronization technology;

[0054] The hierarchical decision-making engine includes a terminal layer, a proximal layer, and a sideline layer, which respectively process hitting event detection and visual tracking. The cognitive closed-loop mechanism dynamically adjusts the training intensity and feedback strategy through the correlation analysis of physiological data and hitting data;

[0055] The terminal layer (racket side) processes hitting event detection at 200 Hz, the proximal layer (UAV side) performs visual target tracking at 30 fps, and the sideline layer (edge server) completes multi-source data fusion.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. Through multi-source data fusion and collaborative processing, the present invention realizes the comprehensive monitoring and analysis of tennis hitting data. The system can capture detailed information at the moment of hitting in real time, combine the physiological data of players and environmental factors, generate accurate hitting quality scores and landing point predictions, and provide scientific training feedback and game strategy support for players. At the same time, the adaptive feedback control function of the system can dynamically adjust the training intensity according to the physical condition of players, effectively improving the training effect and game performance.

[0058] 2. Through various sensors in the intelligent tennis racket, the present invention can capture detailed data such as voiceprint characteristics, grip force distribution, swing trajectory, angular velocity, hitting acceleration, and racket body bending deformation at the moment of hitting in real time, providing a basis for subsequent accurate analysis.

[0059] 3. The present invention fuses the sensor data of the intelligent tennis racket with the visual data of the drone, environmental sensor data, and physiological data of the intelligent bracelet to form a complete hitting dynamics model and player state assessment, making the analysis results more comprehensive and accurate.

[0060] 4. The present invention adopts a distributed intelligent framework based on edge computing with a cross-device cognitive architecture. By fusing multi-modal data, it realizes the understanding of hitting intentions and adaptive feedback control. According to the physiological data and hitting data of players, it dynamically adjusts the training intensity and feedback strategy to avoid overtraining or under-training.

[0061] 5. The environmental sensors in the drone collaboration layer of the present invention can monitor environmental factors such as light and wind speed in real time and analyze their impact on hitting, enabling players to better adapt to different game environments and improving the stability and response ability of the game. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a schematic diagram of the overall monitoring system process of the solution of the present invention;

[0063] Figure 2 It is a schematic diagram of data interaction between the edge computing layer and the cloud / user layer of the present invention;

[0064] Figure 3 It is a schematic diagram of data interaction between the intelligent bracelet and the intelligent tennis racket of the present invention.

[0065] Figure 4 It is a schematic diagram of data interaction between the sensor layer and the drone collaboration layer of the present invention.

[0066] Figure 5 It is a schematic diagram of the communication module interaction process of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0067] To make the technical means, creative features, achieved purposes and effects of the present invention easily understood, the present invention will be further described below in conjunction with specific embodiments.

[0068] Example 1, as Figure 1 and Figure 2 shown, the tennis racket hitting data monitoring system based on intelligent sensors includes: a four-layer collaborative architecture, an intelligent tennis racket, an intelligent bracelet and a communication module;

[0069] Through the four-layer collaborative architecture, combined with the intelligent tennis racket, intelligent bracelet and communication module, comprehensive monitoring and analysis of tennis hitting data are realized.

[0070] The four-layer collaborative architecture consists of a sensor layer, a drone collaboration layer, an edge computing layer and a cloud / user layer;

[0071] During the implementation process, the sensor layer can communicate and exchange data with the drone collaboration layer in real time. The data is transmitted to the edge server through the drone collaboration layer, and after being processed by the edge computing server, dynamic 3D model data is formed and stored in the edge computing server. When needed, the user can use the APP to retrieve it;

[0072] The cloud / user layer receives the processing results of the edge computing layer and presents them to the player in the form of intuitive charts, videos, etc., including information such as hitting quality scores, landing point predictions, and physical fitness status warnings, to assist the player in training and adjusting the game strategy.

[0073] The intelligent tennis racket consists of a distributed sensing architecture, a spatial collaborative data flow module and a cross-device cognitive architecture;

[0074] The spatial collaborative data flow module includes a spatio-temporal calibration engine, a data synchronization protocol and a heterogeneous data fusion interface, which are used to achieve spatio-temporal alignment and dynamic weight distribution between multiple devices;

[0075] The above spatio-temporal calibration engine is explained as follows:

[0076] 1. Core functions

[0077] The spatio-temporal calibration engine is the core technical module for realizing the collaborative work of multiple devices (such as rackets, drones, bracelets) in the intelligent tennis racket system. Its core functions include:

[0078] Time synchronization: Ensure that the data of all devices has a unified timestamp, and eliminate the timing error caused by clock asynchronization.

[0079] Spatial positioning: Establish a global coordinate system and map the local positioning data of different devices to a unified spatial reference system.

[0080] Dynamic Calibration: Real-time correction of spatio-temporal deviations caused by device movement or environmental changes to maintain data consistency.

[0081] 2. Technical Implementation

[0082] (1) Time Synchronization

[0083] Quantum Entanglement Clock Synchronization:

[0084] Using paired atomic clocks (such as rubidium atomic clocks), time synchronization between devices is achieved through the principle of quantum entanglement, with an error less than 1 microsecond (μs).

[0085] Application Scenario: Precise alignment of racket hitting events and drone vision frames to avoid trajectory prediction errors caused by time deviations.

[0086] PTP Protocol (Precision Time Protocol):

[0087] Based on the IEEE 1588 standard, it synchronizes the time of devices within the network through a master-slave clock architecture, supporting nanosecond-level precision.

[0088] Example: The drone obtains UTC time through the GPS module as the master clock source, and the racket and bracelet are synchronized as slave devices.

[0089] (2) Spatial Positioning

[0090] Laser Beacon and TOF Ranging:

[0091] VCSEL Laser Beacon: A vertical-cavity surface-emitting laser (wavelength 850nm) is integrated on top of the racket to emit encoded optical pulses.

[0092] TOF (Time of Flight) Ranging: The drone calculates the relative distance to the racket by the time difference of receiving laser pulses, with an accuracy of ±2 cm.

[0093] Application Scenario: Real-time tracking of the racket's position on the court to assist the drone in adjusting the observation angle.

[0094] SLAM and QR Code Calibration:

[0095] SLAM (Simultaneous Localization and Mapping): The drone constructs a 3D map of the court (accuracy ±5 cm) through visual sensors and lidar.

[0096] QR Code Marking Points: QR codes are arranged at the four corners of the court, and the mapping relationship (rotation matrix R and translation vector T) between the racket's local coordinate system and the drone's global coordinate system is solved through computer vision algorithms (such as the PnP algorithm).

[0097] (3) Dynamic Weight Assignment

[0098] Heterogeneous Data Fusion Interface:

[0099] Dynamically adjust the data weights according to the device reliability (for example, trust the racket sensor during the hitting stage and trust the drone vision during the flight stage).

[0100] Formula example: Weight matrix

[0101] W = diag(0.6, 0.8, 0.3), corresponding to the weights of the racket, drone, and bracelet respectively.

[0102] 3. Practical application example

[0103] Scenario: The drone tracks the hitting trajectory

[0104] Time synchronization: The moment of racket hitting (timestamp t1) is aligned with the drone's captured frame (timestamp t2) through an atomic clock to ensure |t1 - t2| < 1 μs.

[0105] Spatial positioning:

[0106] The laser beacon emits pulses, and the drone calculates the position (x, y, z) of the racket through T0F ranging.

[0107] The SLAM algorithm combines QR code calibration to map the racket coordinates to the global coordinate system.

[0108] Data fusion: The racket sensor data (swing speed, acceleration) and the drone vision data (ball trajectory) are fused under a unified spatio-temporal reference to generate a high-precision hitting model.

[0109] 4. Technical advantages

[0110] The time synchronization error < 1 μs, the spatial positioning error < 2 cm, far exceeding the traditional IMU + GPS solution (the error is usually > 10 cm), the dynamic calibration frequency reaches 30 Hz, adapting to high-speed motion scenarios (such as professional players' serves), and resisting environmental interference (such as light changes, electromagnetic noise) through laser and SLAM technologies.

[0111] 5. Collaboration with other modules

[0112] Link with the edge computing layer: The spatio-temporal calibration results are used as the input for multi-modal data fusion to improve the accuracy of trajectory prediction.

[0113] Collaborate with the tactile feedback module: When the spatio-temporal deviation is detected to exceed the threshold, trigger vibration to prompt the user to calibrate the device.

[0114] The cross-device cognitive architecture consists of a hierarchical decision-making engine and a cognitive closed-loop mechanism, supporting intention prediction and adaptive feedback;

[0115] During the implementation process, a cognitive closed-loop is also added in this embodiment. Through Communication Module 2 and Communication Module 1, real-time communication can be achieved among the drone, the smart bracelet, and the smart tennis racket, and data can be exchanged in real time according to the weight matrix.

[0116] The smart tennis racket is built-in with a rechargeable lithium battery, supporting wired or wireless charging;

[0117] The distributed sensing framework is jointly composed of the racket head area and the racket handle area, and the sensor layer is built inside the racket head area and the racket handle area;

[0118] The racket handle area is also built-in with a data processing module and Communication Module 2 for interacting with the drone cooperation layer;

[0119] The drone cooperation layer is built-in with Communication Module 1 for receiving racket data and environmental perception. The drone cooperation layer communicates with the sensor layer in real time. The drone cooperation layer tracks the 3D trajectories of the ball and the player through a vision sensor, and fuses the data with the sensor layer to establish a hitting dynamics model;

[0120] The sensor layer is a physical layer composed of multiple sensor modules, including:

[0121] A micro MEMS microphone array module, a bioimpedance sensor, and a haptic feedback module integrated inside the racket handle area;

[0122] A motion sensor and a deformation sensor integrated in the racket head area;

[0123] The sensor layer, as a physical layer, is composed of multiple sensor modules, including a micro MEMS microphone array module, a bioimpedance sensor, and a haptic feedback module integrated in the racket handle area, as well as a motion sensor and a deformation sensor integrated in the racket head area. These sensors are used to capture information such as the acoustic fingerprint characteristics, grip force distribution, swing trajectory, angular velocity, hitting acceleration, and racket body bending deformation at the moment of hitting in real time;

[0124] During the hitting process, the micro MEMS microphone array module captures the acoustic fingerprint characteristics of hitting to distinguish the hitting material and wear state; the bioimpedance sensor monitors the grip force distribution and the electrical conductivity of hand sweat in real time to evaluate the grip stability; the haptic feedback module simulates the vibration modes of different hitting types according to the grip force state; the motion sensor captures the swing trajectory, angular velocity, and hitting acceleration; the deformation sensor monitors the racket body bending deformation to evaluate the hitting force transfer efficiency;

[0125] The drone tracks the 3D trajectories of the ball and the player through a vision sensor, and the environmental sensor monitors environmental factors such as light and wind speed, and transmits the data to the edge computing layer in real time.

[0126] The miniature MEMS microphone array module can capture the sound signature characteristics at the moment of hitting the ball during a stroke, and distinguish the hitting material and wear state;

[0127] The bioimpedance sensor can monitor the grip force distribution and the electrical conductivity of hand sweat in real time to evaluate the grip stability of the racket;

[0128] The tactile feedback module communicates with the bioimpedance sensor, and simulates the vibration modes of different hitting types according to the grip force state of the user monitored by the bioimpedance sensor, and prompts the user of the applicable scenarios for the current grip force through different vibration types;

[0129] The motion sensor is used to capture the swing trajectory, angular velocity, and hitting acceleration;

[0130] The deformation sensor is attached to the racket frame to monitor the bending deformation of the racket body and evaluate the hitting force transmission efficiency.

[0131] The drone collaboration layer includes drones, which are built-in with visual sensors, environmental sensors, information processing modules, and communication module one;

[0132] Generate a three-dimensional model of the player's actions by using the SLAM algorithm for the data of the sensor layer and the drone visual data;

[0133] The drone communication module and the racket communication module share the GPS / PPS clock signal in real time to ensure the alignment of data timestamps, and QR code marking points are arranged on the court to establish the mapping relationship between the racket coordinate system and the drone global coordinate system, and the data collected by the sensor layer is uploaded to the drone storage in real time;

[0134] The drone layer includes drones, which are built-in with visual sensors, environmental sensors, information processing modules, and communication module one. The drones track the 3D trajectories of the ball and the player through the visual sensors, and the environmental sensors monitor environmental factors such as light and wind speed, and fuse with the data of the sensor layer to establish a hitting dynamics model.

[0135] The visual sensor synchronously tracks the 3D trajectories of the ball and the player through the real-time communication of communication module one and communication module two, combines the racket data to reconstruct the complete hitting dynamics model, generates 3D spatial coordinates and the hitting data at each position in the 3D spatial coordinates;

[0136] The environmental sensor is used to monitor the light and wind speed at the current position in real time and analyze the impact of the environment on hitting.

[0137] Example two, as Figure 1 and Figure 2 shown, after the above data collection is completed, the collected data is sorted and corrected in the following way. The hitting dynamics model is fused by the sensor layer and the drone visual data, and 3D modeling and correction are performed in the following way:

[0138] After the ball is initially hit, the motion sensor and the deformation sensor calculate the initial ball speed and the rotation speed, and the actual flight trajectory of the ball is tracked by the UAV vision sensor, and the 3D modeling motion trajectory of the racket is corrected and reconstructed;

[0139] The correction process is divided into two steps and is carried out according to the following formula:

[0140] Force parameter correction formula when the intelligent racket hits the ball:

[0141]

[0142] The above formula is divided into two parts, where The rotational lift term, is the gravity correction term, where F spin is the rotational force, ρ is the air density, v is the ball speed, C L is the lift coefficient (C L The lift coefficient is obtained by combining wind tunnel experiments with high-speed camera trajectory inversion. The experimental conditions include a wind speed range of 0 - 15 m / s and a rotation rate of 50 - 200 rad / s), r is the radius, πr 2 is the cross-sectional area, m is the ball mass, g is the acceleration due to gravity, β is the rotation axis tilt factor (0 ≤ β ≤ 1, β = 1 means the rotation axis is perpendicular to the ground), and θ is the hitting elevation angle.

[0143] The motion sensor (IMU) can capture the acceleration (±16g), angular velocity (±2000 dps) and swing trajectory at the moment of hitting the ball in real time;

[0144] The deformation sensor can measure the amount of racket frame bending (0 - 5 mm) and evaluate the force transmission effect;

[0145] Human-machine vision: Use a 4K / 120fps camera and lidar to track the flight trajectory of the ball by the optical flow method and the YOLOv8 model (error < 0.5 m / s).

[0146] C L Calibrated through wind tunnel tests: Under the conditions of wind speed 0 - 15 m / s and rotation rate 50 - 200 rad / s, combined with the high-speed camera inverse trajectory data, fit the lift coefficient curve;

[0147] Establish the C L Nonlinear relationship with wind speed and rotation rate:

[0148]

[0149] where ω is the rotation rate;

[0150] θ is obtained by the binocular vision measurement of the drone for the rotation axis direction, and the elevation angle of the hit is obtained by the racket attitude sensor.

[0151] The 3D trajectory of the ball is generated by the SLAM algorithm, compared with the predicted path of the sensor, and the residual error is calculated.

[0152] The smart bracelet is used to provide the heart rate, breathing rate and body temperature of the player, judge the fatigue level of the player and assist in evaluating the hitting quality. When the smart bracelet detects that the fatigue index exceeds the threshold, the model automatically reduces the weight of the swing speed;

[0153] When the fatigue index exceeds the threshold, α (racket data weight) in the weight matrix is set as α new = 0.8·α old .

[0154] The edge computing layer includes an edge server, and the three-way data is uploaded to the edge server for integration. The data are the hitting data of the smart racket, the scanning data of the drone and the data of the smart bracelet respectively. The edge server generates a real-time hitting quality score, predicted landing point coordinates and physical state warning through multi-source data fusion, and feeds them back to the user layer;

[0155] The edge computing layer consists of an edge server, which is responsible for integrating the hitting data of the smart racket, the scanning data of the drone and the data of the smart bracelet, generating a real-time hitting quality score, predicted landing point coordinates and physical state warning through multi-source data fusion, and feeding them back to the user layer;

[0156] Receives the processing results of the edge computing layer, provides intuitive data display and analysis reports for the players, and assists the players in adjusting their training and game strategies.

[0157] The integration process adopts a hierarchical-time-domain-domain processing framework, dynamically allocates computing resources and is based on the following core fusion formula:

[0158] (1) Space-time alignment model formula:

[0159]

[0160] Among them, t sync is the unified timestamp, t racket is the local time of the racket, P drone is the GPS coordinate of the drone, c is the propagation speed of the radio frequency signal, Δt clock is the atomic clock deviation compensation, P global is the global coordinate system mapping, R and T are the rotation matrix and translation vector (obtained through QR code), λ is the inertial navigation compensation coefficient, δ IMU is the cumulative error of the racket IMU.

[0161] Through the local time t of the racket racket, the GPS coordinates P of the drone drone , the radio frequency signal propagation speed c, and the atomic clock deviation compensation Δt clock , to achieve timestamp alignment between devices (error < 1 μs);

[0162] Using the rotation matrix R, the translation vector T, combined with the accumulated error δ of the racket IMU IMU , to map the racket local coordinate system to the drone global coordinate system (accuracy ±2 cm).

[0163] The above can eliminate the errors caused by device clock asynchronization and spatial reference system differences, ensure the fusion of racket, drone, and bracelet data under the same space-time reference, support millimeter-level hitting point positioning and trajectory reconstruction. For example, the error of sweet spot hit rate analysis is reduced to ±1 mm, and the motion accumulated error is corrected in real time through the inertial navigation compensation coefficient λ to adapt to high-speed swing scenarios.

[0164] (2) Multimodal feature fusion formula:

[0165]

[0166] Among them, is the state estimation vector at time k, F is the state transition matrix, W is the sensor weight matrix, and Z k is the observation vector (drone vision + bracelet Bluetooth broadcast). Additionally:

[0167]

[0168] Among them, x, y, z are the spherical triangle coordinate system, v x , v y , v z are the velocity components, ω is the spherical rotation angular velocity, HR is the heart rate, and FatigueIndex is the fatigue index based on heart rate variability;

[0169] Predict the state at the next moment through the state transition matrix F Combine the sensor weight matrix W to dynamically adjust the data source weight α to avoid overloading misjudgment.

[0170] In the above, the racket data (high-frequency mechanical parameters) provides local details, the drone data (global trajectory) corrects macroscopic deviations, and the bracelet data (physiological state) optimizes the training intensity;

[0171] Additionally, automatically adjust the weight according to the scenario. For example, reduce the racket data weight in the fatigue state (bracelet detects heart rate > 180 bpm) to avoid overloading misjudgment.

[0172] The sensor weight matrix W = diag(α, β, γ) is dynamically allocated, with α = 0.6 during the hitting stage, β = 0.8 during the flight stage, and γ = 0.3 in the fatigue state. diag is a diagonal matrix.

[0173] The following formula is used to optimize the scheduling of the intelligent racket hitting data, drone scanning data, and smart bracelet data resources:

[0174]

[0175] where f i is the frequency of each processor, i = 1 is the racket side, i = 2 is the drone side, i = 3 is the bracelet side, C i is the task computation amount in MIPS, w i is the task priority weight, and η is the energy consumption coefficient.

[0176] The spatial collaborative data flow module is responsible for the synchronization, alignment, and dynamic mapping of multi-source sensor data in the spatial dimension, ensuring data consistency of devices such as the racket, drone, and bracelet in the global coordinate system;

[0177] The cross-device cognitive architecture is based on an edge computing-based distributed intelligent framework. By fusing the multi-modal data of the intelligent racket, drone, and bracelet, it realizes the understanding of hitting intentions and adaptive feedback control.

[0178] The cross-device cognitive architecture is based on an edge computing-based distributed intelligent framework. By fusing multi-modal data, it realizes the understanding of hitting intentions and adaptive feedback control. According to the player's physiological data and hitting data, it dynamically adjusts the training intensity and feedback strategy.

[0179] The spatio-temporal calibration engine realizes dynamic positioning through VCSEL laser beacons and TOF ranging, and the data synchronization protocol uses quantum entanglement clock synchronization technology;

[0180] The hierarchical decision-making engine includes a terminal layer, a proximal layer, and a sideline layer, which respectively process hitting event detection, visual tracking. The cognitive closed-loop mechanism dynamically adjusts the training intensity and feedback strategy through the correlation analysis of physiological data and hitting data;

[0181] The terminal layer (racket side) processes hitting event detection at 200Hz, the proximal layer (drone side) performs visual target tracking at 30fps, and the sideline layer (edge server) completes multi-source data fusion.

[0182] Through multi-source data fusion and collaborative processing, comprehensive monitoring and analysis of tennis hitting data are achieved. The system can capture detailed information at the moment of hitting in real time, combine the physiological data of players and environmental factors, generate accurate hitting quality scores and landing point predictions, and provide scientific training feedback and game strategy support for players. At the same time, the adaptive feedback control function of the system can dynamically adjust the training intensity according to the physical condition of players, effectively improving the training effect and game performance.

[0183] During the implementation of tennis training, players use a tennis racket with a built-in intelligent sensor to practice hitting. The sensor layer of the intelligent tennis racket captures data such as hitting sound patterns, grip strength, swing trajectory, acceleration, and racket body deformation in real time, and transmits the data to the drone collaboration layer through communication module two. The visual sensors in the drone collaboration layer track the 3D trajectories of the ball and the player, and the environmental sensors monitor the light and wind speed of the venue, and upload the data to the edge computing layer in real time. The edge computing layer integrates multi-source data, and generates real-time hitting quality scores, predicted landing point coordinates, and physical condition warnings through the spatio-temporal alignment model formula and the multi-modal feature fusion formula. Players receive feedback information through the mobile device in the user layer, adjust their hitting actions according to the hitting quality score, optimize their hitting strategies based on the landing point prediction, and reasonably arrange the training intensity with reference to the physical condition warning. At the same time, the intelligent bracelet monitors the player's heart rate, breathing frequency, and body temperature. When the fatigue index exceeds the threshold, the system automatically reduces the weight of the swing speed in the model and adjusts the data fusion strategy to provide the player with training suggestions more in line with the current physical condition.

[0184] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A tennis racket hitting data monitoring system based on intelligent sensors, characterized in that Including: A four-layer collaborative architecture, a smart tennis racket, a smart bracelet, and a communication module; The four-layer collaborative architecture consists of a sensor layer, a drone collaboration layer, an edge computing layer, and a cloud / user layer; The smart tennis racket consists of a distributed sensing architecture, a spatial collaborative data stream module, and a cross-device cognitive architecture; The spatial collaborative data stream module includes a spatio-temporal calibration engine, a data synchronization protocol, and a heterogeneous data fusion interface, which are used to achieve spatio-temporal alignment and dynamic weight allocation between multiple devices; The cross-device cognitive architecture consists of a hierarchical decision-making engine and a cognitive closed-loop mechanism, which supports intention prediction and adaptive feedback; The smart tennis racket is built-in with a rechargeable lithium battery and supports wireless charging; The distributed sensing architecture is jointly composed of the racket head area and the racket handle area of the tennis racket, and the sensor layer is built-in inside the racket head area and the racket handle area; The racket handle area is also built-in with a data processing module and a communication module two, which are used to interact with the drone collaboration layer; The drone collaboration layer is built-in with a communication module one, which is used to receive racket data and environmental perception. The drone collaboration layer and the sensor layer communicate in real time. The drone collaboration layer tracks the 3D trajectories of the ball and the player through a vision sensor and fuses the data with the sensor layer to establish a hitting dynamics model.

2. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 1, wherein: The sensor layer is a physical layer composed of multiple sensor modules, and the multiple sensor modules include: A micro MEMS microphone array module, a bio-impedance sensor, and a tactile feedback module integrated inside the racket handle area; A motion sensor and a deformation sensor integrated in the racket head area.

3. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 1, wherein: The micro MEMS microphone array module can capture the sound pattern characteristics at the moment of hitting during hitting to distinguish the hitting material and the wear state; The bio-impedance sensor can monitor the grip force distribution and the electrical conductivity of hand sweat in real time to evaluate the grip stability; The tactile feedback module communicates with the bio-impedance sensor and simulates the vibration modes of different hitting types according to the grip force state of the user monitored by the bio-impedance sensor, and prompts the user of the applicable scenarios of the current grip force through different vibration types; The motion sensor is used to capture the swing trajectory, angular velocity, and hitting acceleration; The deformation sensor is attached to the racket frame to monitor the bending deformation of the racket body and evaluate the hitting force transmission efficiency.

4. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 1, wherein: The drone collaboration layer includes a drone, and the drone is built-in with a vision sensor, an environmental sensor, an information processing module, and a communication module one; Using the SLAM algorithm to generate a three-dimensional model of the player's actions from the data of the sensor layer and the visual data of the drone; The drone communication module and the racket communication module share the GPS / PPS clock signal in real time to ensure the alignment of data timestamps. QR code marking points are arranged on the court to establish the mapping relationship between the racket coordinate system and the drone global coordinate system, and the data collected by the sensor layer is uploaded to the drone storage in real time; The vision sensor synchronously tracks the 3D trajectories of the ball and the player through the real-time communication of Communication Module 1 and Communication Module 2, reconstructs the complete hitting dynamics model in combination with the racket data, and generates the 3D space coordinates and the hitting data at various positions in the 3D space coordinates; The environmental sensor is used to monitor the light and wind speed at the current position in real time and analyze the impact of the environment on hitting.

5. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 4, characterized in that: The hitting dynamics model is fused by the sensor layer and the drone vision data, and 3D modeling and correction are performed in the following ways: The motion sensor and the deformation sensor calculate the initial ball speed and rotation speed after the ball is first hit, and the drone vision sensor tracks the actual flight trajectory of the ball and corrects and reconstructs the 3D modeling motion trajectory of the racket; The correction process is divided into two steps and is carried out according to the following formula: The force parameter correction formula when the intelligent racket hits the ball: The above formula is divided into two parts, where the rotational lift term, is the gravity correction term, where F spin is the rotational force, ρ is the air density, v is the ball speed, C L is the lift coefficient (the CL lift coefficient is obtained by combining wind tunnel experiments with the inversion of the high-speed camera trajectory. The experimental conditions include a wind speed range of 0 - 15 m / s and a rotational rate of 50 - 200 rad / s), r is the radius, πr 2 is the cross-sectional area, m is the ball mass, g is the acceleration due to gravity, β is the rotational axis tilt factor (0 ≤ β ≤ 1, β = 1 indicates that the rotational axis is perpendicular to the ground), and θ is the hitting elevation angle.

6. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 1, characterized in that: The intelligent bracelet is used to provide the heart rate, breathing frequency and body temperature of the player, judge the fatigue degree of the player and assist in evaluating the hitting quality. When the intelligent bracelet detects that the fatigue index exceeds the threshold, the weight of the swing speed of the model is automatically reduced; When the fatigue index exceeds the threshold, α (the weight of racket data) in the weight matrix is adjusted as α new = 0.8·α old .

7. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 1, characterized in that: The edge computing layer includes an edge server, and the three-way data is uploaded to the edge server for integration. The data are the intelligent racket hitting data, the drone scanning data and the intelligent bracelet data respectively. The edge server generates a real-time hitting quality score, a predicted landing point coordinate and a physical fitness status warning through multi-source data fusion, and feeds them back to the user layer; The integration process adopts a hierarchical-time-sharing-domain processing framework, dynamically allocates computing resources and is based on the following core fusion formula: (1) Space-time alignment model formula: Among them, t sync is the unified timestamp, t racket is the local time of the racket, P drone is the GPS coordinates of the drone, c is the propagation speed of the radio frequency signal, Δt clock is the atomic clock deviation compensation, P global is the global coordinate system mapping, R and T are the rotation matrix and translation vector (obtained through the QR code), λ is the inertial navigation compensation coefficient, δ IMU is the cumulative error of the racket IMU. (2) Multimodal feature fusion formula: Among them, is the state estimation vector at time k, F is the state transition matrix, W is the sensor weight matrix, and Z k is the observation vector (drone vision + bracelet Bluetooth broadcast). Additionally: where x, y, z are the spherical triangle coordinate system, v x , v y , v z are the velocity components, ω is the angular velocity of the sphere rotation, HR is the heart rate, and FatigueIndex is the fatigue index based on heart rate variability; the sensor weight matrix W = diag(α, β, γ) is dynamically allocated, α = 0.6 during the hitting stage, β = 0.8 during the flight stage, and γ = 0.3 in the fatigue state, where diag is the diagonal matrix.

8. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 7, characterized in that: The following formula is used to optimize the scheduling of the intelligent racket hitting data, the drone scanning data and the intelligent bracelet data resources: Among them, f i is the frequency of each processor. i = 1 represents the racket side, i = 2 represents the drone side, and i = 3 represents the bracelet side. C i is the task computing volume MIPS, w i is the task priority weight, and η is the energy consumption coefficient.

9. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 1, characterized in that: The spatial collaborative data flow module is responsible for the synchronization, alignment and dynamic mapping of multi-source sensor data in the spatial dimension to ensure the data consistency of devices such as rackets, drones and bracelets in the global coordinate system; The cross-device cognitive architecture is based on the distributed intelligent framework of edge computing. By fusing the multimodal data of intelligent rackets, drones and bracelets, it realizes the understanding of hitting intentions and adaptive feedback control.

10. The tennis racket hitting data monitoring system based on intelligent sensors according to claim 9, characterized in that: The spatio-temporal calibration engine realizes dynamic positioning through VCSEL laser beacons and TOF ranging, and the data synchronization protocol adopts quantum entanglement clock synchronization technology; The hierarchical decision-making engine includes a terminal layer, a proximal layer and a sideline layer, which respectively process hitting event detection and visual tracking. The cognitive closed-loop mechanism dynamically adjusts the training intensity and feedback strategy through the correlation analysis of physiological data and hitting data; The terminal layer (racket end) processes hitting event detection at 200 Hz, the proximal layer (drone end) performs visual target tracking at 30 fps, and the sideline layer (edge server) completes multi-source data fusion.

Images